Free ebooks ==> www.Ebook777.

com

www.Ebook777.com
 Foundations of Cryptography

Cryptography is concerned with the conceptualization, definition, and construction of

computing systems that address security concerns. The design of cryptographic systems
must be based on firm foundations. Foundations of Cryptography presents a rigorous and
systematic treatment of foundational issues: defining cryptographic tasks and solving
new cryptographic problems using existing tools. The emphasis is on the clarification
of fundamental concepts and on demonstrating the feasibility of solving several central
cryptographic problems, as opposed to describing ad hoc approaches.
This second volume contains a rigorous treatment of three basic applications: en-
cryption, signatures, and general cryptographic protocols. It builds on the previous
volume, which provides a treatment of one-way functions, pseudorandomness, and
zero-knowledge proofs. It is suitable for use in a graduate course on cryptography and
as a reference book for experts. The author assumes basic familiarity with the design
and analysis of algorithms; some knowledge of complexity theory and probability is
also useful.

Oded Goldreich is Professor of Computer Science at the Weizmann Institute of Science

and incumbent of the Meyer W. Weisgal Professorial Chair. An active researcher, he
has written numerous papers on cryptography and is widely considered to be one of
the world experts in the area. He is an editor of Journal of Cryptology and SIAM
Journal on Computing and the author of Modern Cryptography, Probabilistic Proofs
and Pseudorandomness.
Free ebooks ==> www.Ebook777.com

www.Ebook777.com
Foundations of Cryptography
II Basic Applications

Oded Goldreich
Weizmann Institute of Science
 Free ebooks ==> www.Ebook777.com

CAMBRIDGE UNIVERSITY PRESS

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi

Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org
Information on this title: www.cambridge.org/9780521119917

© Oded Goldreich 2004

This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.

First published 2004

This digitally printed version 2009

A catalogue record for this publication is available from the British Library

ISBN 978-0-521-83084-3 hardback

ISBN 978-0-521-11991-7 paperback

www.Ebook777.com
To Dana
Free ebooks ==> www.Ebook777.com

www.Ebook777.com
 Contents
II Basic Applications

List of Figures page xi

Preface xiii
Acknowledgments xxi

5 Encryption Schemes 373

5.1. The Basic Setting 374
5.1.1. Private-Key Versus Public-Key Schemes 375
5.1.2. The Syntax of Encryption Schemes 376
5.2. Definitions of Security 378
5.2.1. Semantic Security 379
5.2.2. Indistinguishability of Encryptions 382
5.2.3. Equivalence of the Security Definitions 383
5.2.4. Multiple Messages 389
5.2.5.* A Uniform-Complexity Treatment 394
5.3. Constructions of Secure Encryption Schemes 403
5.3.1.* Stream-Ciphers 404
5.3.2. Preliminaries: Block-Ciphers 408
5.3.3. Private-Key Encryption Schemes 410
5.3.4. Public-Key Encryption Schemes 413
5.4.* Beyond Eavesdropping Security 422
5.4.1. Overview 422
5.4.2. Key-Dependent Passive Attacks 425
5.4.3. Chosen Plaintext Attack 431
5.4.4. Chosen Ciphertext Attack 438
5.4.5. Non-Malleable Encryption Schemes 470
5.5. Miscellaneous 474
5.5.1. On Using Encryption Schemes 474
5.5.2. On Information-Theoretic Security 476
5.5.3. On Some Popular Schemes 477
vii
 Free ebooks ==> www.Ebook777.com
CONTENTS

5.5.4. Historical Notes 478

5.5.5. Suggestions for Further Reading 480
5.5.6. Open Problems 481
5.5.7. Exercises 481

6 Digital Signatures and Message Authentication 497

6.1. The Setting and Definitional Issues 498
6.1.1. The Two Types of Schemes: A Brief Overview 498
6.1.2. Introduction to the Unified Treatment 499
6.1.3. Basic Mechanism 501
6.1.4. Attacks and Security 502
6.1.5.* Variants 505
6.2. Length-Restricted Signature Scheme 507
6.2.1. Definition 507
6.2.2. The Power of Length-Restricted Signature Schemes 508
6.2.3.* Constructing Collision-Free Hashing Functions 516
6.3. Constructions of Message-Authentication Schemes 523
6.3.1. Applying a Pseudorandom Function to the Document 523
6.3.2.* More on Hash-and-Hide and State-Based MACs 531
6.4. Constructions of Signature Schemes 537
6.4.1. One-Time Signature Schemes 538
6.4.2. From One-Time Signature Schemes to General Ones 543
6.4.3.* Universal One-Way Hash Functions and Using Them 560
6.5.* Some Additional Properties 575
6.5.1. Unique Signatures 575
6.5.2. Super-Secure Signature Schemes 576
6.5.3. Off-Line/On-Line Signing 580
6.5.4. Incremental Signatures 581
6.5.5. Fail-Stop Signatures 583
6.6. Miscellaneous 584
6.6.1. On Using Signature Schemes 584
6.6.2. On Information-Theoretic Security 585
6.6.3. On Some Popular Schemes 586
6.6.4. Historical Notes 587
6.6.5. Suggestions for Further Reading 589
6.6.6. Open Problems 590
6.6.7. Exercises 590

7 General Cryptographic Protocols 599

7.1. Overview 600
7.1.1. The Definitional Approach and Some Models 601
7.1.2. Some Known Results 607
7.1.3. Construction Paradigms 609

viii

www.Ebook777.com
 CONTENTS

7.2.* The Two-Party Case: Definitions 615

7.2.1. The Syntactic Framework 615
7.2.2. The Semi-Honest Model 619
7.2.3. The Malicious Model 626
7.3.* Privately Computing (Two-Party) Functionalities 634
7.3.1. Privacy Reductions and a Composition Theorem 636
7.3.2. The OTk1 Protocol: Definition and Construction 640
7.3.3. Privately Computing c1 + c2 = (a1 + a2 ) · (b1 + b2 ) 643
7.3.4. The Circuit Evaluation Protocol 645
7.4.* Forcing (Two-Party) Semi-Honest Behavior 650
7.4.1. The Protocol Compiler: Motivation and Overview 650
7.4.2. Security Reductions and a Composition Theorem 652
7.4.3. The Compiler: Functionalities in Use 657
7.4.4. The Compiler Itself 681
7.5.* Extension to the Multi-Party Case 693
7.5.1. Definitions 694
7.5.2. Security in the Semi-Honest Model 701
7.5.3. The Malicious Models: Overview and Preliminaries 708
7.5.4. The First Compiler: Forcing Semi-Honest Behavior 714
7.5.5. The Second Compiler: Effectively Preventing Abort 729
7.6.* Perfect Security in the Private Channel Model 741
7.6.1. Definitions 742
7.6.2. Security in the Semi-Honest Model 743
7.6.3. Security in the Malicious Model 746
7.7. Miscellaneous 747
7.7.1.* Three Deferred Issues 747
7.7.2.* Concurrent Executions 752
7.7.3. Concluding Remarks 755
7.7.4. Historical Notes 756
7.7.5. Suggestions for Further Reading 757
7.7.6. Open Problems 758
7.7.7. Exercises 759
Appendix C: Corrections and Additions to Volume 1 765
C.1. Enhanced Trapdoor Permutations 765
C.2. On Variants of Pseudorandom Functions 768
C.3. On Strong Witness Indistinguishability 768
C.3.1. On Parallel Composition 769
C.3.2. On Theorem 4.6.8 and an Afterthought 770
C.3.3. Consequences 771
C.4. On Non-Interactive Zero-Knowledge 772
C.4.1. On NIZKs with Efficient Prover Strategies 772
C.4.2. On Unbounded NIZKs 773
C.4.3. On Adaptive NIZKs 774

ix
 Free ebooks ==> www.Ebook777.com
CONTENTS

C.5. Some Developments Regarding Zero-Knowledge 775

C.5.1. Composing Zero-Knowledge Protocols 775
C.5.2. Using the Adversary’s Program in the Proof of Security 780
C.6. Additional Corrections and Comments 783
C.7. Additional Mottoes 784

Bibliography 785
Index 795

Note: Asterisks indicate advanced material.

www.Ebook777.com
 List of Figures

0.1 Organization of this work page xvi

0.2 Rough organization of this volume xvii
0.3 Plan for one-semester course on Foundations of Cryptography xviii
5.1 Private-key encryption schemes: an illustration 375
5.2 Public-key encryption schemes: an illustration 376
6.1 Message-authentication versus signature schemes 500
6.2 Collision-free hashing via block-chaining (for t = 7) 519
6.3 Collision-free hashing via tree-chaining (for t = 8) 522
6.4 Authentication-trees: the basic authentication step 546
6.5 An authentication path for nodes 010 and 011 547
7.1 Secure protocols emulate a trusted party: an illustration 601
7.2 The functionalities used in the compiled protocol 658
7.3 Schematic depiction of a canonical protocol 690

xi
Free ebooks ==> www.Ebook777.com

www.Ebook777.com
 Preface

It is possible to build a cabin with no foundations,

but not a lasting building.
Eng. Isidor Goldreich (1906–1995)

Cryptography is concerned with the construction of schemes that withstand any abuse.
Such schemes are constructed so as to maintain a desired functionality, even under
malicious attempts aimed at making them deviate from their prescribed functionality.
The design of cryptographic schemes is a very difficult task. One cannot rely on
intuitions regarding the typical state of the environment in which the system operates.
For sure, the adversary attacking the system will try to manipulate the environment into
untypical states. Nor can one be content with countermeasures designed to withstand
specific attacks because the adversary (which acts after the design of the system is
completed) will try to attack the schemes in ways that are typically different from the
ones envisioned by the designer. The validity of the foregoing assertions seems self-
evident; still, some people hope that in practice, ignoring these tautologies will not result
in actual damage. Experience shows that these hopes rarely come true; cryptographic
schemes based on make-believe are broken, typically sooner than later.
In view of these assertions, we believe that it makes little sense to make assumptions
regarding the specific strategy that the adversary may use. The only assumptions that
can be justified refer to the computational abilities of the adversary. Furthermore,
it is our opinion that the design of cryptographic systems has to be based on firm
foundations, whereas ad hoc approaches and heuristics are a very dangerous way to
go. A heuristic may make sense when the designer has a very good idea about the
environment in which a scheme is to operate, yet a cryptographic scheme has to operate
in a maliciously selected environment that typically transcends the designer’s view.
This work is aimed at presenting firm foundations for cryptography. The foundations
of cryptography are the paradigms, approaches, and techniques used to conceptualize,
define, and provide solutions to natural “security concerns.” We will present some of
these paradigms, approaches, and techniques, as well as some of the fundamental results

xiii
 Free ebooks ==> www.Ebook777.com
PREFACE

obtained using them. Our emphasis is on the clarification of fundamental concepts and
on demonstrating the feasibility of solving several central cryptographic problems.
Solving a cryptographic problem (or addressing a security concern) is a two-stage
process consisting of a definitional stage and a constructive stage. First, in the defini-
tional stage, the functionality underlying the natural concern is to be identified, and an
adequate cryptographic problem has to be defined. Trying to list all undesired situations
is infeasible and prone to error. Instead, one should define the functionality in terms of
operation in an imaginary ideal model, and require a candidate solution to emulate this
operation in the real, clearly defined model (which specifies the adversary’s abilities).
Once the definitional stage is completed, one proceeds to construct a system that satis-
fies the definition. Such a construction may use some simpler tools, and its security is
proven relying on the features of these tools. In practice, of course, such a scheme may
also need to satisfy some specific efficiency requirements.
This work focuses on several archetypical cryptographic problems (e.g., encryption
and signature schemes) and on several central tools (e.g., computational difficulty,
pseudorandomness, and zero-knowledge proofs). For each of these problems (resp.,
tools), we start by presenting the natural concern underlying it (resp., its intuitive
objective), then define the problem (resp., tool), and finally demonstrate that the problem
may be solved (resp., the tool can be constructed). In the last step, our focus is on demon-
strating the feasibility of solving the problem, not on providing a practical solution. As
a secondary concern, we typically discuss the level of practicality (or impracticality)
of the given (or known) solution.

Computational Difficulty
The specific constructs mentioned earlier (as well as most constructs in this area) can
exist only if some sort of computational hardness exists. Specifically, all these problems
and tools require (either explicitly or implicitly) the ability to generate instances of hard
problems. Such ability is captured in the definition of one-way functions (see further
discussion in Section 2.1). Thus, one-way functions are the very minimum needed for
doing most sorts of cryptography. As we shall see, one-way functions actually suffice for
doing much of cryptography (and the rest can be done by augmentations and extensions
of the assumption that one-way functions exist).
Our current state of understanding of efficient computation does not allow us to prove
that one-way functions exist. In particular, the existence of one-way functions implies
that N P is not contained in BPP ⊇ P (not even “on the average”), which would
resolve the most famous open problem of computer science. Thus, we have no choice
(at this stage of history) but to assume that one-way functions exist. As justification for
this assumption, we may only offer the combined beliefs of hundreds (or thousands) of
researchers. Furthermore, these beliefs concern a simply stated assumption, and their
validity follows from several widely believed conjectures that are central to various
fields (e.g., the conjecture that factoring integers is hard is central to computational
number theory).
Since we need assumptions anyhow, why not just assume what we want (i.e., the
existence of a solution to some natural cryptographic problem)? Well, first we need
xiv

www.Ebook777.com
 PREFACE

to know what we want: As stated earlier, we must first clarify what exactly we want;
that is, we must go through the typically complex definitional stage. But once this stage
is completed, can we just assume that the definition derived can be met? Not really.
Once a definition is derived, how can we know that it can be met at all? The way to
demonstrate that a definition is viable (and so the intuitive security concern can be
satisfied at all) is to construct a solution based on a better-understood assumption (i.e.,
one that is more common and widely believed). For example, looking at the definition
of zero-knowledge proofs, it is not a priori clear that such proofs exist at all (in a
non-trivial sense). The non-triviality of the notion was first demonstrated by presenting
a zero-knowledge proof system for statements regarding Quadratic Residuosity that
are believed to be hard to verify (without extra information). Furthermore, contrary to
prior beliefs, it was later shown that the existence of one-way functions implies that
any NP-statement can be proven in zero-knowledge. Thus, facts that were not at all
known to hold (and were even believed to be false) were shown to hold by reduction to
widely believed assumptions (without which most of modern cryptography collapses
anyhow). To summarize, not all assumptions are equal, and so reducing a complex,
new, and doubtful assumption to a widely believed simple (or even merely simpler)
assumption is of great value. Furthermore, reducing the solution of a new task to the
assumed security of a well-known primitive typically means providing a construction
that, using the known primitive, solves the new task. This means that we not only know
(or assume) that the new task is solvable but also have a solution based on a primitive
that, being well known, typically has several candidate implementations.

Structure and Prerequisites

Our aim is to present the basic concepts, techniques, and results in cryptography. As
stated earlier, our emphasis is on the clarification of fundamental concepts and the rela-
tionship among them. This is done in a way independent of the particularities of some
popular number-theoretic examples. These particular examples played a central role in
the development of the field and still offer the most practical implementations of all
cryptographic primitives, but this does not mean that the presentation has to be linked
to them. On the contrary, we believe that concepts are best clarified when presented
at an abstract level, decoupled from specific implementations. Thus, the most relevant
background for this work is provided by basic knowledge of algorithms (including
randomized ones), computability, and elementary probability theory. Background on
(computational) number theory, which is required for specific implementations of cer-
tain constructs, is not really required here (yet a short appendix presenting the most
relevant facts is included in the first volume so as to support the few examples of
implementations presented here).

Organization of the Work. This work is organized in two parts (see Figure 0.1): Basic
Tools and Basic Applications. The first volume (i.e., [108]) contains an introductory
chapter as well as the first part (Basic Tools), which consists of chapters on computa-
tional difficulty (one-way functions), pseudorandomness, and zero-knowledge proofs.
These basic tools are used for the Basic Applications of the second part (i.e., the current
xv
 Free ebooks ==> www.Ebook777.com
PREFACE

Volume 1: Introduction and Basic Tools

Chapter 1: Introduction
Chapter 2: Computational Difficulty (One-Way Functions)
Chapter 3: Pseudorandom Generators
Chapter 4: Zero-Knowledge Proof Systems
Volume 2: Basic Applications
Chapter 5: Encryption Schemes
Chapter 6: Digital Signatures and Message Authentication
Chapter 7: General Cryptographic Protocols
Figure 0.1: Organization of this work.

volume), which consists of chapters on Encryption Schemes, Digital Signatures and

Message Authentication, and General Cryptographic Protocols.
The partition of the work into two parts is a logical one. Furthermore, it has offered
us the advantage of publishing the first part before the completion of the second part.
Originally, a third part, entitled Beyond the Basics, was planned. That part was to
have discussed the effect of Cryptography on the rest of Computer Science (and, in
particular, complexity theory), as well as to have provided a treatment of a variety
of more advanced security concerns. In retrospect, we feel that the first direction is
addressed in [106], whereas the second direction is more adequate for a collection of
surveys.

Organization of the Current Volume. The current (second) volume consists of three
chapters that treat encryption schemes, digital signatures and message authentication,
and general cryptographic protocols, respectively. Also included is an appendix that pro-
vides corrections and additions to Volume 1. Figure 0.2 depicts the high-level structure
of the current volume. Inasmuch as this volume is a continuation of the first (i.e., [108]),
one numbering system is used for both volumes (and so the first chapter of the cur-
rent volume is referred to as Chapter 5). This allows a simple referencing of sections,
definitions, and theorems that appear in the first volume (e.g., Section 1.3 presents
the computational model used throughout the entire work). The only exception to this
rule is the use of different bibliographies (and consequently a different numbering of
bibliographic entries) in the two volumes.
Historical notes, suggestions for further reading, some open problems, and some
exercises are provided at the end of each chapter. The exercises are mostly designed to
help and test the basic understanding of the main text, not to test or inspire creativity.
The open problems are fairly well known; still, we recommend a check on their current
status (e.g., in our updated notices web site).

Web Site for Notices Regarding This Work. We intend to maintain a web site listing
corrections of various types. The location of the site is

http://www.wisdom.weizmann.ac.il/∼oded/foc-book.html
xvi

www.Ebook777.com
 PREFACE

Chapter 5: Encryption Schemes

The Basic Setting (Sec. 5.1)
Definitions of Security (Sec. 5.2)
Constructions of Secure Encryption Schemes (Sec. 5.3)
Advanced Material (Secs. 5.4 and 5.5.1–5.5.3)
Chapter 6: Digital Signatures and Message Authentication
The Setting and Definitional Issues (Sec. 6.1)
Length-Restricted Signature Scheme (Sec. 6.2)
Basic Constructions (Secs. 6.3 and 6.4)
Advanced Material (Secs. 6.5 and 6.6.1–6.6.3)
Chapter 7: General Cryptographic Protocols
Overview (Sec. 7.1)
Advanced Material (all the rest):
The Two-Party Case (Sec. 7.2–7.4)
The Multi-Party Case (Sec. 7.5 and 7.6)
Appendix C: Corrections and Additions to Volume 1
Bibliography and Index
Figure 0.2: Rough organization of this volume.

Using This Work

This work is intended to serve as both a textbook and a reference text. That is, it is
aimed at serving both the beginner and the expert. In order to achieve this aim, the
presentation of the basic material is very detailed so as to allow a typical undergraduate
in Computer Science to follow it. An advanced student (and certainly an expert) will
find the pace (in these parts) far too slow. However, an attempt was made to allow the
latter reader to easily skip details obvious to him/her. In particular, proofs are typically
presented in a modular way. We start with a high-level sketch of the main ideas and only
later pass to the technical details. Passage from high-level descriptions to lower-level
details is typically marked by phrases such as “details follow.”
In a few places, we provide straightforward but tedious details in indented para-
graphs such as this one. In some other (even fewer) places, such paragraphs provide
technical proofs of claims that are of marginal relevance to the topic of the work.

More advanced material is typically presented at a faster pace and with fewer details.
Thus, we hope that the attempt to satisfy a wide range of readers will not harm any of
them.

Teaching. The material presented in this work, on the one hand, is way beyond what
one may want to cover in a course and, on the other hand, falls very short of what one
may want to know about Cryptography in general. To assist these conflicting needs, we
make a distinction between basic and advanced material and provide suggestions for
further reading (in the last section of each chapter). In particular, sections, subsections,
and subsubsections marked by an asterisk (*) are intended for advanced reading.
xvii
 Free ebooks ==> www.Ebook777.com
PREFACE

Depending on the class, each lecture consists of 50–90 minutes. Lectures

1–15 are covered by the first volume. Lectures 16–28 are covered by the
current (second) volume.

Lecture 1: Introduction, Background, etc. (depending on class)

Lectures 2–5: Computational Difficulty (One-Way Functions)
Main: Definition (Sec. 2.2), Hard-Core Predicates (Sec. 2.5)
Optional: Weak Implies Strong (Sec. 2.3), and Secs. 2.4.2–2.4.4
Lectures 6–10: Pseudorandom Generators
Main: Definitional Issues and a Construction (Secs. 3.2–3.4)
Optional: Pseudorandom Functions (Sec. 3.6)
Lectures 11–15: Zero-Knowledge Proofs
Main: Some Definitions and a Construction (Secs. 4.2.1, 4.3.1, 4.4.1–4.4.3)
Optional: Secs. 4.2.2, 4.3.2, 4.3.3–4.3.4, 4.4.4
Lectures 16–20: Encryption Schemes
Main: Definitions and Constructions (Secs. 5.1, 5.2.1–5.2.4, 5.3.2–5.3.4)
Optional: Beyond Passive Notions of Security (Overview, Sec. 5.4.1)
Lectures 21–24: Signature Schemes
Definitions and Constructions (Secs. 6.1, 6.2.1–6.2.2, 6.3.1.1, 6.4.1–6.4.2)
Lectures 25–28: General Cryptographic Protocols
The Definitional Approach and a General Construction (Overview, Sec. 7.1).
Figure 0.3: Plan for one-semester course on Foundations of Cryptography.

This work is intended to provide all material required for a course on Foundations
of Cryptography. For a one-semester course, the teacher will definitely need to skip all
advanced material (marked by an asterisk) and perhaps even some basic material; see
the suggestions in Figure 0.3. Depending on the class, this should allow coverage of the
basic material at a reasonable level (i.e., all material marked as “main” and some of the
“optional”). This work can also serve as a textbook for a two-semester course. In such
a course, one should be able to cover the entire basic material suggested in Figure 0.3,
and even some of the advanced material.

Practice. The aim of this work is to provide sound theoretical foundations for cryp-
tography. As argued earlier, such foundations are necessary for any sound practice of
cryptography. Indeed, practice requires more than theoretical foundations, whereas the
current work makes no attempt to provide anything beyond the latter. However, given a
sound foundation, one can learn and evaluate various practical suggestions that appear
elsewhere (e.g., in [149]). On the other hand, lack of sound foundations results in an
inability to critically evaluate practical suggestions, which in turn leads to unsound

xviii

www.Ebook777.com
 PREFACE

decisions. Nothing could be more harmful to the design of schemes that need to with-
stand adversarial attacks than misconceptions about such attacks.

Relationship to Another Book by the Author

A frequently asked question refers to the relationship of the current work to my text
Modern Cryptography, Probabilistic Proofs and Pseudorandomness [106]. That text
consists of three brief introductions to the related topics in its title. Specifically, Chapter 1
of [106] provides a brief (i.e., 30-page) summary of the current work. The other two
chapters of [106] provide a wider perspective on two topics mentioned in the current
work (i.e., Probabilistic Proofs and Pseudorandomness). Further comments on the latter
aspect are provided in the relevant chapters of the first volume of the current work
(i.e., [108]).

A Comment Regarding the Current Volume

There are no privileges without duties.

Adv. Klara Goldreich-Ingwer (1912–2004)

Writing the first volume was fun. In comparison to the current volume, the definitions,
constructions, and proofs in the first volume were relatively simple and easy to write.
Furthermore, in most cases, the presentation could safely follow existing texts. Conse-
quently, the writing effort was confined to reorganizing the material, revising existing
texts, and augmenting them with additional explanations and motivations.
Things were quite different with respect to the current volume. Even the simplest
notions defined in the current volume are more complex than most notions treated in the
first volume (e.g., contrast secure encryption with one-way functions or secure protocols
with zero-knowledge proofs). Consequently, the definitions are more complex, and
many of the constructions and proofs are more complex. Furthermore, in most cases,
the presentation could not follow existing texts. Indeed, most effort had to be (and was)
devoted to the actual design of constructions and proofs, which were only inspired by
existing texts.
The mere fact that writing this volume required so much effort may imply that this
volume will be very valuable: Even experts may be happy to be spared the hardship of
trying to understand this material based on the original research manuscripts.

xix
Free ebooks ==> www.Ebook777.com

www.Ebook777.com
 Acknowledgments

. . . very little do we have and inclose which we can call our own in the
deep sense of the word. We all have to accept and learn, either from our
predecessors or from our contemporaries. Even the greatest genius would
not have achieved much if he had wished to extract everything from inside
himself. But there are many good people, who do not understand this,
and spend half their lives wondering in darkness with their dreams of
originality. I have known artists who were proud of not having followed
any teacher and of owing everything only to their own genius. Such fools!
Goethe, Conversations with Eckermann, 17.2.1832

First of all, I would like to thank three remarkable people who had a tremendous
influence on my professional development: Shimon Even introduced me to theoretical
computer science and closely guided my first steps. Silvio Micali and Shafi Goldwasser
led my way in the evolving foundations of cryptography and shared with me their
constant efforts for further developing these foundations.
I have collaborated with many researchers, yet I feel that my collaboration with
Benny Chor and Avi Wigderson had the most important impact on my professional
development and career. I would like to thank them both for their indispensable contri-
bution to our joint research and for the excitement and pleasure I had when collaborating
with them.
Leonid Levin deserves special thanks as well. I had many interesting discussions
with Leonid over the years, and sometimes it took me too long to realize how helpful
these discussions were.
Special thanks also to four of my former students, from whom I have learned a lot
(especially regarding the contents of this volume): to Boaz Barak for discovering the
unexpected power of non-black-box simulations, to Ran Canetti for developing defini-
tions and composition theorems for secure multi-party protocols, to Hugo Krawczyk
for educating me about message authentication codes, and to Yehuda Lindell for signif-
icant simplification of the construction of a posteriori CCA (which enables a feasible
presentation).
xxi
 Free ebooks ==> www.Ebook777.com
ACKNOWLEDGMENTS

Next, I’d like to thank a few colleagues and friends with whom I had significant
interaction regarding Cryptography and related topics. These include Noga Alon,
Hagit Attiya, Mihir Bellare, Ivan Damgard, Uri Feige, Shai Halevi, Johan Hastad,
Amir Herzberg, Russell Impagliazzo, Jonathan Katz, Joe Kilian, Eyal Kushilevitz,
Yoad Lustig, Mike Luby, Daniele Micciancio, Moni Naor, Noam Nisan, Andrew
Odlyzko, Yair Oren, Rafail Ostrovsky, Erez Petrank, Birgit Pfitzmann, Omer Reingold,
Ron Rivest, Alon Rosen, Amit Sahai, Claus Schnorr, Adi Shamir, Victor Shoup,
Madhu Sudan, Luca Trevisan, Salil Vadhan, Ronen Vainish, Yacob Yacobi, and David
Zuckerman.
Even assuming I did not forget people with whom I had significant interaction on
topics touching upon this book, the list of people I’m indebted to is far more extensive.
It certainly includes the authors of many papers mentioned in the reference list. It also
includes the authors of many Cryptography-related papers that I forgot to mention, and
the authors of many papers regarding the Theory of Computation at large (a theory
taken for granted in the current book).
Finally, I would like to thank Boaz Barak, Alex Healy, Vlad Kolesnikov, Yehuda
Lindell, and Minh-Huyen Nguyen for reading parts of this manuscript and pointing out
various difficulties and errors.

xxii

www.Ebook777.com
 CHAPTER FIVE

Encryption Schemes

Up to the 1970s, Cryptography was understood as the art of building encryption

schemes, that is, the art of constructing schemes allowing secret data exchange over
insecure channels. Since the 1970s, other tasks (e.g., signature schemes) have been
recognized as falling within the domain of Cryptography (and even being at least as
central to Cryptography). Yet the construction of encryption schemes remains, and is
likely to remain, a central enterprise of Cryptography.
In this chapter we review the well-known notions of private-key and public-key
encryption schemes. More importantly, we define what is meant by saying that such
schemes are secure. This definitional treatment is a cornerstone of the entire area,
and much of this chapter is devoted to various aspects of it. We also present several
constructions of secure (private-key and public-key) encryption schemes. It turns out
that using randomness during the encryption process (i.e., not only at the key-generation
phase) is essential to security.

Organization. Our main treatment (i.e., Sections 5.1–5.3) refers to security under
“passive” (eavesdropping) attacks. In contrast, in Section 5.4, we discuss notions of se-
curity under active attacks, culminating in robustness against chosen ciphertext attacks.
Additional issues are discussed in Section 5.5.

Teaching Tip. We suggest to focus on the basic definitional treatment (i.e., Sections 5.1
and 5.2.1–5.2.4) and on the the feasibility of satisfying these definitions (as demon-
started by the simplest constructions provided in Sections 5.3.3 and 5.3.4.1). The
overview to security under active attacks (i.e., Section 5.4.1) is also recommended.
We assume that the reader is familiar with the material in previous chapters (and
specifically with Sections 2.2, 2.4, 2.5, 3.2–3.4, and 3.6). This familiarity is important
not only because we use some of the notions and results presented in these sections but
also because we use similar proof techniques (and do so while assuming that this is not
the reader’s first encounter with these techniques).

373
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

5.1. The Basic Setting

Loosely speaking, encryption schemes are supposed to enable private exchange of

information between parties that communicate over an insecure channel. Thus, the basic
setting consists of a sender, a receiver, and an insecure channel that may be tapped by
an adversary. The goal is to allow the sender to transfer information to the receiver,
over the insecure channel, without letting the adversary figure out this information.
Thus, we distinguish between the actual (secret) information that the receiver wishes to
transmit and the message(s) sent over the insecure communication channel. The former
is called the plaintext, whereas the latter is called the ciphertext. Clearly, the ciphertext
must differ from the plaintext or else the adversary can easily obtain the plaintext by
tapping the channel. Thus, the sender must transform the plaintext into a corresponding
ciphertext such that the receiver can retrieve the plaintext from the ciphertext, but the
adversary cannot do so. Clearly, something must distinguish the receiver (who is able
to retrieve the plaintext from the corresponding ciphertext) from the adversary (who
cannot do so). Specifically, the receiver knows something that the adversary does not
know. This thing is called a key.
An encryption scheme consists of a method of transforming plaintexts into cipher-
texts and vice versa, using adequate keys. These keys are essential to the ability to effect
these transformations. Formally, these transformations are performed by corresponding
algorithms: an encryption algorithm that transforms a given plaintext and an adequate
(encryption) key into a corresponding ciphertext, and a decryption algorithm that given
the ciphertext and an adequate (decryption) key recovers the original plaintext. Actu-
ally, we need to consider a third algorithm, namely, a probabilistic algorithm used to
generate keys (i.e., a key-generation algorithm). This algorithm must be probabilistic
(or else, by invoking it, the adversary obtains the very same key used by the receiver).
We stress that the encryption scheme itself (i.e., the aforementioned three algorithms)
may be known to the adversary, and the scheme’s security relies on the hypothesis that
the adversary does not know the actual keys in use.1
In accordance with these principles, an encryption scheme consists of three
algorithms. These algorithms are public (i.e., known to all parties). The two obvious
algorithms are the encryption algorithm, which transforms plaintexts into ciphertexts,
and the decryption algorithm, which transforms ciphertexts into plaintexts. By these
principles, it is clear that the decryption algorithm must employ a key that is known
to the receiver but is not known to the adversary. This key is generated using a third
algorithm, called the key-generator. Furthermore, it is not hard to see that the encryp-
tion process must also depend on the key (or else messages sent to one party can be
read by a different party who is also a potential receiver). Thus, the key-generation
algorithm is used to produce a pair of (related) keys, one for encryption and one for de-
cryption. The encryption algorithm, given an encryption-key and a plaintext, produces
a ciphertext that when fed to the decryption algorithm, together with the corresponding

1 In fact, in many cases, the legitimate interest may be served best by publicizing the scheme itself, because this
allows an (independent) expert evaluation of the security of the scheme to be obtained.

374

www.Ebook777.com
 5.1 THE BASIC SETTING

plaintext plaintext
ciphertext
X
E D X

K K

Sender’s protected region Receiver’s protected region

ADVERSARY

The key K is known to both receiver and sender, but is unknown to

the adversary. For example, the receiver may generate K at random
and pass it to the sender via a perfectly-private secondary channel (not
shown here).
Figure 5.1: Private-key encryption schemes: an illustration.

decryption-key, yields the original plaintext. We stress that knowledge of the decryption-
key is essential for the latter transformation.

5.1.1. Private-Key Versus Public-Key Schemes

A fundamental distinction between encryption schemes refers to the relation between
the aforementioned pair of keys (i.e., the encryption-key and the decryption-key). The
simpler (and older) notion assumes that the encryption-key equals the decryption-key.
Such schemes are called private-key (or symmetric).

Private-Key Encryption Schemes. To use a private-key scheme, the legitimate parties

must first agree on the secret key. This can be done by having one party generate the
key at random and send it to the other party using a (secondary) channel that (unlike
the main channel) is assumed to be secure (i.e., it cannot be tapped by the adversary). A
crucial point is that the key is generated independently of the plaintext, and so it can be
generated and exchanged prior to the plaintext even being determined. Assuming that
the legitimate parties have agreed on a (secret) key, they can secretly communicate
by using this key (see illustration in Figure 5.1): The sender encrypts the desired
plaintext using this key, and the receiver recovers the plaintext from the corresponding
ciphertext (by using the same key). Thus, private-key encryption is a way of extending
a private channel over time: If the parties can use a private channel today (e.g., they
are currently in the same physical location) but not tomorrow, then they can use the
private channel today to exchange a secret key that they may use tomorrow for secret
communication.
A simple example of a private-key encryption scheme is the one-time pad. The
secret key is merely a uniformly chosen sequence of n bits, and an n-bit long ci-
phertext is produced by XORing the plaintext, bit-by-bit, with the key. The plaintext
is recovered from the ciphertext in the same way. Clearly, the one-time pad provides

375
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

plaintext plaintext
ciphertext
X
E D X

e
e e d
Sender’s protected region Receiver’s protected region
ADVERSARY

The key-pair (e, d) is generated by the receiver, who posts the

encryption-key e on a public media, while keeping the decryption-key
d secret.
Figure 5.2: Public-key encryption schemes: an illustration.

absolute security. However, its usage of the key is inefficient; or, put in other words,
it requires keys of length comparable to the total length (or information contents) of
the data being communicated. By contrast, the rest of this chapter will focus on en-
cryption schemes in which n-bit long keys allow for the secure communication of
data having an a priori unbounded (albeit polynomial in n) length. In particular, n-bit
long keys allow for significantly more than n bits of information to be communicated
securely.

Public-Key Encryption Schemes. A new type of encryption schemes emerged in

the 1970s. In these so-called public-key (or asymmetric) encryption schemes, the
decryption-key differs from the encryption-key. Furthermore, it is infeasible to find the
decryption-key, given the encryption-key. These schemes enable secure communication
without the use of a secure channel. Instead, each party applies the key-generation
algorithm to produce a pair of keys. The party (denoted P) keeps the decryption-key,
denoted d P , secret and publishes the encryption-key, denoted e P . Now, any party can
send P private messages by encrypting them using the encryption-key e P . Party P can
decrypt these messages by using the decryption-key d P , but nobody else can do so.
(See illustration in Figure 5.2.)

5.1.2. The Syntax of Encryption Schemes

We start by defining the basic mechanism of encryption schemes. This definition says
nothing about the security of the scheme (which is the subject of the next section).

Definition 5.1.1 (encryption scheme): An encryption scheme is a triple, (G, E, D),

of probabilistic polynomial-time algorithms satisfying the following two conditions:

1. On input 1n , algorithm G (called the key-generator) outputs a pair of bit strings.

2. For every pair (e, d) in the range of G(1n ), and for every α ∈ {0, 1}∗ , algorithms E
376

www.Ebook777.com
 5.1 THE BASIC SETTING

(encryption) and D (decryption) satisfy

Pr[D(d, E(e, α)) = α] = 1

where the probability is taken over the internal coin tosses of algorithms E and D.

The integer n serves as the security parameter of the scheme. Each (e, d) in the range
of G(1n ) constitutes a pair of corresponding encryption/decryption keys. The string
E(e, α) is the encryption of the plaintext α ∈ {0, 1}∗ using the encryption-key e, whereas
D(d, β) is the decryption of the ciphertext β using the decryption-key d.

We stress that Definition 5.1.1 says nothing about security, and so trivial (insecure)
def def
algorithms may satisfy it (e.g., E(e, α) = α and D(d, β) = β). Furthermore, Defini-
tion 5.1.1 does not distinguish private-key encryption schemes from public-key ones.
The difference between the two types is introduced in the security definitions: In a
public-key scheme the “breaking algorithm” gets the encryption-key (i.e., e) as an ad-
ditional input (and thus e = d follows), while in private-key schemes e is not given to
the “breaking algorithm” (and thus, one may assume, without loss of generality, that
e = d).
We stress that this definition requires the scheme to operate for every plaintext,
and specifically for plaintext of length exceeding the length of the encryption-key.
(This rules out the information theoretic secure “one-time pad” scheme mentioned
earlier.)

Notation. In the rest of this text, we write E e (α) instead of E(e, α) and Dd (β) instead
of D(d, β). Sometimes, when there is little risk of confusion, we drop these subscripts.
Also, we let G 1 (1n ) (resp., G 2 (1n )) denote the first (resp., second) element in the
pair G(1n ). That is, G(1n ) = (G 1 (1n ), G 2 (1n )). Without loss of generality, we may
assume that |G 1 (1n )| and |G 2 (1n )| are polynomially related to n, and that each of these
integers can be efficiently computed from the other. (In fact, we may even assume that
|G 1 (1n )| = |G 2 (1n )| = n; see Exercise 6.)

Comments. Definition 5.1.1 may be relaxed in several ways without significantly harm-
ing its usefulness. For example, we may relax Condition (2) and allow a negligible de-
cryption error (e.g., Pr[Dd (E e (α)) = α] < 2−n ). Alternatively, one may postulate that
Condition (2) holds for all but a negligible measure of the key-pairs generated by G(1n ).
At least one of these relaxations is essential for some suggestions of (public-key) en-
cryption schemes.
Another relaxation consists of restricting the domain of possible plaintexts (and
ciphertexts). For example, one may restrict Condition (2) to α’s of length (n), where
 : N → N is some fixed function. Given a scheme of the latter type (with plaintext
length ), we may construct a scheme as in Definition 5.1.1 by breaking plaintexts into
blocks of length (n) and applying the restricted scheme separately to each block. (Note
that security of the resulting scheme requires that the security of the length-restricted
scheme be preserved under multiple encryptions with the same key.) For more details
see Sections 5.2.4 and 5.3.2.
377
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

5.2. Definitions of Security

In this section we present two fundamental definitions of security and prove their equiv-
alence. The first definition, called semantic security, is the most natural one. Semantic
security is a computational-complexity analogue of Shannon’s definition of perfect pri-
vacy (which requires that the ciphertext yield no information regarding the plaintext).
Loosely speaking, an encryption scheme is semantically secure if it is infeasible to
learn anything about the plaintext from the ciphertext (i.e., impossibility is replaced
by infeasibility). The second definition has a more technical flavor. It interprets se-
curity as the infeasibility of distinguishing between encryptions of a given pair of
messages. This definition is useful in demonstrating the security of a proposed encryp-
tion scheme and for the analysis of cryptographic protocols that utilize an encryption
scheme.
We stress that the definitions presented in Section 5.2.1 go far beyond saying that it
is infeasible to recover the plaintext from the ciphertext. The latter statement is indeed a
minimal requirement for a secure encryption scheme, but we claim that it is far too weak
a requirement. For example, one should certainly not use an encryption scheme that
leaks the first part of the plaintext (even if it is infeasible to recover the entire plaintext
from the ciphertext). In general, an encryption scheme is typically used in applications
where even obtaining partial information on the plaintext may endanger the security
of the application. The question of which partial information endangers the security
of a specific application is typically hard (if not impossible) to answer. Furthermore,
we wish to design application-independent encryption schemes, and when doing so
it is the case that each piece of partial information may endanger some application.
Thus, we require that it be infeasible to obtain any information about the plaintext
from the ciphertext. Moreover, in most applications the plaintext may not be uniformly
distributed, and some a priori information regarding it may be available to the adversary.
We thus require that the secrecy of all partial information be preserved also in such a
case. That is, given any a priori information on the plaintext, it is infeasible to obtain
any (new) information about the plaintext from the ciphertext (beyond what is feasible
to obtain from the a priori information on the plaintext). The definition of semantic
security postulates all of this.

Security of Multiple Plaintexts. Continuing the preceding discussion, the definitions

are presented first in terms of the security of a single encrypted plaintext. However,
in many cases, it is desirable to encrypt many plaintexts using the same encryption-
key, and security needs to be preserved in these cases, too. Adequate definitions and
discussions are deferred to Section 5.2.4.

A Technical Comment: Non-Uniform Complexity Formulation. To simplify the ex-

position, we define security in terms of non-uniform complexity (see Section 1.3.3 of
Volume 1). Namely, in the security definitions we expand the domain of efficient adver-
saries (and algorithms) to include (explicitly or implicitly) non-uniform polynomial-size
circuits, rather than only probabilistic polynomial-time machines. Likewise, we make
378

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

no computational restriction regarding the probability distribution from which messages

are taken, nor regarding the a priori information available on these messages. We note
that employing such a non-uniform complexity formulation (rather than a uniform one)
may only strengthen the definitions, yet it does weaken the implications proven between
the definitions because these (simpler) proofs make free usage of non-uniformity. A
uniform-complexity treatment is provided in Section 5.2.5.

5.2.1. Semantic Security

A good disguise should not reveal the person’s height.
Shafi Goldwasser and Silvio Micali, 1982

Loosely speaking, semantic security means that nothing can be gained by looking
at a ciphertext. Following the simulation paradigm, this means that whatever can be
efficiently learned from the ciphertext can also be efficiently learned from scratch (or
from nothing).

5.2.1.1. The Actual Definitions

To be somewhat more accurate, semantic security means that whatever can be efficiently
computed from the ciphertext can be efficiently computed when given only the length
of the plaintext. Note that this formulation does not rule out the possibility that the
length of the plaintext can be inferred from the ciphertext. Indeed, some information
about the length of the plaintext must be revealed by the ciphertext (see Exercise 4).
We stress that other than information about the length of the plaintext, the ciphertext is
required to yield nothing about the plaintext.
In the actual definitions, we consider only information regarding the plaintext (rather
than information regarding the ciphertext and/or the encryption-key) that can be ob-
tained from the ciphertext. Furthermore, we restrict our attention to functions (rather
than randomized processes) applied to the plaintext. We do so because of the intuitive
appeal of this special case, and are comfortable doing so because this special case im-
plies the general one (see Exercise 13). We augment this formulation by requiring that
the infeasibility of obtaining information about the plaintext remain valid even in the
presence of other auxiliary partial information about the same plaintext. Namely, what-
ever can be efficiently computed from the ciphertext and additional partial information
about the plaintext can be efficiently computed given only the length of the plaintext and
the same partial information. In the definition that follows, the information regarding the
plaintext that the adversary tries to obtain is represented by the function f, whereas the
a priori partial information about the plaintext is represented by the function h. The in-
feasibility of obtaining information about the plaintext is required to hold for any
distribution of plaintexts, represented by the probability ensemble {X n }n∈N .
Security holds only for plaintexts of length polynomial in the security parameter. This
is captured in the following definitions by the restriction |X n | ≤ poly(n), where “poly”
represents an arbitrary (unspecified) polynomial. Note that we cannot hope to provide
computational security for plaintexts of unbounded length or for plaintexts of length
379
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

that is exponential in the security parameter (see Exercise 3). Likewise, we restrict the
functions f and h to be polynomially-bounded, that is, | f (z)|, |h(z)| ≤ poly(|z|).
The difference between private-key and public-key encryption schemes is manifested
in the definition of security. In the latter case, the adversary (which is trying to obtain
information on the plaintext) is given the encryption-key, whereas in the former case
it is not. Thus, the difference between these schemes amounts to a difference in the
adversary model (considered in the definition of security). We start by presenting the
definition for private-key encryption schemes.

Definition 5.2.1 (semantic security – private-key): An encryption scheme, (G, E, D),

is semantically secure (in the private-key model) if for every probabilistic polynomial-
time algorithm A there exists a probabilistic polynomial-time algorithm A such that
for every probability ensemble {X n }n∈N , with |X n | ≤ poly(n), every pair of polynomi-
ally bounded functions f, h : {0, 1}∗ → {0, 1}∗ , every positive polynomial p and all
sufficiently large n


Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )

 1
< Pr A (1n , 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

(The probability in these terms is taken over X n as well as over the internal coin tosses
of either algorithms G, E, and A or algorithm A .)

We stress that all the occurrences of X n in each of the probabilistic expressions re-
fer to the same random variable (see the general convention stated in Section 1.2.1
in Volume 1). The security parameter 1n is given to both algorithms (as well as to the
functions h and f ) for technical reasons.2 The function h provides both algorithms with
partial information regarding the plaintext X n . Furthermore, h also makes the defini-
tion implicitly non-uniform; see further discussion in Section 5.2.1.2. In addition, both
algorithms get the length of X n . These algorithms then try to guess the value f (1n , X n );
namely, they try to infer information about the plaintext X n . Loosely speaking, in a se-
mantically secure encryption scheme the ciphertext does not help in this inference task.
That is, the success probability of any efficient algorithm (i.e., algorithm A) that is given
the ciphertext can be matched, up to a negligible fraction, by the success probability of
an efficient algorithm (i.e., algorithm A ) that is not given the ciphertext at all.
Definition 5.2.1 refers to private-key encryption schemes. To derive a definition of
security for public-key encryption schemes, the encryption-key (i.e., G 1 (1n )) should
be given to the adversary as an additional input.

2 The auxiliary input 1n is used for several purposes. First, it allows smooth transition to fully non-uniform
formulations (e.g., Definition 5.2.3) in which the (polynomial-size) adversary depends on n. Thus, it is good to
provide A (and thus also A ) with 1n . Once this is done, it is natural to allow also h and f to depend on n. In
fact, allowing h and f to explicitly depend on n facilitates the proof of Proposition 5.2.7. In light of the fact
that 1n is given to both algorithms, we may replace the input part 1|X n | by |X n |, because the former may be
recovered from the latter in poly(n)-time.

380

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Definition 5.2.2 (semantic security – public-key): An encryption scheme, (G, E, D),

is semantically secure (in the public-key model) if for every probabilistic polynomial-
time algorithm A, there exists a probabilistic polynomial-time algorithm A such that
for every {X n }n∈N , f, h, p, and n as in Definition 5.2.1


Pr A(1n , G 1 (1n ), E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )

 1
< Pr A (1n , 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

Recall that (by our conventions) both occurrences of G 1 (1n ), in the first probabilistic
expression, refer to the same random variable. We comment that it is pointless to give
the random encryption-key (i.e., G 1 (1n )) to algorithm A (because the task as well as
the main inputs of A are unrelated to the encryption-key, and anyhow A could generate
a random encryption-key by itself).

Terminology. For sake of simplicity, we refer to an encryption scheme that is seman-

tically secure in the private-key (resp., public-key) model as a semantically secure
private-key (resp., public-key) encryption scheme.
The reader may note that a semantically secure public-key encryption scheme cannot
employ a deterministic encryption algorithm; that is, E e (x) must be a random variable
rather than a fixed string. This is more evident with respect to the equivalent Defini-
tion 5.2.4. See further discussion following Definition 5.2.4.

5.2.1.2. Further Discussion of Some Definitional Choices

We discuss several secondary issues regarding Definitions 5.2.1 and 5.2.2. The in-
terested reader is also referred to Exercises 16, 17, and 19, which present additional
variants of the definition of semantic security.

Implicit Non-Uniformity of the Definitions. The fact that h is not required to be

computable makes these definitions non-uniform. This is the case because both algo-
rithms are given h(1n , X n ) as auxiliary input, and the latter may account for arbitrary
(polynomially bounded) advice. For example, letting h(1n , ·) = an ∈ {0, 1}poly(n) means
that both algorithms are supplied with (non-uniform) advice (as in one of the com-
mon formulations of non-uniform polynomial-time; see Section 1.3.3). In general, the
function h can code both information regarding its main input and non-uniform ad-
vice depending on the security parameter (i.e., h(1n , x) = (h  (x), an )). We comment
that these definitions are equivalent to allowing A and A to be related families of non-
uniform circuits, where by related we mean that the circuits in the family A = {An }n∈N
can be efficiently computed from the corresponding circuits in the family A = {An }n∈N .
For further discussion, see Exercise 9.

Lack of Computational Restrictions Regarding the Function f. We do not even

require that the function f be computable. This seems strange at first glance because
(unlike the situation with respect to the function h, which codes a priori information
381
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

given to the algorithms) the algorithms are asked to guess the value of f (at a plaintext
implicit in the ciphertext given only to A). However, as we shall see in the sequel (see
also Exercise 13), the actual technical content of semantic security is that the proba-
bility ensembles {(1n , E(X n ), 1|X n | , h(1n , X n ))}n and {(1n , E(1|X n | ), 1|X n | , h(1n , X n ))}n
are computationally indistinguishable (and so whatever A can compute can also be
computed by A ). Note that the latter statement does not refer to the function f , which
explains why we need not make any restriction regarding f.

Other Modifications of No Impact. Actually, inclusion of a priori information re-

garding the plaintext (represented by the function h) does not affect the definition of
semantic security: Definition 5.2.1 remains intact if we restrict h to only depend on
the security parameter (and so only provide plaintext-oblivious non-uniform advice).
(This can be shown in various ways; e.g., see Exercise 14.1.) Also, the function f can
be restricted to be a Boolean function having polynomial-size circuits, and the random
variable X n may be restricted to be very “dull” (e.g., have only two strings in its sup-
port): See proof of Theorem 5.2.5. On the other hand, Definition 5.2.1 implies stronger
forms discussed in Exercises 13, 17 and 18.

5.2.2. Indistinguishability of Encryptions

A good disguise should not allow a mother to distinguish her own children.
Shafi Goldwasser and Silvio Micali, 1982

The following technical interpretation of security states that it is infeasible to distinguish

the encryptions of two plaintexts (of the same length). That is, such ciphertexts are
computationally indistinguishable as defined in Definition 3.2.7. Again, we start with
the private-key variant.

Definition 5.2.3 (indistinguishability of encryptions – private-key): An encryption

scheme, (G, E, D), has indistinguishable encryptions (in the private-key model) if
for every polynomial-size circuit family {Cn }, every positive polynomial p, all suffi-
ciently large n, and every x, y ∈ {0, 1}poly(n) (i.e., |x| = |y|),


 1
| Pr C n (E G 1 (1n ) (x)) = 1 − Pr Cn (E G 1 (1n ) (y)) = 1 | <
p(n)

The probability in these terms is taken over the internal coin tosses of algorithms G
and E.

Note that the potential plaintexts to be distinguished can be incorporated into the circuit
Cn . Thus, the circuit models both the adversary’s strategy and its a priori information:
See Exercise 11.
Again, the security definition for public-key encryption schemes is derived by adding
the encryption-key (i.e., G 1 (1n )) as an additional input to the potential distinguisher.
382

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Definition 5.2.4 (indistinguishability of encryptions – public-key): An encryption

scheme, (G, E, D), has indistinguishable encryptions (in the public-key model) if for
every polynomial-size circuit family {Cn }, and every p, n, x, and y as in Definition 5.2.3


 1
| Pr Cn (G 1 (1n ), E G 1 (1n ) (x)) = 1 − Pr Cn (G 1 (1n ), E G 1 (1n ) (y)) = 1 | <
p(n)

Terminology. We refer to an encryption scheme that has indistinguishable encryptions

in the private-key (resp., public-key) model as a ciphertext-indistinguishable private-
key (resp., public-key) encryption scheme.

Failure of Deterministic Encryption Algorithms. A ciphertext-indistinguishable

public-key encryption scheme cannot employ a deterministic encryption algorithm (i.e.,
E e (x) cannot be a fixed string). The reason is that for a public-key encryption scheme
with a deterministic encryption algorithm E, given an encryption-key e and a pair of
candidate plaintexts (x, y), one can easily distinguish E e (x) from E e (y) (by merely
applying E e to x and comparing the result to the given ciphertext). In contrast, in case
the encryption algorithm itself is randomized, the same plaintext can be encrypted
in many exponentially different ways, under the same encryption-key. Furthermore,
the probability that applying E e twice to the same message (while using independent
randomization in E e ) results in the same ciphertext may be exponentially vanishing.
(Indeed, as shown in Section 5.3.4, public-key encryption schemes having indistin-
guishable encryptions can be constructed based on any trapdoor permutation, and these
schemes employ randomized encryption algorithms.)

5.2.3. Equivalence of the Security Definitions

The following theorem is stated and proven for private-key encryption schemes. A
similar result holds for public-key encryption schemes (see Exercise 12).

Theorem 5.2.5 (equivalence of definitions – private-key): A private-key encryption

scheme is semantically secure if and only if it has indistinguishable encryptions.

Let (G, E, D) be an encryption scheme. We formulate a proposition for each of the two
directions of this theorem. Each proposition is in fact stronger than the corresponding
direction stated in Theorem 5.2.5. The more useful direction is stated first: It asserts
that the technical interpretation of security, in terms of ciphertext-indistinguishability,
implies the natural notion of semantic security. Thus, the following proposition yields
a methodology for designing semantically secure encryption schemes: Design and
prove your scheme to be ciphertext-indistinguishable, and conclude (by applying the
proposition) that it is semantically secure. The opposite direction (of Theorem 5.2.5)
establishes the “completeness” of the latter methodology, and more generally asserts
that requiring an encryption scheme to be ciphertext-indistinguishable does not rule
out schemes that are semantically secure.
383
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Proposition 5.2.6 (useful direction: “indistinguishability” implies “security”): Sup-

pose that (G, E, D) is a ciphertext-indistinguishable private-key encryption scheme.
Then (G, E, D) is semantically secure. Furthermore, Definition 5.2.1 is satisfied by
using A = M A , where M is a fixed oracle machine; that is, there exists a single M
such that for every A letting A = M A will do.

Proposition 5.2.7 (opposite direction: “security” implies “indistinguishability”): Sup-

pose that (G, E, D) is a semantically secure private-key encryption scheme. Then
(G, E, D) has indistinguishable encryptions. Furthermore, the conclusion holds even
if the definition of semantic security is restricted to the special case satisfying the
following four conditions:
1. The random variable X n is uniformly distributed over a set containing two strings;
2. The value of h depends only on the length of its input or alternatively h(1n , x) =
h  (n), for some h  ;
3. The function f is Boolean and is computable by a family of (possibly non-uniform)
polynomial-size circuits;
4. The algorithm A is deterministic.
In addition, no computational restrictions are placed on algorithm A (i.e., A can be
any function), and moreover A may depend on {X n }n∈N , h, f , and A.

Observe that the four itemized conditions limit the scope of the four universal quantifiers
in Definition 5.2.1, whereas the last sentence removes a restriction on the existential
quantifier (i.e., removes the complexity bound on A ) and reverses the order of quanti-
fiers allowing the existential quantifier to depend on all universal quantifiers (rather than
only on the last one). Thus, each of these modifications makes the resulting definition
potentially weaker. Still, combining Propositions 5.2.7 and 5.2.6, it follows that a weak
version of Definition 5.2.1 implies (an even stronger version than) the one stated in
Definition 5.2.1.

5.2.3.1. Proof of Proposition 5.2.6

Suppose that (G, E, D) has indistinguishable encryptions. We will show that (G, E, D)
is semantically secure by constructing, for every probabilistic polynomial-time algo-
rithm A, a probabilistic polynomial-time algorithm A such that the condition in Defi-
nition 5.2.1 holds. That is, for every {X n }n∈N , f and h, algorithm A guesses f (1n , X n )
from (1n , 1|X n | , h(1n , X n )) essentially as well as A guesses f (1n , X n ) from E(X n ) and
(1n , 1|X n | , h(1n , X n )). Our construction of A consists of merely invoking A on input
(1n , E(1|X n | ), 1|X n | , h(1n , X n )), and returning whatever A does. That is, A invokes A
with a dummy encryption rather than with an encryption of X n (which A expects to
get, but A does not have). Intuitively, the indistinguishability of encryptions implies
that A behaves nearly as well when invoked by A (and given a dummy encryption) as
when given the encryption of X n , and this establishes that A is adequate with respect
to A. The main issue in materializing this plan is to show that the specific formulation
of indistinguishability of encryptions indeed supports the implication (i.e., implies that
384

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

A guesses f (1n , X n ) essentially as well when given a dummy encryption as when given
the encryption of X n ). Details follow.

The construction of A : Let A be an algorithm that tries to infer partial information (i.e.,
the value f (1n , X n )) from the encryption of the plaintext X n (when also given 1n , 1|X n |
and a priori information h(1n , X n )). Intuitively, on input E(α) and (1n , 1|α| , h(1n , α)),
algorithm A tries to guess f (1n , α). We construct a new algorithm, A , that performs
essentially as well without getting the input E(α). The new algorithm consists of invok-
ing A on input E G 1 (1n ) (1|α| ) and (1n , 1|α| , h(1n , α)), and outputting whatever A does.
That is, on input (1n , 1|α| , h(1n , α)), algorithm A proceeds as follows:
1. A invokes the key-generator G (on input 1n ), and obtains an encryption-key e ←
G 1 (1n ).
2. A invokes the encryption algorithm with key e and (“dummy”) plaintext 1|α| , ob-
taining a ciphertext β ← E e (1|α| ).
3. A invokes A on input (1n , β, 1|α| , h(1n , α)), and outputs whatever A does.

Observe that A is described in terms of an oracle machine that makes a single oracle
call to (any given) A, in addition to invoking the fixed algorithms G and E. Furthermore,
the construction of A depends neither on the functions h and f nor on the distribution
of plaintexts to be encrypted (represented by the probability ensembles {X n }n∈N ). Thus,
A is probabilistic polynomial-time whenever A is probabilistic polynomial-time (and
regardless of the complexity of h, f , and {X n }n∈N ).
Indistinguishability of encryptions will be used to prove that A performs essentially
as well as A. Specifically, the proof will use a reducibility argument.

Claim 5.2.6.1: Let A be as in the preceding construction. Then, for every {X n }n∈N , f ,
h, and p as in Definition 5.2.1, and all sufficiently large n’s


Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )

 1
< Pr A (1n , 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

def
Proof: To simplify the notations, let us incorporate 1|α| into h n (α) = h(1n , α) and let
def
f n (α) = f (1n , α). Also, we omit 1n from the inputs given to A, shorthanding A(1n , c, v)
by A(c, v). Using the definition of A , we rewrite the claim as asserting


Pr A(E G 1 (1n ) (X n ), h n (X n )) = f n (X n ) (5.1)

 1
< Pr A(E G 1 (1n ) (1|X n | ), h n (X n )) = f n (X n ) +
p(n)
Intuitively, Eq. (5.1) follows from the indistinguishability of encryptions. Otherwise,
by fixing a violating value of X n and hardwiring the corresponding values of h n (X n )
and f n (X n ), we get a small circuit that distinguishes an encryption of this value of X n
from an encryption of 1|X n | . Details follow.
385
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Assume toward the contradiction that for some polynomial p and infinitely many
n’s Eq. (5.1) is violated. Then, for each such n, we have E[n (X n )] > 1/ p(n), where
def 


n (x) =  Pr A(E G 1 (1n ) (x), h n (x)) = f n (x) − Pr A(E G 1 (1n ) (1|x| ), h n (x)) = f n (x) 
We use an averaging argument to single out a string xn in the support of X n such that
n (xn ) ≥ E[n (X n )]: That is, let x n ∈ {0, 1}poly(n) be a string for which the value of
n (·) is maximum, and so n (xn ) > 1/ p(n). Using this x n , we introduce a circuit Cn ,
which incorporates the fixed values f n (xn ) and h n (xn ), and distinguishes the encryption
of xn from the encryption of 1|xn | . The circuit Cn operates as follows. On input β = E(α),
the circuit C n invokes A(β, h n (xn )) and outputs 1 if and only if A outputs the value
f n (xn ). Otherwise, Cn outputs 0.
This circuit is indeed of polynomial size because it merely incorporates strings of
polynomial length (i.e., f n (xn ) and h n (xn )) and emulates a polynomial-time computation
(i.e., that of A). (Note that the circuit family {Cn } is indeed non-uniform since its
definition is based on a non-uniform selection of xn ’s as well as on a hardwiring of
(possibly uncomputable) corresponding strings h n (xn ) and f n (x n ).) Clearly,



Pr C n (E G 1 (1n ) (α)) = 1 = Pr A(E G 1 (1n ) (α), h n (xn )) = f n (x n ) (5.2)
Combining Eq. (5.2) with the definition of n (xn ), we get



 Pr Cn (E G 1 (1n ) (xn )) = 1 − Pr Cn (E G 1 (1n ) (1|xn | )) = 1  = n (x n )
1
>
p(n)
This contradicts our hypothesis that E has indistinguishable encryptions, and the claim
follows.
We have just shown that A performs essentially as well as A, and so Proposition 5.2.6
follows.

Comments. The fact that we deal with a non-uniform model of computation allows
the preceding proof to proceed regardless of the complexity of f and h. All that
our definition of C n requires is the hardwiring of the values of f and h on a single
string, and this can be done regardless of the complexity of f and h (provided that
| f n (xn )|, |h n (x n )| ≤ poly(n)).
When proving the public-key analogue of Proposition 5.2.6, algorithm A is defined
exactly as in the present proof, but its analysis is slightly different: The distinguishing
circuit, considered in the analysis of the performance of A , obtains the encryption-key
as part of its input and passes it to algorithm A (upon invoking the latter).

5.2.3.2. Proof of Proposition 5.2.7

Intuitively, indistinguishability of encryption (i.e., of the encryptions of xn and yn ) is
a special case of semantic security in which f indicates one of the plaintexts and h
does not distinguish them (i.e., f (1n , z) = 1 iff z = x n and h(1n , xn ) = h(1n , yn )). The
only issue to be addressed by the actual proof is that semantic security refers to uniform
386

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

(probabilistic polynomial-time) adversaries, whereas indistinguishability of encryption

refers to non-uniform polynomial-size circuits. This gap is bridged by using the func-
tion h to provide the algorithms in the semantic-security formulation with adequate
non-uniform advice (which may be used by the machine in the indistinguishability of
encryption formulation).
The actual proof is by a reducibility argument. We show that if (G, E, D) has dis-
tinguishable encryptions, then it is not semantically secure (not even in the restricted
sense mentioned in the furthermore-clause of the proposition). Toward this end, we
assume that there exists a (positive) polynomial p and a polynomial-size circuit family
{Cn }, such that for infinitely many n’s there exists x n , yn ∈ {0, 1}poly(n) so that


 1
 Pr Cn (E G n) (x n )) = 1 − Pr Cn (E G 1 (1n ) (yn )) = 1  > (5.3)
1 (1
p(n)
Using these sequences of Cn ’s, xn ’s and yn ’s, we define {X n }n∈N , f and h (referred to
in Definition 5.2.1) as follows:
r The probability ensemble {X n }n∈N is defined such that X n is uniformly distributed
over {xn , yn }.
r The (Boolean) function f is defined such that f (1n , xn ) = 1 and f (1n , yn ) = 0, for
every n. Note that f (1n , X n ) = 1 with probability 1/2 and equals 0 otherwise.
r The function h is defined such that h(1n , X n ) equals the description of the circuit
Cn . Note that h(1n , X n ) = Cn with probability 1, and thus h(1n , X n ) reveals no
information on the value of X n .
Note that X n , f , and h satisfy the restrictions stated in the furthermore-clause of the
proposition. Intuitively, Eq. (5.3) implies violation of semantic security with respect to
the X n , h, and f . Indeed, we will present a (deterministic) polynomial-time algorithm
A that, given Cn = h(1n , X n ), guesses the value of f (1n , X n ) from the encryption
of X n , and does so with probability non-negligibly greater than 1/2. This violates
(even the restricted form of) semantic security, because no algorithm, regardless of its
complexity, can guess f (1n , X n ) with probability greater than 1/2 when only given
1|X n | (because given the constant values 1|X n | and h(1n , X n ), the value of f (1n , X n ) is
uniformly distributed over {0, 1}). Details follow.
Let us assume, without loss of generality, that for infinitely many n’s


 1
Pr C n (E G 1 (1n ) (xn )) = 1 > Pr Cn (E G 1 (1n ) (yn )) = 1 + (5.4)
p(n)

Claim 5.2.7.1: There exists a (deterministic) polynomial-time algorithm A such that

for infinitely many n’s

 1 1
Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n ) > +
2 2 p(n)

Proof: The desired algorithm A merely uses C n = h(1n , X n ) to distinguish E(x n ) from
E(yn ), and thus given E(X n ) it produces a guess for the value of f (1n , X n ). Specifically,
on input β = E(α) (where α is in the support of X n ) and (1n , 1|α| , h(1n , α)), algorithm A
387
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

recovers Cn = h(1n , α), invokes C n on input β, and outputs 1 if Cn outputs 1 (otherwise,

A outputs 0).3
def
It is left to analyze the success probability of A. Letting m = |x n | = |yn |, h n (α) =
def
h(1n , α) and f n (α) = f (1n , α), we have


Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h n (X n )) = f n (X n )
1

= · Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h n (X n )) = f n (X n ) | X n = x n
2
1

+ · Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h n (X n )) = f n (X n ) | X n = yn
2
1

= · Pr A(1n , E G 1 (1n ) (xn ), 1|xn | , C n ) = 1
2
1

+ · Pr A(1n , E G 1 (1n ) (yn ), 1|yn | , C n ) = 0
2
1 


= · Pr C n (E G 1 (1n ) (xn )) = 1 + 1 − Pr Cn (E G 1 (1n ) (yn )) = 1
2
1 1
> +
2 2 p(n)
where the inequality is due to Eq. (5.4).

In contrast, as aforementioned, no algorithm (regardless of its complexity) can guess

f (1n , X n ) with success probability above 1/2, when given only 1|X n | and h(1n , X n ). That
is, we have the following:
Fact 5.2.7.2: For every n and every algorithm A

 1
Pr A (1n , 1|X n | , h(1n , X n )) = f (1n , X n ) ≤ (5.5)
2

Proof: Just observe that the output of A , on its constant input values 1n , 1|X n | and
h(1n , X n ), is stochastically independent of the random variable f (1n , X n ), which in
turn is uniformly distributed in {0, 1}. Eq. (5.5) follows (and equality holds in case A
always outputs a value in {0, 1}).
Combining Claim 5.2.7.1 and Fact 5.2.7.2, we reach a contradiction to the hypothesis
that the scheme is semantically secure (even in the restricted sense mentioned in the
furthermore-clause of the proposition). Thus, the proposition follows.

Comment. When proving the public-key analogue of Proposition 5.2.7, algorithm A

is defined as in the current proof except that it passes the encryption-key, given to it as
part of its input, to the circuit Cn . The rest of the proof remains intact.

3 We comment that the value “1” output by Cn is an indication that α is more likely to be xn , whereas the
def
output of A is a guess of f (α). This point may be better stressed by redefining f such that f (1n , x n ) = x n and
def
f (1 , x) = yn if x = x n , and having A output x n if C n outputs 1 and output yn otherwise.
n

388

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

5.2.4. Multiple Messages

Definitions 5.2.1–5.2.4 only refer to the security of an encryption scheme that is used
to encrypt a single plaintext (per generated key). Since the plaintext may be longer than
the key, these definitions are already non-trivial, and an encryption scheme satisfying
them (even in the private-key model) implies the existence of one-way functions (see
Exercise 2). Still, in many cases, it is desirable to encrypt many plaintexts using the
same encryption-key. Loosely speaking, an encryption scheme is secure in the multiple-
message setting if analogous definitions (to Definitions 5.2.1–5.2.4) also hold when
polynomially many plaintexts are encrypted using the same encryption-key.
We show that in the public-key model, security in the single-message setting
(discussed earlier) implies security in the multiple-message setting (defined in
Section 5.2.4.1). We stress that this is not necessarily true for the private-key model.

5.2.4.1. Definitions
For a sequence of strings x = (x (1) , ..., x (t) ), we let E e (x) denote the sequence of the
t results that are obtained by applying the randomized process E e to the t strings
x (1) , ..., x (t) , respectively. That is, E e (x) = (E e (x (1) ), ..., E e (x (t) )). We stress that in
each of these t invocations, the randomized process E e utilizes independently chosen
random coins. For the sake of simplicity, we consider the encryption of (polynomi-
ally) many plaintexts of the same (polynomial) length (rather than the encryption of
plaintexts of various lengths as discussed in Exercise 20). The number of plaintexts
as well as their total length (in unary) are given to all algorithms either implicitly or
explicitly.4

Definition 5.2.8 (semantic security – multiple messages):

For private-key: An encryption scheme, (G, E, D), is semantically secure for mul-
tiple messages in the private-key model if for every probabilistic polynomial-
time algorithm A, there exists a probabilistic polynomial-time algorithm A such
(1) (t(n)) (1)
that for every probability ensemble {X n = (X n , ..., X n )}n∈N , with |X n | = · · · =
(t(n))
|X n | ≤ poly(n) and t(n) ≤ poly(n), every pair of polynomially bounded functions
f, h : {0, 1}∗ → {0, 1}∗ , every positive polynomial p and all sufficiently large n
 
Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A (1n , t(n), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

For public-key: An encryption scheme, (G, E, D), is semantically secure for multiple
messages in the public-key model if for A, A , t, {X n }n∈N , f, h, p, and n, as in the

4 For example, A can infer the number of plaintexts from the number of ciphertexts, whereas A is given this
number explicitly. Given the number of the plaintexts as well as their total length, both algorithms can infer the
length of each plaintext.

389
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

private-key case, it holds that

 
Pr A(1n , G 1 (1n ), E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A (1n , t(n), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

(The probability in these terms is taken over X n as well as over the internal coin tosses
of the relevant algorithms.)

We stress that the elements of X n are not necessarily independent; they may depend on
one another. Note that this definition also covers the case where the adversary obtains
some of the plaintexts themselves. In this case it is still infeasible for him/her to obtain
information about the missing plaintexts (see Exercise 22).

Definition 5.2.9 (indistinguishability of encryptions – multiple messages):

For private-key: An encryption scheme, (G, E, D), has indistinguishable encryptions
for multiple messages in the private-key model if for every polynomial-size cir-
cuit family {Cn }, every positive polynomial p, all sufficiently large n, and every
x1 , ..., x t(n) , y1 , ..., yt(n) ∈ {0, 1}poly(n) , with t(n) ≤ poly(n), it holds that


 1
| Pr Cn (E G 1 (1n ) (x̄)) = 1 − Pr C n (E G 1 (1n ) ( ȳ)) = 1 | <
p(n)
where x̄ = (x1 , ..., x t(n) ) and ȳ = (y1 , ..., yt(n) ).
For public-key: An encryption scheme, (G, E, D), has indistinguishable encryp-
tions for multiple messages in the public-key model if for t, {Cn }, p, n, and
x1 , ..., x t(n) , y1 , ..., yt(n) as in the private-key case


 1
| Pr Cn (G 1 (1n ), E G 1 (1n ) (x̄)) = 1 − Pr C n (G 1 (1n ), E G 1 (1n ) ( ȳ)) = 1 | <
p(n)

The equivalence of Definitions 5.2.8 and 5.2.9 can be established analogously to the
proof of Theorem 5.2.5.

Theorem 5.2.10 (equivalence of definitions – multiple messages): A private-key (resp.,

public-key) encryption scheme is semantically secure for multiple messages if and only
if it has indistinguishable encryptions for multiple messages.

Thus, proving that single-message security implies multiple-message security for one
definition of security yields the same for the other. We may thus concentrate on the
ciphertext-indistinguishability definitions.

5.2.4.2. The Effect on the Public-Key Model

We first consider public-key encryption schemes.
390

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Theorem 5.2.11 (single-message security implies multiple-message security): A

public-key encryption scheme has indistinguishable encryptions for multiple messages
(i.e., satisfies Definition 5.2.9 in the public-key model) if and only if it has indistinguish-
able encryptions for a single message (i.e., satisfies Definition 5.2.4).

Proof: Clearly, multiple-message security implies single-message security as a special

case. The other direction follows by adapting the proof of Theorem 3.2.6 to the current
setting.
Suppose, toward the contradiction, that there exist a polynomial t, a polynomial-size
circuit family {Cn }, and a polynomial p, such that for infinitely many n’s, there exists
x1 , ..., x t(n) , y1 , ..., yt(n) ∈ {0, 1}poly(n) so that


 1
 Pr Cn (G 1 (1n ), E G 1 (1n ) (x̄)) = 1 − Pr Cn (G 1 (1n ), E G 1 (1n ) ( ȳ)) = 1  >
p(n)
where x̄ = (x1 , ..., xt(n) ) and ȳ = (y1 , ..., yt(n) ). Let us consider such a generic n and
the corresponding sequences x1 , ..., x t(n) and y1 , ..., yt(n) . We use a hybrid argument.
Specifically, define
def
h̄ (i) = (x 1 , ..., x i , yi+1 , ..., yt(n) )
def
and Hn(i) = (G 1 (1n ), E G 1 (1n ) (h̄ (i) ))
(0) (t(n))
Since Hn = (G 1 (1n ), E G 1 (1n ) ( ȳ)) and Hn = (G 1 (1n ), E G 1 (1n ) (x̄)), it follows that
there exists an i ∈ {0, ..., t(n) − 1} so that


 1
 Pr C n (H (i) ) = 1 − Pr C n (H (i+1) ) = 1  > (5.6)
n n
t(n) · p(n)
We show that Eq. (5.6) yields a polynomial-size circuit that distinguishes the encryption
of xi+1 from the encryption of yi+1 , and thus derive a contradiction to security in the
single-message setting. Specifically, we construct a circuit Dn that incorporates the
circuit Cn as well as the index i and the strings x 1 , ..., xi+1 , yi+1 , ..., yt(n) . On input an
encryption-key e and (corresponding) ciphertext β, the circuit Dn operates as follows:
r For every j ≤ i, the circuit Dn generates an encryption of x j using the encryption-
key e. Similarly, for every j ≥ i + 2, the circuit Dn generates an encryption of y j
using the encryption-key e.
Let us denote the resulting ciphertexts by β1 , ..., βi , βi+2 , ..., βt(n) . That is, β j ←
E e (x j ) for j ≤ i and β j ← E e (y j ) for j ≥ i + 2.
r Finally, Dn invokes Cn on input the encryption-key e and the sequence of ciphertexts
β1 , ..., βi , β, βi+2 , ..., βt(n) , and outputs whatever Cn does.
We stress that the construction of Dn relies in an essential way on the fact that the
encryption-key is given to Dn as input.
We now turn to the analysis of the circuit Dn . Suppose that β is a (random)
encryption of xi+1 with (random) key e; that is, β = E e (xi+1 ). Then, Dn (e, β) ≡
(i+1)
Cn (e, E e (h̄ (i+1) )) = C n (Hn ), where X ≡ Y means that the random variables X
and Y are identically distributed. Similarly, for β = E e (yi+1 ), we have Dn (e, β) ≡
391
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

(i)
Cn (e, E e (h̄ (i) )) = C n (Hn ). Thus, by Eq. (5.6), we have


 Pr Dn (G 1 (1n ), E G (1n ) (yi+1 )) = 1
1

 1
−Pr Dn (G 1 (1n ), E G 1 (1n ) (xi+1 )) = 1  >
t(n) · p(n)

in contradiction to our hypothesis that (G, E, D) is a ciphertext-indistinguishable

public-key encryption scheme (in the single-message sense). The theorem
follows.

Discussion. The fact that we are in the public-key model is essential to this proof. It
allows the circuit Dn to form encryptions relative to the same encryption-key used in
the ciphertext given to it. In fact, as previously stated (and proven next), the analogous
result does not hold in the private-key model.

5.2.4.3. The Effect on the Private-Key Model

In contrast to Theorem 5.2.11, in the private-key model, ciphertext-indistinguishability
for a single message does not necessarily imply ciphertext-indistinguishability for mul-
tiple messages.

Proposition 5.2.12: Suppose that there exist pseudorandom generators (robust against
polynomial-size circuits). Then, there exists a private-key encryption scheme that sat-
isfies Definition 5.2.3 but does not satisfy Definition 5.2.9.

Proof: We start with the construction of the desired private-key encryption scheme. The
encryption/decryption key for security parameter n is a uniformly distributed n-bit long
string, denoted s. To encrypt a ciphertext, x, the encryption algorithm uses the key s
as a seed for a (variable-output) pseudorandom generator, denoted g, that stretches
seeds of length n into sequences of length |x|. The ciphertext is obtained by a bit-by-bit
exclusive-or of x and g(s). Decryption is done in an analogous manner.
We first show that this encryption scheme satisfies Definition 5.2.3. Intuitively,
this follow from the hypothesis that g is a pseudorandom generator and the fact that
x ⊕ U|x| is uniformly distributed over {0, 1}|x| . Specifically, suppose toward the contra-
diction that for some polynomial-size circuit family {Cn }, a polynomial p, and infinitely
many n’s

1
| Pr[Cn (x ⊕ g(Un )) = 1] − Pr[Cn (y ⊕ g(Un )) = 1] | >
p(n)

where Un is uniformly distributed over {0, 1}n and |x| = |y| = m = poly(n). On the
other hand,

Pr[Cn (x ⊕ Um ) = 1] = Pr[Cn (y ⊕ Um ) = 1]

392

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Thus, without loss of generality

1
| Pr[Cn (x ⊕ g(Un )) = 1] − Pr[Cn (x ⊕ Um ) = 1] | >
2 · p(n)
Incorporating x into the circuit Cn , we obtain a circuit that distinguishes Um from g(Un ),
in contradiction to our hypothesis (regarding the pseudorandomness of g).
Next, we observe that this encryption scheme does not satisfy Definition 5.2.9.
Specifically, given the ciphertexts of two plaintexts, one may easily retrieve the
exclusive-or of the corresponding plaintexts. That is,

Es (x1 ) ⊕ E s (x2 ) = (x1 ⊕ g(s)) ⊕ (x2 ⊕ g(s)) = x1 ⊕ x 2

This clearly violates Definition 5.2.8 (e.g., consider f (x1 , x2 ) = x1 ⊕ x 2 ) as well as

Definition 5.2.9 (e.g., consider any x̄ = (x 1 , x2 ) and ȳ = (y1 , y2 ) such that x1 ⊕ x2 =
y1 ⊕ y2 ). Viewed in a different way, note that any plaintext-ciphertext pair yields a
corresponding prefix of the pseudorandom sequence, and knowledge of this prefix
violates the security of additional plaintexts. That is, given the encryption of a known
plaintext x1 along with the encryption of an unknown plaintext x2 , we can retrieve
x 2 .5

Discussion. The single-message security of the scheme used in the proof of Propo-
sition 5.2.12 was proven by considering an ideal version of the scheme in which the
pseudorandom sequence is replaced by a truly random sequence. The latter scheme
is secure in an information-theoretic sense, and the security of the actual scheme fol-
lowed by the indistinguishability of the two sequences. As we show in Section 5.3.1, this
construction can be modified to yield a private-key “stream-cipher” that is secure for
multiple message encryptions. All that is needed in order to obtain multiple-message
security is to make sure that (as opposed to this construction) the same portion of the
pseudorandom sequence is never used twice.

An Alternative Proof of Proposition 5.2.12. Given an arbitrary private-key encryption

scheme (G, E, D), consider the following private-key encryption scheme (G  , E  , D  ):
r G  (1n ) = ((k, r ), (k, r )), where (k, k) ← G(1n ) and r is uniformly selected in
{0, 1}|k| ;
r E  (x) = (E k (x), k ⊕ r ) with probability 1/2 and E  (x) = (E k (x), r ) other-
(k,r ) (k,r )
wise;
r and D  z) = Dk (y).
(k,r ) (y,

If (G, E, D) is secure, then so is (G  , E  , D  ) (with respect to a single message); how-

ever, (G  , E  , D  ) is not secure with respect to two messages. For further discussion see
Exercise 21.

5 On input the ciphertexts β1 and β2 , knowing that the first plaintext is x 1 , we first retrieve the pseudorandom
def
sequence (i.e., it is just r = β1 ⊕ x1 ), and next retrieve the second plaintext (i.e., by computing β2 ⊕ r ).

393
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

5.2.5.* A Uniform-Complexity Treatment

As stated at the beginning of this section, the non-uniform complexity formulation
was adopted in this chapter for the sake of simplicity. In contrast, in this subsection,
we outline an alternative definitional treatment of security based on a uniform (rather
than a non-uniform) complexity formulation. We stress that by uniform or non-uniform
complexity treatment of cryptographic primitives, we refer merely to the modeling of the
adversary. The honest (legitimate) parties are always modeled by uniform complexity
classes (most commonly probabilistic polynomial-time).
The notion of efficiently constructible probability ensembles, defined in Section 3.2.3
of Volume 1, is central to the uniform-complexity treatment. Recall that an ensemble,
X = {X n }n∈N , is said to be polynomial-time constructible if there exists a probabilistic
polynomial-time algorithm S so that for every n, the random variables S(1n ) and X n
are identically distributed.

5.2.5.1. The Definitions

We present only the definitions of security for multiple messages; the single-message
variant can be easily obtained by setting the polynomial t (in Definitions 5.2.13 and
5.2.14) to be identically 1. Likewise, we present the public-key version, and the private-
key analogue can be obtained by omitting G 1 (1n ) from the inputs to the various
algorithms.
The uniformity of the following definitions is reflected in the complexity of the
inputs given to the algorithms. Specifically, the plaintexts are taken from polynomial-
time constructible ensembles and so are the auxiliary inputs given to the algo-
rithms. For example, in the following definition we require the ensemble {X n }
to be polynomial-time constructible and the function h to be polynomial-time
computable.

Definition 5.2.13 (semantic security – uniform-complexity version): An encryption

scheme, (G, E, D), is uniformly semantically secure in the public-key model if for
every two polynomials t, , and every probabilistic polynomial-time algorithm A there
exists a probabilistic polynomial-time algorithm A such that for every polynomial-
(1) (t(n)) (i)
time constructible ensemble {X n = (X n , ..., X n )}n∈N , with |X n | = (n), every
polynomial-time computable h : {0, 1} → {0, 1} , every f : {0, 1} → {0, 1}∗ , every
∗ ∗ ∗

positive polynomial p, and all sufficiently large n’s

 
Pr A(1n , G 1 (1n ), E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A (1n , t(n), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

def
where E e (x) = (E e (x (1) ), ..., E e (x (t(n)) )), for x = (x (1) , ..., x (t(n)) ) ∈ {0, 1}t(n)·(n) , is as
in Definition 5.2.8.
394

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Again, we stress that X n is a sequence of random variables, which may depend on

one another. Note that Definition 5.2.13 is seemingly weaker than the corresponding
non-uniform definition (i.e., Definition 5.2.8). We stress that even here (i.e., in
the uniform-complexity setting) no computational limitations are placed on the
function f .

Definition 5.2.14 (indistinguishability of encryptions – uniform-complexity version):

An encryption scheme, (G, E, D), has uniformly indistinguishable encryptions in the
public-key model if for every two polynomials t, , every probabilistic polynomial-time
def
algorithm D  , every polynomial-time constructible ensemble T = {T n = X n Y n Z n }n∈N ,
(1) (t(n)) (1) (t(n)) (i) (i)
with X n = (X n , ..., X n ), Y n = (Yn , ..., Yn ), and |X n | = |Yn | = (n), it holds
that


| Pr D  (1n , Z n , G 1 (1n ), E G 1 (1n ) (X n )) = 1

 1
− Pr D  (1n , Z n , G 1 (1n ), E G 1 (1n ) (Y n )) = 1 | <
p(n)
for every positive polynomial p and all sufficiently large n’s. (The probability in these
terms is taken over T n = X n Y n Z n as well as over the internal coin tosses of the relevant
algorithms.)

The random variable Z n represented a priori information about the plaintexts for which
encryptions should be distinguished. A special case of interest is when Z n = X n Y n .
Uniformity is captured in the requirement that D  be a probabilistic polynomial-time
algorithm (rather than a family of polynomial-size circuits) and that the ensemble {T n =
X n Y n Z n }n∈N be polynomial-time constructible. Recall that in the non-uniform case (i.e.,
Definition 5.2.9), the random variable Z n can be incorporated in the distinguishing
circuit C n (and thus be eliminated).6 Thus, Definition 5.2.14 is seemingly weaker than
the corresponding non-uniform definition (i.e., Definition 5.2.9).

5.2.5.2. Equivalence of the Multiple-Message Definitions

We prove the equivalence of the uniform-complexity definitions (presented earlier) for
(multiple-message) security.

Theorem 5.2.15 (equivalence of definitions – uniform treatment): A public-key en-

cryption scheme satisfies Definition 5.2.13 if and only if it satisfies Definition 5.2.14.
Furthermore, this holds even if Definition 5.2.14 is restricted to the special case where
Z n = X n Y n . Similarly, the equivalence holds even if Definition 5.2.13 is restricted to
the special case where f is polynomial-time computable.

An analogous result holds for the private-key model. The important direction of the
theorem holds also for the single-message version (this is quite obvious from the

6 Furthermore, in the case of non-uniform distinguishers, the auxiliary input 1n is redundant.

395
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

following proof). In the other direction, we seem to use the multiple-message ver-
sion (of semantic security) in an essential way. An alterative treatment is provided in
Exercise 23.

Proof Sketch: Again, we start with the more important direction (i.e., “indistinguisha-
bility” implies semantic security). Specifically, assuming that (G, E, D) has indistin-
guishable encryptions in the uniform sense, even merely in the special case where
Z n = X n Y n , we show that it is semantically secure in the uniform sense. Our construc-
tion of algorithm A is analogous to the construction used in the non-uniform treatment.
Specifically, on input (1n , t(n), 1|α| , h(1n , α)), algorithm A generates a random encryp-
tion of a dummy sequence of plaintexts (i.e., 1|α| ), feeds it to A, and outputs whatever
A does.7 That is,

A (1n , t(n), 1|α| , u) = A(1n , G 1 (1n ), E G 1 (1n ) (1|α| ), 1|α| , u) (5.7)

As in the non-uniform case, the analysis of algorithm A reduces to the following claim.
Claim 5.2.15.1: For every two polynomials t and , every polynomial-time constructible
(1) (t(n)) (i)
ensemble {X n }n∈N , with X n = (X n , ..., X n ) and |X n | = (n), every polynomial-
time computable h, every positive polynomial p, and all sufficiently large n’s
 
Pr A(1n , G 1 (1n ), E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A(1n , G 1 (1n ), E G 1 (1n ) (1|X n | ), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

Proof Sketch: Analogously to the non-uniform case, assuming toward the contradiction
that the claim does not hold yields an algorithm that distinguishes encryptions of X n
from encryptions of Y n = 1|X n | , when getting auxiliary information Z n = X n Y n =
X n 1|X n | . Thus, we derive a contradiction to Definition 5.2.14 (even under the special
case postulated in the theorem).
We note that the auxiliary information that is given to the distinguishing algorithm
replaces the hardwiring of auxiliary information that was used in the non-uniform
case (and is not possible in the uniform-complexity model). Specifically, rather than
using a hardwired value of h (at some non-uniformly fixed sequence), the distinguish-
ing algorithm will use the auxiliary information Z n = X n 1|X n | in order to compute
def
h n (X n ) = (1n , 1|X n | , h(1n , X n )), which it will pass to A. Indeed, we rely on the hypoth-
esis that h is efficiently computable.
The actual proof is quite simple in case the function f is also polynomial-
time computable (which is not the case in general). In this special case, on input
(1n , e, z, E e (α)), where z = (x, 1|x| ) and α ∈ {x, 1|x| } for x ← X n , the distinguishing
algorithm computes u = h(1n , x) and v = f (1n , x), invokes A, and outputs 1 if and
only if A(1n , e, E e (α), 1|x| , u) = v.
7 More accurately, algorithm A proceeds as follows. Using t(n), the algorithm breaks 1|α| into a sequence of t(n)
equal-length (unary) strings, using 1n it generates a random encryption-key, and using this key it generates the
corresponding sequence of encryptions.

396

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

(Tedious comment: In case α = 1|x| , we actually mean that α is a sequence of

t(n) strings of the form 1(n) , where t and  are as in x = (x (1) , ..., x (t(n)) ) ∈
({0, 1}(n) )t(n) .)

The proof becomes more involved in the case where f is not polynomial-time
computable.8 Again, the solution is in realizing that indistinguishability of encryp-
tion postulates a similar output profile (of A) in both cases, where the two cases
correspond to whether A is given an encryption of x or an encryption of 1x (for
x ← X n ). In particular, no value can occur as the output of A in one case with non-
negligibly higher probability than in the other case. To clarify the point, for every fixed
x, we define n,v (x) to be the difference between Pr[A(G 1 (1n ), E G 1 (1n ) (x), h n (x)) = v]
def
and Pr[A(G 1 (1n ), E G 1 (1n ) (1|x| ), h n (x)) = v], where h n (x) = (1n , 1|x| , h(1n , x)) and the
probability space is over the internal coin tosses of algorithms G, E, and A. Taking
the expectation over X n , the contradiction hypothesis means that E[n, f (1n , X n ) (X n )] >
1/ p(n), and so with probability at least 1/2 p(n) over the choice of x ← X n we have
n, f (1n , x) (x) > 1/2 p(n). The problem is that, given x (and 1n ), we cannot even approx-
imate n, f (1n , x) (x), because we do not have the value f (1n , x) (and we cannot compute
def
it). Instead, we let n (x) = maxv∈{0,1}poly(n) {n,v (x)}, and observe that E[n (X n )] ≥
E[n, f (1n , X n ) (X n )] > 1/ p(n). Furthermore, given (1n , x), we can (efficiently) approx-
imate n (x) as well as find a value v such that n,v (x) > n (x) − (1/2 p(n)), with
probability at least 1 − 2−n .

On approximating n (x) and finding an adequate v: Let q(n) be a bound on the

length of f (1n , x). Our goal is to approximate n (x), which is the maximum of
n,v (x) taken over all v ∈ {0, 1}q(n) , as well as find a value v for which n,v (x) is
close to n (x). For each fixed v, it is easy to approximate n,v (x), but we cannot
afford to seperately compute each of these approximations. Yet we can efficiently
compute an implicit representation of all the 2q(n) approximations, where all but
polynomially many of the n,v (x)’s will be approximated by zero. This is possible
because the n,v (x)’s are the differences between corresponding two sequences
of positive numbers (where each sequence has a sum equal to one). Specifically,
def
we obtain m = O((n + q(n)) · p(n)2 ) outputs of A(G 1 (1n ), E G 1 (1n ) (x), h n (x)) and
m outputs of A(G 1 (1n ), E G 1 (1n ) (1|x| ), h n (x)), where in each of the invocations we
use new coin tosses for algorithms A, G, and E. For each v, the quantity n,v (x)
is approximated by the difference between the fraction of times that v occurs as
output in the first case and the fraction of times it occurs as output in the second
case. Indeed, at most, 2m values may occur as outputs, and for all other v’s the
quantity n,v (x) is implicitly approximated by zero. Let us denote by   n,v (x) the
approximation computed (explicitly or implicitly) for n,v (x). Note that for every
fixed v, the probability that |n,v (x) −  n,v (x)| > 1/4 p(n) is at most 2−(n+q(n)) ;
−n
hence, with probability at least 1 − 2 , |n,v (x) −   n,v (x)| ≤ 1/4 p(n) holds for
all v’s. Having computed all these approximations, we just select a string ṽ for
which the approximated quantity  n, ṽ (x) is the largest. To analyze the quality of

8 Unlike in the non-uniform treatment, here we cannot hardwire values (such as the values of h and f on good
sequences) into the algorithm D  , because D  is required to be uniform.

397
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

our selection, let us denote by vn a string s that maximizes n,s (x) (i.e., n,vn (x) =
n (x)). Then, with probability at least 1 − 2−n , the string ṽ satisfies
 n, ṽ (x) − (1/4 p(n))
n, ṽ (x) ≥ 
 n,vn (x) − (1/4 p(n))
≥
≥ n,vn (x) − (1/4 p(n)) − (1/4 p(n))

where the first and last inequalities are due to the quality of our approximations, and
the second inequality is due to the fact that ṽ maximizes  n,· (x). Thus, n, ṽ (x) ≥
n (x) − (1/2 p(n)).

Thus, on input (1n , e, z, E e (α)), where z = (x, 1|x| ), the distinguisher, denoted D  ,
operates in two stages.
1. In the first stage, D  ignores the ciphertext E e (α). Using z, algorithm D  recovers
def
x, and computes u = h n (x) = (1n , 1|x| , h(1n , x)). Using x and u, algorithm D  esti-
mates n (x), and finds a value v as noted. That is, with probability at least 1 − 2−n ,
it holds that n,v (x) > n (x) − (1/2 p(n)).
2. In the second stage (using u and v, as determined in the first stage), algorithm D 
invokes A, and outputs 1 if and only if A(e, E e (α), u) = v.
Let Vn (x) be the value found in the first stage of algorithm A (i.e., obliviously of the
ciphertext E e (α)). The reader can easily verify that

 n 
 Pr D (1 , G 1 (1n ), Z n , E G (1n ) (X n )) = 1
1
 

− Pr D  (1n , G 1 (1n ), Z n , E G 1 (1n ) (1 X n )) = 1 


= E n, Vn (X n ) (X n )
  1
≥ 1 − 2−n · E n (X n ) − − 2−n
2 p(n)

 2 1
> E n (X n ) − >
3 p(n) 3 p(n)
where the first inequality is due to the quality of the first stage (and the 2−n factors ac-
count for the probability that the value found in that stage is bad). Thus, we have derived
a probabilistic polynomial-time algorithm (i.e., D  ) that distinguishes encryptions of X n
from encryptions of Y n = 1|X n | , when getting auxiliary information Z n = X n 1|X n | . By
hypothesis, {X n } is polynomial-time constructible, and it follows that so is {X n Y n Z n }
Thus, we derive contradiction to Definition 5.2.14 (even under the special case postu-
lated in the theorem), and the claim follows.
Having established the important direction, we now turn to the opposite one. That is,
we assume that (G, E, D) is (uniformly) semantically secure and prove that it has (uni-
formly) indistinguishable encryptions. Again, the proof is by contradiction. However,
the proof is more complex than in the non-uniform case, because here “distinguishable
encryptions” means distinguishing between two plaintext-distributions (rather than be-
tween two fixed sequences of plaintexts), when also given a possibly related auxiliary
398

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

input Z n . Thus, it seems that we need to incorporate Z n into the input given to the
(semantic-security) adversary, and the only way to do so seems to be by letting Z n be
part of the a priori information given to that adversary (i.e., letting h(plaintext) = Z n ).
Indeed, this will be part of the construction presented next.
Suppose, without loss of generality, that there exists a probabilistic polynomial-time
def
algorithm D  , a polynomial-time constructible ensemble T = {T n = X n Y n Z n }n∈N (as
in Definition 5.2.14), a positive polynomial p, and infinitely many n’s such that


Pr D  (Z n , G 1 (1n ), E G 1 (1n ) (X n )) = 1

 1
> Pr D  (Z n , G 1 (1n ), E G 1 (1n ) (Y n )) = 1 | +
p(n)
Let t(n) and (n) be such that X n (resp., Y n ) consists of t(n) strings, each of length
(n). Suppose, without loss of generality, that |Z n | = m(n) · (n), and parse Z n into
) ∈ ({0, 1}(n) )m(n) such that Z n = Z n · · · Z n
(1) (m(n)) (1) (m(n))
Z n = (Z n , ..., Z n . We define an
def
auxiliary polynomial-time constructible ensemble Q = {Q n }n∈N such that
0(n) Z n X n Y n with probability 1
Qn = 2
(5.8)
1(n) Z n Y n X n with probability 1
2

That is, Q n is a sequence of 1 + m(n) + 2t(n) strings, each of length (n), that contains
Z n X n Y n in addition to a bit (encoded in the (n)-bit long prefix) indicating whether or
not the order of X n and Y n is switched. We define the function f to be equal to this
“switch”-indicator bit, and the function h to provide all information in Q n except this
switch bit. That is, we define f and h as follows:
r We define f (1n , q) def
= f n (q), where f n returns the first bit of its input; that is,
f n (σ (n) zαβ) = σ , for (z, α, β) ∈ ({0, 1}l(n) )m(n)+2t(n) .
r We define h(1n , q) def = h n (q), where h n reorders the suffix of its input according to
the first bit; that is, h n (0(n) zαβ) = zαβ and h n (1(n) zαβ) = zβα. Thus, h(1n , Q n ) =
Z n X n Y n , where Z n X n Y n is determined by T n = X n Y n Z n (and is independent of
the switch-case chosen in Eq. (5.8)).
We stress that both h and f are polynomial-time computable.
We will show that the distinguishing algorithm D  (which distinguishes E(X n ) from
E(Y n ), when also given Z n ≡ Z n ) can be transformed into a polynomial-time algo-
rithm A that guesses the value of f (1n , Q n ), from the encryption of Q n (and the value
of h(1n , Q n )), and does so with probability non-negligibly greater than 1/2. This vio-
lates semantic security, since no algorithm (regardless of its running time) can guess
f (1n , Q n ) with probability greater than 1/2 when only given h(1n , Q n ) and 1|Q n | (be-
cause, conditioned on the value of h(1n , Q n ) (and 1|Q n | ), the value of f (1n , Q n ) is
uniformly distributed over {0, 1}).
On input (e, E e (α), 1|α| , h(1n , α)), where α = σ (n) z u v ∈ ({0, 1}l(n) )1+m(n)+2t(n)
equals either (0(n) , z, x, y) or (1(n) , z, y, x), algorithm A proceeds in two stages:
1. In the first stage, algorithm A ignores the ciphertext E e (α). It first extracts
x, y and z ≡ z out of h(1n , α) = z x y, and approximates n (z, x, y), which is
399
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

defined to equal



Pr D  (z, G 1 (1n ), E G 1 (1n ) (x)) = 1 − Pr D  (z, G 1 (1n ), E G 1 (1n ) (y)) = 1
Specifically, using O(n · p(n)2 ) samples, algorithm A obtains an approximation, de-
noted  n (z, x, y), such that | n (z, x, y) − n (z, x, y)| < 1/3 p(n) with probability
at least 1 − 2−n .
Algorithm A sets ξ = 1 if   n (z, x, y) > 1/3 p(n), sets ξ = −1 if  n (z, x, y) <

−1/3 p(n), and sets ξ = 0 otherwise (i.e., |n (z, x, y)| ≤ 1/3 p(n)). Intuitively, ξ
indicates the sign of   n (z, x, y), provided that the absolute value of the latter is
large enough, and is set to zero otherwise. In other words, with overwhelmingly high
probability, ξ indicates whether the value of Pr[D  (z, ·, E · (x)) = 1] is significantly
greater, smaller, or about the same as Pr[D  (z, ·, E · (y)) = 1].
In case ξ = 0, algorithm A halts with an arbitrary reasonable guess (say a randomly
selected bit). (We stress that all this is done obliviously of the ciphertext E e (α),
which is only used next.)
2. In the second stage, algorithm A extracts the last block of ciphertexts (i.e., E e (v))
out of E e (α) = E e (σ (n) z u v), and invokes D  on input (z, e, E e (v)), where z is as
extracted in the first stage. Using the value of ξ as determined in the first stage,
algorithm A decides (i.e., determines its output bit) as follows:
r In case ξ = 1, algorithm A outputs 1 if and only if the output of D  is 1.
r In case ξ = −1, algorithm A outputs 0 if and only if the output of D  is 1.
That is, ξ = 1 (resp., ξ = −1) indicates that D  is more (resp., less) likely to output
1 when given the encryption of x than when given the encryption of y.

Claim 5.2.15.2: Let p, Q n , h, f , and A be as in Eq. (5.8) and the text that follows it.

 1 1
Pr A(G 1 (1n ), E G 1 (1n ) (Q n ), h(1n , Q n )) = f (1n , Q n ) > +
2 7 · p(n)

Proof Sketch: We focus on the case in which the approximation of n (z, x, y) computed
by (the first stage of) A is within 1/3 p(n) of the correct value. Thus, in case ξ = 0, the
sign of ξ agrees with the sign of n (z, x, y). It follows that for every possible (z, x, y)
such that ξ = 1 (it holds that n (z, x, y) > 0 and) the following holds:


Pr A(G 1 (1n ), E G 1 (1n ) (Q n ), h(1n , Q n )) = f (1n , Q n ) | (Z n , X n , X n ) = (z, x, y)
1

= · Pr A(G 1 (1n ), E G 1 (1n ) (0(n) , z, x, y), h n (0(n) , z, x, y)) = 0
2
1

+ · Pr A(G 1 (1n ), E G 1 (1n ) (1(n) , z, y, x), h n (1(n) , z, y, x)) = 1
2
1

= · Pr D  (z, G 1 (1n ), E G 1 (1n ) (y)) = 0
2
1

+ · Pr D  (z, G 1 (1n ), E G 1 (1n ) (x)) = 1
2
1
= · (1 + n (z, x, y))
2
400

www.Ebook777.com
 5.2 DEFINITIONS OF SECURITY

Similarly, for every possible (z, x, y) such that ξ = −1 (it holds that n (z, x, y) < 0
and) the following holds:


Pr A(G 1 (1n ), E G 1 (1n ) (Q n ), h(1n , Q n )) = f (1n , Q n ) | (Z n , X n , X n ) = (z, x, y)
1
= · (1 − n (z, x, y))
2
Thus, in both cases where ξ = 0, algorithm A succeeds with probability
1 + ξ · n (z, x, y) 1 + |n (z, x, y)|
=
2 2
and in case ξ = 0 it succeeds with probability 1/2, which is (artificially) lower-bounded
by (1 + |n (z, x, y)| − (2/3 p(n)))/2 (because |n (z, x, y)| ≤ 2/3 p(n) for ξ = 0).9
Thus, ignoring the negligible probability that the approximation deviated from the
correct value by more than 1/3 p(n), the overall success probability of algorithm A is

1 + |n (Z n , X n , Y n )| − (2/3 p(n)) 1 + E[n (Z n , X n , Y n )] − (2/3 p(n))

E ≥
2 2
1 + (1/ p(n)) − (2/3 p(n)) 1 1
> = +
2 2 6 p(n)
where the last inequality follows by the contradiction hypothesis (i.e., that
E[n (Z n , X n , Y n )] > p(n)
1
). The claim follows (because the negligible probability ig-
nored in the last [displayed] inequality is certainly upper-bounded by (1/6 p(n)) −
(1/7 p(n))).

This completes the proof of the opposite direction.

Discussion. The proof of the first (i.e., important) direction holds also in the single-
message setting. In general, for any function t, in order to prove that semantic security
holds with respect to t-long sequences of ciphertexts, we just use the hypothesis that t-
long message-sequences have indistinguishable encryptions. In contrast, the proof of the
second (i.e., opposite) direction makes an essential use of the multiple-message setting.
In particular, in order to prove that t-long message-sequences have indistinguishable
encryptions, we use the hypothesis that semantic security holds with respect to (1 +
m + 2t)-long sequences of ciphertexts, where m depends on the length of the auxiliary
input in the claim of ciphertext-indistinguishability. Thus, even if we only want to
establish ciphertext-indistinguishability in the single-message setting, we do so by
using semantic security in the multiple-message setting. Furthermore, we use the fact
that given a sequence of ciphertexts, we can extract a certain subsequence of ciphertexts.

9 This analysis looks somewhat odd but is nevertheless valid. Our aim is to get a “uniform” expression for
the success probability of A in all cases (i.e., for all values of ξ ). In case |ξ | = 1, we have the lower-bound
(1 + |n (z, x, y)|)/2, which is certainly lower-bounded by (1 + |n (z, x, y)| − (2/3 p(n)))/2, whereas in case
ξ = 0 we artificially lower-bound 1/2 by the same expression. Once we have such a “uniform” expression, we
may take expectation over it (without breaking it to cases).

401
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

5.2.5.3. Single-Message Versus Multiple-Message Definitions

As in the non-uniform case, for the public-key model, single-message security implies
multiple-message security. Again, this implication does not hold in the private-key
model. The proofs of both statements are analogous to the proofs provided in the non-
uniform case. Specifically:

1. For the public-key model, single-message uniform-indistinguishability of encryp-

tions implies multiple-message uniform-indistinguishability of encryptions, which
in turn implies multiple-message uniform-semantic security.
In the proof of this result, we use the fact that all hybrids are polynomial-time
constructible, and that we may select a random pair of neighboring hybrids (as
in the proof of Theorem 3.2.6). We also use the fact that an ensemble of triplets,
{T n = X n Y n Z n }n∈N , with X n = (X n , ..., X n ), Y n = (Yn , ..., Yn ), as in Defi-
(1) (t(n)) (1) (t(n))

nition 5.2.14, induces an ensemble of triplets, {Tn = X n Yn Z n }n∈N , for the case t ≡ 1.
Specifically, we shall use X n = X n , Yn = Yn , and Z n = (X n , Y n , Z n , i), where i
(i) (i)

is uniformly distributed in {1, ..., t(n)}.

2. For the private-key model, single-message uniform-indistinguishability of encryp-
tions does not imply multiple-message uniform-indistinguishability of encryptions.
The proof is exactly as in the non-uniform case.

5.2.5.4. The Gain of a Uniform Treatment

Suppose that one is content with the uniform-complexity level of security, which is what
we advocate in the following paragraph. Then the gain in using the uniform-complexity
treatment is that a uniform-complexity level of security can be obtained using only
uniform-complexity assumptions (rather than non-uniform-complexity assumptions).
Specifically, the results presented in Section 5.3 are based on non-uniform assumptions
such as the existence of functions that cannot be inverted by polynomial-size circuits
(rather than by probabilistic polynomial-time algorithms). These non-uniform assump-
tions are used in order to satisfy the non-uniform definitions presented in Sections 5.2.1
and 5.2.2. Using any of these constructions, while making the analogous uniform as-
sumptions, yields encryption schemes with the analogous uniform-complexity security.
(We stress that this is no coincidence, but is rather an artifact of these construction being
proven secure via a uniform reducibility argument.)
However, something is lost when relying on these (seemingly weaker) uniform-
complexity assumptions. Namely, the security we obtain is only against the (seemingly
weaker) uniform adversaries. We believe that this loss in security is immaterial. In par-
ticular, schemes secure against probabilistic polynomial-time adversaries can be used in
any setting consisting of probabilistic polynomial-time machines with inputs generated
by probabilistic polynomial-time procedures. We believe that the cryptographic setting
is such a case. That is, we believe that the world should be modeled as a probabilistic
polynomial-time procedure; namely, each object in the world was generated at some
point by a uniform and efficient procedure (possibly depending on previously existing
402

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

objects), rather than handed down from heaven (where it might have been selected
non-uniformly or using non-recursive procedures).

5.2.5.5. An Alternative Treatment

An alternative uniform-complexity treatment of security (in the current passive setting)
can be derived from the treatment of security under chosen plaintext attacks (presented
in Section 5.4.3). Specifically, the definitions presented in Section 5.4.3.1 should be
modified as follows:
r Replace the input pair (e, z), which is given to the attacker’s first part (i.e., A1 ), by 1n .
That is, eliminate the (non-uniform) auxiliary input z, and omit the encryption-key
e (also in the public-key version).
r Remove the encryption oracle (i.e., E e ) from the definitions; that is, model the
attacker by an ordinary (probabilistic polynomial-time) algorithm, rather than by an
oracle machine.
Consequently, the definition of semantic security (Definition 5.4.8) can be simplified
by using A1 = A1 and omitting Condition 2 (which refers to the distributions produced
by A1 and A1 ). Doing so requires a minor change in the first part of the proof of
Theorem 5.4.11 (i.e., letting A2 rather than A1 generate a random encryption-key).
In the resulting definitions, the first part of the attacker is confined to an oblivious
selection of a challenge template (i.e., the challenge template is selected independently
of the encryption-key), whereas the second part of the attacker is given an adequate
challenge (and nothing else). In the case of semantic security, this means that the adver-
sary first selects the “application” that consists of the plaintext distribution, the partial
information function h, and the desired information function f . These three objects
are represented by circuits. Next, a plaintext x is selected according to the specified
distribution, and the adversary is given a corresponding ciphertext (i.e., E e (x)), as well
as the corresponding partial information h(x).

5.3. Constructions of Secure Encryption Schemes

In this section we present constructions of secure private-key and public-key encryption

schemes. Here and throughout this section security means semantic security in the
multiple-message setting. Recall that this is equivalent to ciphertext-indistinguishability
(in the multiple-message setting). Also recall that for public-key schemes it suffices to
prove ciphertext-indistinguishability in the single-message setting. Following are the
main results of this section:
r Using any (non-uniformly robust) pseudorandom function, one can construct secure
private-key encryption schemes. Recall that the former can be constructed using any
(non-uniformly strong) one-way function.
r Using any (non-uniform strong) trapdoor one-way permutation, one can construct
secure public-key encryption schemes.
403
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

In addition, we review some popular suggestions for private-key and public-key en-
cryption schemes.

Probabilistic Encryption. Before starting, we recall that a secure public-key encryp-

tion scheme must employ a probabilistic (i.e., randomized) encryption algorithm. Oth-
erwise, given the encryption-key as (additional) input, it is easy to distinguish the
encryption of the all-zero message from the encryption of the all-ones message. The
same holds for private-key encryption schemes when considering the multi-message
setting.10 For example, using a deterministic (private-key) encryption algorithm allows
the adversary to distinguish two encryptions of the same message from the encryptions
of a pair of different messages. Thus, the common practice of using pseudorandom per-
mutations as “block-ciphers” (see definition in Section 5.3.2) is not secure (again, one
can distinguish two encryptions of the same message from encryptions of two different
messages). This explains the linkage between our security definitions and randomized
(aka probabilistic) encryption schemes. Indeed, all our encryption schemes will employ
randomized encryption algorithms.11

5.3.1.* Stream-Ciphers
It is common practice to use “pseudorandom generators” as a basis for private-key
stream-ciphers (see definition in Section 5.3.1.1). Specifically, the pseudorandom gen-
erator is used to produce a stream of bits that are XORed with the corresponding
plaintext bits to yield corresponding ciphertext bits. That is, the generated pseudoran-
dom sequence (which is determined by the a priori shared key) is used as a “one-time
pad” instead of a truly random sequence, with the advantage that the generated se-
quence may be much longer than the key (whereas this is not possible for a truly
random sequence). This common practice is indeed sound, provided one actually uses
pseudorandom generators (as defined in Section 3.3 of Volume 1), rather than programs
that are called “pseudorandom generators” but actually produce sequences that are easy
to predict (such as the linear congruential generator or some modifications of it that
output a constant fraction of the bits of each resulting number).
As we shall see, by using any pseudorandom generator one may obtain a secure
private-key stream-cipher that allows for the encryption of a stream of plaintext bits.
We note that such a stream-cipher does not conform to our formulation of an encryption
scheme (i.e., as in Definition 5.1.1), because in order to encrypt several messages one is
required to maintain a counter (to prevent reusing parts of the pseudorandom “one-time
pad”). In other words, we obtain a secure encryption scheme with a variable state that
is modified after the encryption of each message. We stress that constructions of secure

10 We note that this does not hold with respect to private-key schemes in the single-message setting (or for the
augmented model of state-based ciphers discussed in Section 5.3.1). In such a case, the private-key can be
augmented to include a seed for a pseudorandom generator, the output of which can be used to eliminate
randomness from the encryption algorithm. (Question: Why does the argument fail in the public-key setting
and in the multi-message private-key setting?)
11 The (private-key) stream-ciphers discussed in Section 5.3.1 are an exception, but (as further explained in Sec-

tion 5.3.1) these schemes do not adhere to our (basic) formulation of encryption schemes (as in Definition 5.1.1).

404

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

and stateless encryption schemes (i.e., conforming with Definition 5.1.1) are known
and are presented in Sections 5.3.3 and 5.3.4. The traditional interest in stream-ciphers
is due to efficiency considerations. We discuss this issue at the end of Section 5.3.3.
But before doing so, let us formalize the discussion.

5.3.1.1. Definitions
We start by extending the simple mechanism of encryption schemes (as presented
in Definition 5.1.1). The key-generation algorithm remains unchanged, but both the
encryption and decryption algorithm take an additional input and emit an additional
output, corresponding to their state before and after the operation. The length of the state
is not allowed to grow by too much during each application of the encryption algorithm
(see Item 3 in Definition 5.3.1), or else the efficiency of the entire “repeated encryption”
process cannot be guaranteed. For the sake of simplicity, we incorporate the key in the
state of the corresponding algorithm. Thus, the initial state of each of the algorithms is
set to equal its corresponding key. Furthermore, one may think of the intermediate states
as updated values of the corresponding key. For clarity, the reader may consider the
special case in which the state contains the initial key, the number of times the scheme
was invoked (or the total number of bits in such invocations), and auxiliary information
that allows a speedup of the computation of the next ciphertext (or plaintext).
For simplicity, we assume that the decryption algorithm (i.e., D) is deterministic
(otherwise formulating the reconstruction condition would be more complex). Intu-
itively, the main part of the reconstruction condition (i.e., Item 2 in Definition 5.3.1)
is that the (proper) iterative encryption–decryption process recovers the original plain-
texts. The additional requirement in Item 2 is that the state of the decryption algorithm
is updated correctly so long as it is fed with strings of length equal to the length of
the valid ciphertexts. The reason for this additional requirement is discussed following
Definition 5.3.1. We comment that in traditional stream-ciphers, the plaintexts (and ci-
phertexts) are individual bits or blocks of a fixed number of bits (i.e., |α (i) | = |β (i) | = 
for all i’s).

Definition 5.3.1 (state-based cipher – the mechanism): A state-based encryption

scheme is a triple, (G, E, D), of probabilistic polynomial-time algorithms satisfying
the following three conditions:

1. On input 1n , algorithm G outputs a pair of bit strings.

2. For every pair (e(0) , d (0) ) in the range of G(1n ), and every sequence of plaintexts α (i) ’s,
the following holds: If (e(i) , β (i) ) ← E(e(i−1) , α (i) ) and (d (i) , γ (i) ) ← D(d (i−1) , β (i) ),
for i = 1, 2, ..., then γ (i) = α (i) for every i. Furthermore, for every i and every β ∈
{0, 1}|β | , it holds that D(d (i−1) , β) = (d (i) , ·). That is, d (i) is actually determined by
(i)

d (i−1) and |β (i) |.12

12 Alternatively, we may decompose the decryption (resp., encryption) algorithm into two algorithms, where the
first takes care of the actual decryption (resp., encryption) and the second takes care of updating the state. For
details see Exercise 24.

405
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

3. There exists a polynomial p such that for every pair (e(0) , d (0) ) in the range of
G(1n ), and every sequence of α (i) ’s and e(i) ’s as in Item 2, it holds that |e(i) | ≤
|e(i−1) | + |α (i) | · p(n). Similarly for the d (i) ’s.

That is, as in Definition 5.1.1, the encryption–decryption process operates properly

(i.e., the decrypted message equals the plaintext), provided that the corresponding
algorithms get the corresponding keys (or states). Note that in Definition 5.3.1, the
keys are modified by the encryption–decryption process, and so correct decryption
requires holding the correctly updated decryption-key. We stress that the furthermore-
clause in Item 2 guarantees that the decryption-key is correctly updated so long as the
decryption process is fed with strings of the correct lengths (but not necessarily with the
correct ciphertexts). This extra requirement implies that given the initial decryption-
key and the current ciphertext, as well as the lengths of all previous ciphertexts (which
may be actually incorporated in the current ciphertext), one may recover the current
plaintext. This fact is interesting for two reasons:

A theoretical reason: It implies that without loss of generality (albeit with possible
loss in efficiency), the decryption algorithm may be stateless. Furthermore, without
loss of generality (again, with possible loss in efficiency), the state of the encryption
algorithm may consist of the initial encryption-key and the lengths of the plaintexts
encrypted so far.
A practical reason: It allows for recovery from the loss of some of the ciphertexts. That
is, assuming that all ciphertexts have the same (known) length (which is typically
the case in the relevant applications), if the receiver knows (or is given) the total
number of ciphertexts sent so far, then it can recover the current plaintext from the
current ciphertext, even if some of the previous ciphertexts were lost. See the special
provision in Construction 5.3.3.

We comment that in Construction 5.3.3, it holds that |e(i) | ≤ |e(0) | + log2 ij=1 |α ( j) |,
which is much stronger than the requirement in Item 3 (of Definition 5.3.1).
We stress that Definition 5.3.1 refers to the encryption of multiple messages (and
meaningfully extends Definition 5.1.1 only when considering the encryption of multiple
messages). However, Definition 5.3.1 by itself does not explain why one should encrypt
the ith message using the updated encryption-key e(i−1) , rather than reusing the initial
encryption-key e(0) in all encryptions (where decryption is done by reusing the initial
decryption-key d (0) ). Indeed, the reason for updating these keys is provided by the
following security definition that refers to the encryption of multiple messages, and
holds only in case the encryption-keys in use are properly updated (in the multiple-
message encryption process). Here we present only the semantic security definition for
private-key schemes.

Definition 5.3.2 (semantic security – state-based cipher): For a state-based encryp-

tion scheme, (G, E, D), and any x = (x (1) , ..., x (t) ), we let E e (x) = (y (1) , ..., y (t) ) be
def
the result of the following t-step (possibly random) process, where e(0) = e. For
i = 1, ..., t, we let (e(i) , y (i) ) ← E(e(i−1) , x (i) ), where each of the t invocations E utilizes
406

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

independently chosen random coins. The scheme (G, E, D) is semantically secure in

the state-based private-key model if for every polynomial t and every probabilistic
polynomial-time algorithm A there exists a probabilistic polynomial-time algorithm A
(1) (t(n))
such that for every {X n = (X n , ..., X n )}n∈N , f , h, p, and n as in Definition 5.2.8,
it holds that
 
Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A (1n , t(n), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

Note that Definition 5.3.2 (only) differs from Definition 5.2.8 in the preamble defin-
ing the random variable E e (x), which mandates that the encryption-key e(i−1) is used
in the ith encryption. Furthermore, Definition 5.3.2 guarantees nothing regarding
an encryption process in which the plaintext sequence x (1) , ..., x (t) is encrypted by
E(e, x (1) ), E(e, x (2) ), ..., E(e, x (t) ) (i.e., the initial encryption-key e itself is used in all
encryptions, as in Definition 5.2.8).

5.3.1.2. A Sound Version of a Common Practice

Using any (on-line) pseudorandom generator, one can easily construct a secure state-
based private-key encryption scheme. Recall that on-line pseudorandom generators are
a special case of variable-output pseudorandom generators (see Section 3.3.3), in which
a hidden state is maintained and updated so as to allow generation of the next output
bit in time polynomial in the length of the initial seed, regardless of the number of bits
generated so far. Specifically, the next (hidden) state and output bit are produced by
applying a (polynomial-time computable) function g : {0, 1}n → {0, 1}n+1 to the current
state (i.e., s  σ ← g(s), where s is the current state, s  is the next state and σ is the next
output bit). The suggested state-based private-key encryption scheme will be initialized
with a key equal to the seed of such a generator, and will maintain and update a state
allowing it to quickly produce the next output bit of the generator. The stream of
plaintext bits will be encrypted by XORing these bits with the corresponding output
bits of the generator.

Construction 5.3.3 (how to construct stream-ciphers [i.e., state-based private-key

encryption schemes]): Let g be a polynomial-time computable function such that
|g(s)| = |s| + 1 for all s ∈ {0, 1}∗ .

Key-generation and initial state: On input 1n , uniformly select s ∈ {0, 1}n , and output
the key-pair (s, s). The initial state of each algorithm is set to (s, 0, s).
(We maintain the initial key s and a step-counter in order to allow recovery from
loss of ciphertexts.)
Encrypting the next plaintext bit x with state (s, t, s  ): Let s  σ = g(s  ), where |s  | =
|s  | and σ ∈ {0, 1}. Output the ciphertext bit x ⊕ σ , and set the new state to (s, t +
1, s  ).
407
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Decrypting the ciphertext bit y with state (s, t, s  ): Let s  σ = g(s  ), where |s  | = |s  |
and σ ∈ {0, 1}. Output the plaintext bit y ⊕ σ , and set the new state to (s, t + 1, s  ).
Special recovery procedure: When notified that some ciphertext bits may have been
lost and that the current ciphertext bit has index t  , the decryption procedure first
recovers the correct current state, denoted st  , to be used in decryption instead of s  .
def
This can be done by computing si σi = g(si−1 ), for i = 1, ..., t  , where s0 = s.13
Note that both the encryption and decryption algorithms are deterministic, and that the
state after encryption of t bits has length 2n + log2 t < 3n (for t < 2n ).
Recall that g (as in Construction 5.3.3) is called a next-step function of an on-
p
line pseudorandom generator if for every polynomial p the ensemble {G n }n∈N is
p
pseudorandom (with respect to polynomial-size circuits), where G n is defined by the
following random process:
Uniformly select s0 ∈ {0, 1}n ;
For i = 1 to p(n), let si σi ← g(si−1 ), where σi ∈ {0, 1} (and si ∈ {0, 1}n );
Output σ1 σ2 · · · σ p(n) .
Also recall that if g is itself a pseudorandom generator, then it constitutes a next-step
function of an on-line pseudorandom generator (see Exercise 21 of Chapter 3). We
have:

Proposition 5.3.4: Suppose that g is a next-step function of an on-line pseudoran-

dom generator. Then Construction 5.3.3 constitutes a secure state-based private-key
encryption scheme.

Proof Idea: Consider an ideal version of Construction 5.3.3 in which a truly random
sequence is used instead of the output produced by the on-line pseudorandom gener-
ator defined by g. The ideal version coincides with the traditional one-time pad, and
thus is perfectly secure. The security of the actual Construction 5.3.3 follows by the
pseudorandomness of the on-line generator.

5.3.2. Preliminaries: Block-Ciphers

Many encryption schemes are conveniently presented by first constructing a restricted
type of encryption scheme that we call a block-cipher.14 In contrast to encryption
schemes (as defined in Definition 5.1.1), block-ciphers (as defined in Definition 5.3.5)
are only required to operate on plaintexts of a specific length (which is a function of
the security parameter). As we shall see, given a secure block-cipher, we can easily
construct a (general) secure encryption scheme.

13 More generally, if the decryption procedure holds the state at time t < t  then it needs only compute st+1 , ..., st  .
14 In using the term block-cipher, we abuse standard terminology by which a block-cipher must, in addition to op-
erating on plaintext of specific length, produce ciphertexts of a length that equals the length of the corresponding
plaintexts. We comment that the latter cannot be semantically secure; see Exercise 25.

408

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

5.3.2.1. Definitions
We start by considering the syntax (cf. Definition 5.1.1).

Definition 5.3.5 (block-cipher): A block-cipher is a triple, (G, E, D), of probabilistic

polynomial-time algorithms satisfying the following two conditions:

1. On input 1n , algorithm G outputs a pair of bit strings.

2. There exists a polynomially bounded function  : N → N, called the block length, so
that for every pair (e, d) in the range of G(1n ), and for each α ∈ {0, 1}(n) , algorithms
E and D satisfy

Pr[Dd (E e (α)) = α] = 1

Typically, we use either (n) = (n) or (n) = 1. Analogously to Definition 5.1.1, this
definition does not distinguish private-key encryption schemes from public-key ones.
The difference between the two types is captured in the security definitions, which are
essentially as before, with the modification that we only consider plaintexts of length
(n). For example, the analogue of Definition 5.2.8 (for private-key schemes) reads:

Definition 5.3.6 (semantic security – private-key block-ciphers): A block-cipher,

(G, E, D), with block length  is semantically secure (in the private-key model) if for
every probabilistic polynomial-time algorithm A there exists a probabilistic polynomial-
time algorithm A such that for every ensemble {X n = (X n , ..., X n )}n∈N , with
(1) (t(n))
(1) (t(n))
|X n | = · · · = |X n | = (n) and t(n) ≤ poly(n), every pair of polynomially bounded
functions f, h, every positive polynomial p, and all sufficiently large n, it holds that
 
Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< Pr A (1n , t(n), 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

where E e (x (1) , ..., x (t) ) = (E e (x (1) ), ..., E e (x (t) )), as in Definition 5.2.8.

Note that, in case  is polynomial-time computable, we can omit the auxiliary input
1|X n | = 1t(n)·(n) , because it can be reconstructed from the security parameter n and the
value t(n).

5.3.2.2. Transforming Block-Ciphers into General Encryption Schemes

There are obvious ways of transforming a block-cipher into a general encryption
scheme. The basic idea is to break the plaintexts (for the resulting scheme) into blocks
and encode each block separately by using the block-cipher. Thus, the security of the
block-cipher (in the multiple-message settings) implies the security of the resulting
encryption scheme. The only technicality we need to deal with is how to encrypt plain-
texts of length that is not an integer multiple of the block-length (i.e., of (n)). This
409
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

is easily resolved by padding the last block (while indicating the end of the actual
plaintext).15

Construction 5.3.7 (from block-ciphers to general encryption schemes): Let

(G, E, D) be a block-cipher with block length function . We construct an encryp-
tion scheme, (G  , E  , D  ), as follows. The key-generation algorithm, G  , is identical to
G. To encrypt a message α (with encryption-key e generated under security parameter
n), we break it into consecutive blocks of length (n), while possibly augmenting the
last block. Let α1 , ..., αt be the resulting blocks. Then

E e (α) = (|α|, E e (α1 ), ..., E e (αt ))

def

To decrypt the ciphertext (m, β1 , ..., βt ) (with decryption-key d), we let αi = Dd (βi )
for i = 1, ..., t, and let the plaintext be the m-bit long prefix of the concatenated string
α1 · · · αt .

This construction yields ciphertexts that reveal the exact length of the plaintext. Recall
that this is not prohibited by the definitions of security, and that we cannot hope to totally
hide the plaintext length. However, we can easily construct encryption schemes that hide
some information about the length of the plaintext; see examples in Exercise 5. Also,
note that the above construction applies even to the special case where  is identically 1.

Proposition 5.3.8: Let (G, E, D) and (G  , E  , D  ) be as in Construction 5.3.7. Suppose

that the former is a secure private-key16 (resp., public-key) block-cipher. Then the latter
is a secure private-key (resp., public-key) encryption scheme.

Proof Sketch: The proof is by a reducibility argument. Assuming toward the contra-
diction that the encryption scheme (G  , E  , D  ) is not secure, we conclude that neither
is (G, E, D), contradicting our hypothesis. Specifically, we rely on the fact that in
both schemes, security means security in the multiple-message setting. Note that in
case the security of (G  , E  , D  ) is violated via t(n) messages of length L(n), the se-
curity of (G, E, D) is violated via t(n) · L(n)/(n) messages of length (n). Also,
the argument may utilize any of the two notions of security (i.e., semantic security or
ciphertext-indistinguishability).

5.3.3. Private-Key Encryption Schemes

Secure private-key encryption schemes can be easily constructed using any efficiently
computable pseudorandom function ensemble (see Section 3.6). Specifically, we present
a block-cipher with block length (n) = n. The key-generation algorithm consists of

15 We choose to use a very simple indication of the end of the actual plaintext (i.e., to include its length in the
ciphertext). In fact, it suffices to include the length of the plaintext modulo (n). Another natural alternative
is to use a padding of the form 10((n)−|α|−1) mod (n) , while observing that no padding is ever required in case
(n) = 1.
16 Recall that throughout this section security means security in the multiple-message setting.

410

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

selecting a seed, denoted s, for such a function, denoted f s . To encrypt a message

x ∈ {0, 1}n (using key s), the encryption algorithm uniformly selects a string r ∈ {0, 1}n
and produces the ciphertext (r, x ⊕ f s (r )). To decrypt the ciphertext (r, y) (using key
s), the decryption algorithm just computes y ⊕ f s (r ). Formally, we have:

Construction 5.3.9 (a private-key block-cipher based on pseudorandom functions):

Let F = {Fn } be an efficiently computable function ensemble and let I and V be the
algorithms associated with it. That is, I (1n ) selects a function with distribution Fn and
V (s, x) returns f s (x), where f s is the function associated with the string s. We define a
private-key block-cipher, (G, E, D), with block length (n) = n as follows:

Key-generation: G(1n ) = (k, k), where k ← I (1n ).

Encrypting plaintext x ∈ {0, 1}n (using the key k): E k (x) = (r, V (k, r ) ⊕ x), where r
is uniformly chosen in {0, 1}n .
Decrypting ciphertext (r, y) (using the key k): Dk (r, y) = V (k, r ) ⊕ y.

Clearly, for every k (in the range of I (1n )) and x ∈ {0, 1}n ,

Dk (E k (x)) = Dk (Un , f k (Un ) ⊕ x) = f k (Un ) ⊕ ( f k (Un ) ⊕ x) = x

We assume that F is pseudorandom with respect to polynomial-size circuits, meaning

that no polynomial-size circuit having “oracle gates” can distinguish the case in which
the answers are provided by a random function from the case in which the answers are
provided by a function in F. Alternatively, one may consider probabilistic polynomial-
time oracle machines that obtain a non-uniform polynomially long auxiliary input.
That is, for every probabilistic polynomial-time oracle machine M, for every pair of
positive polynomials p and q, and for all sufficiently large n’s and all z ∈ {0, 1} p(n) ,


φ 
 1
 Pr M (z) = 1 − Pr M f I (1n ) (z) = 1  <
q(n)

where φ is a uniformly selected function mapping {0, 1}n to {0, 1}n .

Analogously to Corollary 3.6.7, such (non-uniformly strong) pseudorandom functions
can be constructed using any non-uniformly strong one-way function.

Proposition 5.3.10: Let F and (G, E, D) be as in Construction 5.3.9, and suppose

that F is pseudorandom with respect to polynomial-size circuits. Then (G, E, D) is
secure.

The proof of Proposition 5.3.10 follows. Combining Propositions 5.3.8 and 5.3.10 (with
a non-uniform version of Corollary 3.6.7), we obtain:

Theorem 5.3.11: If there exist (non-uniformly strong) one-way functions, then there
exist secure private-key encryption schemes.
411
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

The converse holds too; see Exercise 2.

Proof of Proposition 5.3.10: The proof consists of two steps (suggested as a general
methodology in Section 3.6):

1. Prove that an idealized version of the scheme, in which one uses a uniformly selected
function φ : {0, 1}n → {0, 1}n , rather than the pseudorandom function f s , is secure
(in the sense of ciphertext-indistinguishability).
2. Conclude that the real scheme (as presented in Construction 5.3.9) is secure (because
otherwise one could distinguish a pseudorandom function from a truly random one).
Specifically, in the ideal version, the messages x (1) , ..., x (t) are encrypted by
(r (1) , φ(r (1) ) ⊕ x (1) ), ..., (r (t) , φ(r (t) ) ⊕ x (t) ), where the r ( j) ’s are independently and uni-
formly selected, and φ is a random function. Thus, with probability greater than
1 − t 2 · 2−n , the r ( j) ’s are all distinct, and so the values φ(r ( j) ) ⊕ x ( j) are independently
and uniformly distributed, regardless of the x ( j) ’s. It follows that the ideal version is
ciphertext-indistinguishable; that is, for any x (1) , ..., x (t) and y (1) , ..., y (t) , the statisti-
(1) (1) (t) (t)
cal difference between the distributions (Un , φ(Un ) ⊕ x (1) ), ..., (Un , φ(Un ) ⊕ x (t) )
(1) (1) (t) (t) −n
and (Un , φ(Un ) ⊕ y ), ..., (Un , φ(Un ) ⊕ y ) is at most t · 2 .
(1) (t) 2

Now, if the actual scheme is not ciphertext-indistinguishable, then for some sequence
of r ( j) ’s and v ( j) ’s, a polynomial-size circuit can distinguish the φ(r ( j) ) ⊕ v ( j) ’s from
the f s (r ( j) ) ⊕ v ( j) ’s, where φ is random and f s is pseudorandom.17 But this contra-
dicts the hypothesis that polynomial-size circuits cannot distinguish between the two
cases.

Discussion. Note that we could have gotten rid of the randomization if we had al-
lowed the encryption algorithm to be history dependent (as discussed in Section 5.3.1).
Specifically, in such a case, we could have used a counter in the role of r . Further-
more, if the encryption scheme is used for fifo communication between the parties and
both can maintain the counter-value, then there is no need for the sender to send the
counter-value. However, in the latter case, Construction 5.3.3 is preferable (because the
adequate pseudorandom generator may be more efficient than a pseudorandom function
as used in Construction 5.3.9). We note that in case the encryption scheme is not used
for fifo communication and one may need to decrypt messages with arbitrary varying
counter-values, it is typically better to use Construction 5.3.9. Furthermore, in many
cases it may be preferable to select a value (i.e., r ) at random, rather, than rely on a
counter that must be stored in a reliable manner between applications (of the encryption
algorithm).
The ciphertexts produced by Construction 5.3.9 are longer than the corresponding
plaintexts. This is unavoidable in the case of secure (history-independent) encryption
schemes (see Exercise 25). In particular, the common practice of using pseudorandom

17 The v ( j) ’s either equal the x ( j) ’s or the y ( j) ’s, whereas the r ( j) ’s are random (or are fixed by an averaging
argument). The conclusion follows by considering the actual encryptions of the x ( j) ’s and the y ( j) ’s versus
their ideal encryptions. Since the actual encryptions are distinguishable whereas the ideals are not, the actual
encryption of either the x ( j) ’s or the y ( j) ’s must be distinguishable from the corresponding ideal version.

412

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

permutations as block-ciphers18 is not secure (e.g., one can distinguish two encryptions
of the same message from encryptions of two different messages).
Recall that by combining Constructions 5.3.7 and 5.3.9 (and referring to Proposi-
tions 5.3.8 and 5.3.10), we obtain a (full-fledged) private-key encryption scheme. A
more efficient scheme is obtained by a direct combination of the ideas underlying both
constructions:

Construction 5.3.12 (a private-key encryption scheme based on pseudorandom func-

tions): Let F = {Fn } be as in Construction 5.3.9 (that is, F = {Fn } is an efficiently
computable function ensemble) and I and V be the selection and evaluation algo-
rithms associated with it (e.g., V (s, x) = f s (x)). We define a private-key encryption
scheme, (G, E, D), as follows:

Key-generation: G(1n ) = (k, k), where k ← I (1n ).

Encrypting plaintext α ∈ {0, 1}∗ (using the key k): Break α into consecutive blocks of
length n, while possibly augmenting the last block. Let α1 , ..., αt be the resulting
blocks. Associate {0, 1}n with the set of integer residues modulo 2n , select uniformly
r ∈ {0, 1}n , and compute ri = r + i mod 2n , for i = 1, ..., t. Finally, form the ci-
phertext (r, |α|, V (k, r1 ) ⊕ α1 , ..., V (k, rt ) ⊕ αt ). That is,

E k (x) = (r, |α|, f k (r + 1 mod 2n ) ⊕ α1 , ..., f k (r + t mod 2n ) ⊕ αt )

Decrypting ciphertext (r, m, y1 , ..., yt ) (using the key k): For i = 1, ..., t, compute
αi = V (k, (r + i mod 2n )) ⊕ yi , and output the m-bit long prefix of α1 · · · αt . That
is, Dk (r, m, y1 , ..., yt ) is the m-bit long prefix of

(V (k, (r + 1 mod 2n )) ⊕ y1 ) · · · (V (k, (r + t mod 2n )) ⊕ yt )

= ( f k (r + 1 mod 2n ) ⊕ y1 ) · · · ( f k (r + t mod 2n ) ⊕ yt )

Clearly, Construction 5.3.12 constitutes a secure private-key encryption scheme (pro-

vided that F is pseudorandom with respect to polynomial-size circuits). See Exercise 26.

5.3.4. Public-Key Encryption Schemes

As mentioned earlier, randomization during the encryption process can be avoided in
private-key encryption schemes that employ a varying state (not allowed in our basic
Definition 5.1.1). In the case of public-key encryption schemes, randomization during
the encryption process is essential (even if the encryption scheme employs a varying
state). Thus, the randomized encryption paradigm plays an even more pivotal role in the
construction of public-key encryption schemes. To demonstrate this paradigm, we start
with a very simple (and quite wasteful) construction. But before doing so, we recall the
notion of trapdoor permutations.

18 That is, letting E k (x) = pk (x), where pk is the permutation associated with the string k.

413
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Trapdoor permutations. All our constructions employ a collection of trapdoor per-

mutations, as in Definition 2.4.5. Recall that such a collection, { pα }α , comes with four
probabilistic polynomial-time algorithms, denoted here by I, S, F, and B (for index,
sample, forward, and backward ), such that the following (syntactic) conditions hold:
1. On input 1n , algorithm I selects a random n-bit long index α of a permutation pα ,
along with a corresponding trapdoor τ ;
2. On input α, algorithm S samples the domain of pα , returning a random element in
it;
3. For x in the domain of pα , given α and x, algorithm F returns pα (x) (i.e., F(α, x) =
pα (x));
4. For y in the range of pα , if (α, τ ) is a possible output of I (1n ), then given τ and y,
algorithm B returns pα−1 (y) (i.e., B(τ, y) = pα−1 (y)).
The hardness condition refers to the difficulty of inverting pα on a random element of
its range, when given only the range-element and α. That is, let I1 (1n ) denote the first
element in the output of I (1n ) (i.e., the index); then for every polynomial-size circuit
family {Cn }, every polynomial p and all sufficiently large n’s
1
Pr[Cn (I1 (1n ), p I1 (1n ) (S(I1 (1n )))) = S(I1 (1n ))] <
p(n)
Namely, Cn fails to invert pα on pα (x), where α and x are selected by I and S as in the
previous paragraph. Recall that the collection can be easily modified to have a hard-core
predicate (see Theorem 2.5.2). For simplicity, we continue to refer to the collection as
{ pα }, and let b denote the corresponding hard-core predicate.

5.3.4.1. Simple Schemes

We are now ready to present a very simple (alas quite wasteful) construction of a secure
public-key encryption scheme. Actually, we present a block-cipher with block-length
 ≡ 1.

Construction 5.3.13 (a simple public-key block-cipher scheme): Let { pα }, I, S, F, B,

and b be as in the paragraph entitled “trapdoor permutations.”
Key-generation: The key-generation algorithm consists of selecting at random a per-
mutation pα together with a trapdoor τ for it: The permutation (or rather its de-
scription) serves as the public-key, whereas the trapdoor serves as the private-key.
That is, G(1n ) = I (1n ), which means that the index-trapdoor pair generated by I is
associated with the key-pair of G.
Encryption: To encrypt a bit σ, using the encryption-key α, the encryption algorithm
randomly selects an element, r, in the domain of pα , and produces the ciphertext
( pα (r ), σ ⊕ b(r )). That is, E α (σ ) = (F(α, r ), σ ⊕ b(r )), where r ← S(α).
Decryption: To decrypt the ciphertext (y, ς), using the decryption-key τ , the decryption
algorithm just computes ς ⊕ b( pα−1 (y)), where the inverse is computed using the
trapdoor τ of pα . That is, Dτ (y, ς ) = ς ⊕ b(B(τ, y)).
414

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

Clearly, for every possible (α, τ ) output of G and for every σ ∈ {0, 1}, it holds that
Dτ (E α (σ )) = Dτ (F(α, S(α)), σ ⊕ b(S(α)))
= (σ ⊕ b(S(α))) ⊕ b(B(τ, F(α, S(α))))
= σ ⊕ b(S(α)) ⊕ b( pα−1 ( pα (S(α))))
= σ ⊕ b(S(α)) ⊕ b(S(α)) = σ
The security of this public-key encryption scheme follows from the (non-uniform) one-
way feature of the collection { pα } (or rather from the hypothesis that b is a corresponding
hard-core predicate).

Proposition 5.3.14: Suppose that b is a (non-uniformly strong) hard-core of the collec-

tion { pα }. Then Construction 5.3.13 constitutes a secure public-key block-cipher (with
block-length  ≡ 1).

Proof: Recall that by the equivalence theorems (i.e., Theorems 5.2.5 and 5.2.11), it
suffices to show single-message ciphertext-indistinguishability. Furthermore, by the
fact that here there are only two plaintexts (i.e., 0 and 1), it suffices to show that one
cannot distinguish the encryptions of these two plaintexts. That is, all we need to prove
is that, given the encryption-key α, it is infeasible to distinguish E α (0) = ( pα (r ), b(r ))
from E α (1) = ( pα (r ), 1 ⊕ b(r )), where r ← S(α). But this is easily implied by the
hypothesis that b is a hard-core of the collection { pα }. Details follow.
Recall that by saying that b is a hard-core of { pα }, we mean that for every polynomial-
size circuit family {Cn }, every polynomial p and all sufficiently large n’s
1 1
Pr[Cn (I1 (1n ), p I1 (1n ) (S(I1 (1n )))) = b(S(I1 (1n )))] < + (5.9)
2 p(n)
However, analogously to the second proof of Theorem 3.4.1, it can be shown that this
implies that for every polynomial-size circuit family {C n }, every polynomial p  , and all
sufficiently large n’s
1
|Pr[Cn (α, pα (r ), b(r )) = 1] − Pr[Cn (α, pα (r ), 1 ⊕ b(r )) = 1]| <
p (n)
where α ← I1 (1n ) and r ← S(α). Thus, (α, E α (0)) is computationally indistinguishable
from (α, E α (1)), and the proposition follows.
Using Propositions 5.3.8 and 5.3.14, and recalling that Theorem 2.5.2 applies also to
collections of one-way permutations and to the non-uniform setting, we obtain:

Theorem 5.3.15: If there exist collections of (non-uniformly hard) trapdoor permuta-

tions, then there exist secure public-key encryption schemes.

A generalization. As admitted earlier, Construction 5.3.13 is quite wasteful. Specif-

ically, it is wasteful in bandwidth, which is defined to be the relationship between the
length of the plaintext and the length of the ciphertext. In Construction 5.3.13, the
relationship between these lengths equals the security parameter (i.e., the length of
415
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

description of individual elements in the domain of the permutation). However, the

idea underlying Construction 5.3.13 can yield efficient public-key schemes, provided
we use trapdoor permutations having hard-core functions with large range (see Sec-
tion 2.5.3). To demonstrate the point, we use the following assumption relating to the
RSA collection of trapdoor permutations (cf. Subsections 2.4.3 and 2.4.4).
Large Hard-Core Conjecture for RSA: The first n/2 least-significant bits of the argu-
ment constitute a (non-uniformly strong) hard-core function of the RSA function when
applied with n-bit long moduli.
We stress that the conjecture is not known to follow from the assumption that the RSA
collection is (non-uniformly) hard to invert. What is known to hold under the latter as-
sumption is only that the first O(log n) least-significant bits of the argument constitute
a (non-uniformly strong) hard-core function of RSA (with n-bit long moduli). Still, if
the large hard-core conjecture holds, then one obtains a secure public-key encryption
scheme with efficiency comparable to that of “plain RSA” (see the following discussion).
Furthermore, this scheme is related (but not identical) to the common practice of ran-
domly padding messages (using padding equal in length to the message) before encrypt-
ing them (by applying the RSA function).19 That is, we consider the following scheme:

Construction 5.3.16 (Randomized RSA – a public-key block-cipher scheme): This

scheme employs the RSA collection of trapdoor permutations (cf. Subsections 2.4.3
and 2.4.4). The following description is, however, self-contained.

Key-generation: The key-generation algorithm consists of selecting at random two

n-bit primes, P and Q, setting N = P · Q, selecting at random a pair (e, d) such that
e · d ≡ 1 (mod (P − 1) · (Q − 1)), and outputting the pair ((N , e), (N , d)), where
(N , e) is the encryption-key and (N , d) is the decryption-key. That is, ((N , e),
(N , d)) ← G(1n ), where N , e, and d are as specified here.
(Note that N is 2n-bits long.)
Encryption: To encrypt an n-bit string σ (using the encryption-key (N , e)), the encryp-
tion algorithm randomly selects an element r ∈ {0, ..., N − 1}, and produces the
ciphertext (r e mod N , σ ⊕ lsb(r )), where lsb(r ) denotes the n least-significant bits
of r . That is, E (N ,e) (σ ) = (r e mod N , σ ⊕ lsb(r )).
Decryption: To decrypt the ciphertext (y, ς) ∈ {0, ..., N − 1} × {0, 1}n (using the
decryption-key (N , d)), the decryption algorithm just computes ς ⊕ lsb(y d mod N ),
where lsb(·) is as in the Encryption procedure. That is, D(N ,d) (y, ς) = ς ⊕
lsb( y d mod N ).

The bandwidth of this scheme is much better than in Construction 5.3.13: A plaintext of
length n is encrypted via a ciphertext of length 2n + n = 3n. Furthermore, Randomized
RSA is almost as efficient as “plain RSA” (or the RSA function itself).

19 The conjectured security of the common practice relies on a seemingly stronger assumption; that is, the as-
sumption is that for every x ∈ {0, ..., 2n − 1}, given (N , e) as generated in Construction 5.3.16, it is infeasible
to distinguish r e mod N from (x + s2n )e mod N , where r (resp., s) is uniformly distributed in {0, ..., N − 1}
(resp., in {0, ..., N /2n  − 1}).

416

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

To see that Randomized RSA satisfies the syntactic requirements of an encryption

scheme, consider any possible output of G(1n ), denoted ((N , e), (N , d)), and any σ ∈
{0, 1}n . Then, for any r ∈ {0, ..., N − 1}, it holds that

D(N ,d) (E (N ,e) (σ )) = D(N ,d) ((r e mod N ) , σ ⊕ lsb(r ))

= (σ ⊕ lsb(r )) ⊕ lsb((r e mod N )d mod N )
= σ ⊕ lsb(r ) ⊕ lsb(r ed mod N ) = σ

where the last equality is due to r ed ≡ r (mod N ). The security of Randomized RSA
(as a public-key encryption scheme) follows from the large hard-core conjecture for
RSA, analogously to the proof of Proposition 5.3.14.

Proposition 5.3.17: Suppose that the large hard-core conjecture for RSA does hold.
Then Construction 5.3.16 constitutes a secure public-key block-cipher (with block-
length (n) = n).

Proof Sketch: Recall that by the equivalence theorems (i.e., Theorems 5.2.5 and 5.2.11),
it suffices to show single-message ciphertext-indistinguishability. Considering any
two strings x and y, we need to show that ((N , e), r e mod N , x ⊕ lsb(r )) and
((N , e), r e mod N , y ⊕ lsb(r )) are indistinguishable, where N , e and r are selected
at random as in the construction. It suffices to show that for every fixed x, the distribu-
tions ((N , e), r e mod N , x ⊕ lsb(r )) and ((N , e), r e mod N , x ⊕ s) are indistinguish-
able, where s ∈ {0, 1}n is uniformly distributed, independently of anything else. The
latter claim follows from the hypothesis that the n least-significant bits are a hard-core
function for RSA with moduli of length 2n.

Discussion. We wish to stress that encrypting messages by merely applying the RSA
function to them (without randomization) yields an insecure encryption scheme. Un-
fortunately, this procedure (previously referred to as “plain RSA”) is quite common in
practice. The fact that plain RSA is definitely insecure is a special case of the fact that
any public-key encryption scheme that employs a deterministic encryption algorithm
is insecure. We warn that the fact that in such deterministic encryption schemes one
can distinguish encryptions of two specific messages (e.g., the all-zero message and
the all-one message) is not “merely of theoretical concern”; it may seriously endanger
some applications! In contrast, Randomized RSA (as defined in Construction 5.3.16)
may be secure, provided a quite reasonable conjecture (i.e., the large hard-core con-
jecture for RSA) holds. We comment that the more common practice of applying the
RSA function to a randomly padded version of the plaintext is secure if and only if a
seemingly stronger (and yet reasonable) assumption holds; see footnote 19. Thus, the
latter practice is far superior to using the RSA function directly (i.e., without random-
ization): The randomized version is likely to be secure, whereas the non-randomized
(or plain) version is definitely insecure.
We note that Construction 5.3.16 (or, alternatively, Construction 5.3.13) generalizes
to any collection of trapdoor permutations having a corresponding large hard-core
function. Suppose that { pα } is such a collection, and h (or rather {h α }) is a corresponding
hard-core function (resp., a corresponding collection of hard-core functions), such
417
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

that any element in the domain of pα is mapped by h (or h α ) to an (|α|)-bit long

string. Then we can encrypt an (|α|)-bit long plaintext, x, by ( pα (r ), h(r ) ⊕ x) (resp.,
( pα (r ), h α (r ) ⊕ x)), where r ← S(α) (as in Construction 5.3.13). This yields a secure
public-key encryption scheme with bandwidth related to the ratio of (|α|) over the
length of the description of an individual element in the domain of pα .

5.3.4.2. An Alternative Scheme

An alternative construction of a public-key encryption scheme is presented in Con-
struction 5.3.18. Rather than encrypting each plaintext bit (or block of bits) by an
independently selected element in the domain of the trapdoor permutation (as done in
Construction 5.3.13), we select only one such element (for the entire plaintext) and
generate from it additional bits, one per each bit of the plaintext. These additional bits
are determined by successive applications of the trapdoor permutation, and only the
last result is included in the ciphertext. In a sense, the construction of this encryption
scheme augments the construction of a pseudorandom generator based on one-way
permutations (i.e., Construction 3.4.4).

Construction 5.3.18 (a public-key encryption scheme): Let { pα }, I, S, F, B, and b be

−(i+1)
as in Construction 5.3.13. We use the notation pαi+1 (x) = pα ( pαi (x)) and pα (x) =
−1 −i
pα ( pα (x)).
Key-generation: The key-generation algorithm consists of selecting at random a per-
mutation pα together with a trapdoor, exactly as in Construction 5.3.13. That is,
G(1n ) = I (1n ), which means that the index-trapdoor pair generated by I is associ-
ated with the key-pair of G.
Encryption: To encrypt a string σ , using the encryption-key α, the encryption algorithm
randomly selects an element, r , in the domain of pα and produces the ciphertext
(|σ |)
( pα|σ | (r ), σ ⊕ G α (r )), where
def −1
α (r ) = b(r ) · b( pα (r )) · · · b( pα (r ))
G () (5.10)
(|σ |)
That is, E α (σ ) = ( pα|σ | (S(α)), σ ⊕ G α (S(α))).
Decryption: To decrypt the ciphertext (y, ς), using the decryption-key τ , the decryption
(|ς |)
algorithm just computes ς ⊕ G α ( pα−|ς | (y)), where the inverse is computed using
(|ς |)
the trapdoor τ of pα . That is, Dτ (y, ς) = ς ⊕ G α ( pα−|ς | (y)).
We stress that this encryption scheme is a full-fledged one (rather than a block-cipher).
Its bandwidth tends to 1 with the length of the plaintext; that is, a plaintext of length
 = poly(n) is encrypted via a ciphertext of length m + , where m denotes the length
of the description of individual elements in the domain of pα . Clearly, for every possible
(α, τ ) output of G (and r ← S(α)), it holds that
Dτ (E α (σ )) = Dτ ( pα|σ | (r ) , σ ⊕ G (|σ |)
α (r ))
(|σ |)
|) (|σ |) −|σ ⊕G α
= (σ ⊕ G (|σ
α (r )) ⊕ G α ( pα
(r )|
( pα|σ | (r )))
|) (|σ |)
= σ ⊕ G (|σ
α (r ) ⊕ G α (r ) = σ

418

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

The security of this public-key encryption scheme follows from the (non-uniform)
one-way feature of the collection { pα }, but here we restrict the sampling algorithm
S to produce almost uniform distribution over the domain (so that this distribution is
preserved under successive applications of pα ).

Proposition 5.3.19: Suppose that b is a (non-uniformly strong) hard-core of the trap-

door collection { pα }. Furthermore, suppose that this trapdoor collection utilizes a
domain sampling algorithm S so that the statistical difference between S(α) and
the uniform distribution over the domain of pα is negligible in terms of |α|. Then
Construction 5.3.18 constitutes a secure public-key encryption scheme.

Proof: Again, we prove single-message ciphertext-indistinguishability. It suffices

(|σ |)
to show that for every σ , the distributions (α, pα|σ | (S(α)), σ ⊕ G α (S(α))) and
(α, pα|σ | (S(α)), σ ⊕ s) are indistinguishable, where s ∈ {0, 1}|σ | is uniformly dis-
tributed, independently of anything else. The latter claim holds by a minor extension
to Proposition 3.4.6: This proposition refers to the case where S(α) is uniform over the
domain of pα , but can be extended to the case in which there is a negligible statistical
difference between the distributions.

Details: We need to prove that for every polynomial  and every sequence of
def (n)
pairs (σn , σn ) ∈ {0, 1}(n) × {0, 1}(n) , the distributions Dn = (α, pα (S(α)), σn ⊕
def (n)
Dn = σn ⊕
((n)) ((n))
G α (S(α)))and (α, pα (S(α)), G α (S(α))) are indistinguishable,
where α ← I1 (1n ). We prove this in two steps:
def
1. We first prove that for every sequence of σn ’s, the distributions Dn =
(n) ((n)) def (n)
(α, pα (S(α)), σn ⊕ and Rn = (α,
G α (S(α))) σn ⊕ U(n) ) are
pα (S(α)),
indistinguishable, where U(n) denotes a random variable uniformly distributed
over {0, 1}(n) and α ← I1 (1n ).
Suppose first that S(α) is uniform over the domain of pα . Then the indistin-
guishability of {Dn }n∈N and {Rn }n∈N follows directly from Proposition 3.4.6
(as adapted to circuits): The adapted form refers to the indistinguishability of
(n) ((n)) (n)
(α, pα (S(α)), G α (S(α))) and (α, pα (S(α)), U(n) ), and yields the desired
claim by noting that σn can be incorporated in the prospective distinguisher.
The extension (to the case that S(α) has negligible statistical difference to the
uniform distribution over the domain of pα ) is straightforward.
2. Applying the previous item to Dn and Rn = (α, pα(n) (S(α)), σn ⊕ U(n) ), we
def

conclude that {Dn }n∈N and {Rn }n∈N are indistinguishable. Similarly, {Dn }n∈N
 
def (n)
and {Rn }n∈N , where Rn = (α, pα (S(α)), σn ⊕ U(n) ), are indistinguishable.
Furthermore, {Rn }n∈N and {Rn }n∈N are identically distributed. Thus, {Dn }n∈N
and {Dn }n∈N are indistinguishable.

The proposition follows.

An instantiation. Assuming that factoring Blum Integers (i.e., products of two primes
each congruent to 3 (mod 4)) is hard, one may use the modular squaring function
(which induces a permutation over the quadratic residues modulo the product of these
419
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

integers) in the role of the trapdoor permutation used in Construction 5.3.18. This yields
a secure public-key encryption scheme with efficiency comparable to that of plain RSA
(see further discussion latter in this section).

Construction 5.3.20 (the Blum-Goldwasser Public-Key Encryption Scheme): Consult

Appendix A in Volume 1 for the relevant number-theoretic background, and note that
for P ≡ 3 (mod 4) the number (P + 1)/4 is an integer. For simplicity, we present a
block-cipher with arbitrary block-length (n) ≤ poly(n); a full-fledged encryption
scheme can be derived by an easy modification (see Exercise 27 ).

Key-generation: The key-generation algorithm consists of selecting at random two n-

bit primes, P and Q, each congruent to 3 mod 4, and outputting the pair (N , (P, Q)),
where N = P · Q.
Actually, for sake of efficiency, the key-generator also computes

d P = ((P + 1)/4)(n) mod P − 1 (in {0, ..., P − 2})

d Q = ((Q + 1)/4)(n) mod Q − 1 (in {0, ..., Q − 2})
c P = Q · (Q −1 mod P) (in {0, ..., N − Q})
−1
c Q = P · (P mod Q) (in {0, ..., N − P})

It outputs the pair (N , T ), where N serves as the encryption-key and T =

(P, Q, N , c P , d P , c Q , d Q ) serves as decryption-key.
Encryption: To encrypt the message σ ∈ {0, 1}(n) , using the encryption-key N :

1. Uniformly select s0 ∈ {1, ..., N }.

(Note that if GCD(s0 , N ) = 1, then s02 mod N is a uniformly distributed quadratic
residue modulo N .)
2. For i = 1, .., (n) + 1, compute si ← si−12
mod N and bi = lsb(si ), where lsb(s)
is the least-significant bit of s.

The ciphertext is (s(n)+1 , ς), where ς = σ ⊕ b1 b2 · · · b(n) .

Decryption: To decrypt the ciphertext (r, ς) using the decryption-key T =
(P, Q, N , c P , d P , c Q , d Q ), one first retrieves s1 and then computes the bi ’s as in
the Encryption procedure. Instead of successively extracting modular square roots
(n) times, we extract the 2(n) -th root, which can be done as efficiently as extracting
a single square root. Extracting the 2(n) -th root modulo N is done by extracting the
corresponding root modulo P and modulo Q (by raising to power d P modulo P and
d Q modulo Q, respectively) and combining the results via the Chinese Reminder
Theorem:

1. Let s  ← r d P mod P, and s  ← r d Q mod Q.

2. Let s1 ← c P · s  + c Q · s  mod N .
3. For i = 1, ..., (n), compute bi = lsb(si ) and si+1 ← si2 mod N .

The plaintext is ς ⊕ b1 b2 · · · b(n) .

420

www.Ebook777.com
 5.3 CONSTRUCTIONS OF SECURE ENCRYPTION SCHEMES

Again, one can easily verify that this construction constitutes an encryption scheme:
The main fact to verify is that the value of s1 as reconstructed in the decryption stage
equals the value used in the encryption stage. This follows by combining the Chinese
Reminder Theorem with the fact that for every quadratic residue s mod N , it holds that
(n) (n)
s ≡ (s 2 mod N )d P (mod P) and s ≡ (s 2 mod N )d Q (mod Q).
Details: Recall that for a prime P ≡ 3 (mod 4), and every quadratic residue r , we
have r (P+1)/2 ≡ r (mod P). Thus, for every quadratic residue s (modulo N ) and
every , we have

  ((P+1)/4)
(s 2 mod N )d P ≡ s 2 mod N (mod P)

≡ s ((P+1)/2) (mod P)
≡s (mod P)
2
Similarly, (s mod N ) ≡ s (mod Q). Finally, observing that c P and c Q are as
dQ

in the Chinese Reminder Theorem,20 we conclude that s1 as recovered in Step 2

of the decryption process equals s1 as first computed in Step 2 of the encryption
process.

Encryption amounts to (n) + 1 modular multiplications, whereas decryption amounts

to (n) + 2 such multiplications and 2 modular exponentiations (relative to half-sized
moduli). Counting modular exponentiations with respect to n-bit moduli as O(n) (i.e.,
at least n, typically 1.5n, and at most 2n) modular multiplications (with respect to
n-bit moduli), we conclude that the entire encryption–decryption process requires
work comparable to 2(n) + 3n modular multiplications. For comparison to (Ran-
domized) RSA, note that encrypting/decrypting (n)-bit messages (in Randomized
RSA) amounts to (n)/n modular exponentiations, and so the total work is compa-
rable to 2 · ((n)/n) · 1.5n = 3(n) modular multiplications (for general exponent e, or
((n)/n) · (2 + 1.5n) ≈ 1.5(n) modular multiplications in case e = 3).
The security of the Blum-Goldwasser scheme (i.e., Construction 5.3.20) follows
immediately from Proposition 5.3.19 and the fact that the least-significant bit (i.e., lsb)
is a hard-core for the modular squaring function. Recalling that inverting the latter is
computationally equivalent to factoring, we get:

Corollary 5.3.21: Suppose that factoring is infeasible in the sense that for every
polynomial-size circuit {Cn }, every positive polynomial p, and all sufficiently large
n’s
1
Pr[Cn (Pn · Q n ) = Pn ] <
p(n)
where Pn and Q n are uniformly distributed n-bit long primes. Then Construction 5.3.20
constitutes a secure public-key encryption scheme.

Thus, the conjectured infeasibility of factoring (which is a necessary condition for secu-
rity of RSA) yields a secure public-key encryption scheme with efficiency comparable

20 That is, i ≡ c P · (i mod P) + c Q · (i mod Q) (mod N ), for every integer i.

421
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

to that of (plain or Randomized) RSA. In contrast, recall that plain RSA itself is not
secure (as it employs a deterministic encryption algorithm), whereas Randomized RSA
(i.e., Construction 5.3.16) is not known to be secure under a standard assumption such
as intractability of factoring (or even of inverting the RSA function).21

5.4.* Beyond Eavesdropping Security

Our treatment so far has referred only to a “passive” attack in which the adversary
merely eavesdrops on the line over which ciphertexts are being sent. Stronger types
of attacks, culminating in the so-called Chosen Ciphertext Attack, may be possible in
various applications. Specifically, in some settings it is feasible for the adversary to
make the sender encrypt a message of the adversary’s choice, and in some settings the
adversary may even make the receiver decrypt a ciphertext of the adversary’s choice.
This gives rise to chosen plaintext attacks and to chosen ciphertext attacks, respectively,
which are not covered by the security definitions considered in previous sections. Thus,
our main goal in this section is to provide a treatment of such types of “active” attacks.
In addition, we also discuss the related notion of non-malleable encryption schemes
(see Section 5.4.5).

5.4.1. Overview
We start with an overview of the type of attacks and results considered in the current
(rather long) section.

5.4.1.1. Types of Attacks

The following mini-taxonomy of attacks is certainly not exhaustive.

Passive attacks. We first reconsider passive attacks as referred to in the definitions

given in previous sections. In the case of public-key schemes we distinguish two sub-
cases:
1. A key-oblivious, passive attack, as captured in the aforementioned definitions. By
“key-obliviousness” we refer to the postulation that the choice of plaintext does not
depend on the public-key.
2. A key-dependent, passive attack, in which the choice of plaintext may depend on the
public-key.
(In Definition 5.2.2, the choice of plaintext means the random variable X n , whereas in
Definition 5.2.4, it means the pair (xn , yn ). In both these definitions, the choice of the
plaintext is key-oblivious.)

21 Recall that Randomized RSA is secure provided that the n/2 least-significant bits constitute a hard-core function
for n-bit RSA moduli. This is a reasonable conjecture, but it seems stronger than the conjecture that RSA is hard
to invert: Assuming that RSA is hard to invert, we only know that the O(log n) least-significant bits constitute
a hard-core function for n-bit moduli.

422

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

Chosen Plaintext Attacks. Here the attacker may obtain encryptions of plaintexts of
its choice (under the key being attacked). Indeed, such an attack does not add power in
the case of public-key schemes.

Chosen Ciphertext Attacks. Here the attacker may obtain decryptions of ciphertexts
of its choice (under the key being attacked). That is, the attacker is given oracle access
to the decryption function corresponding to the decryption-key in use. We distinguish
two types of such attacks.
1. In an a priori chosen ciphertext attack, the attacker is given access to the decryption
oracle only prior to being presented with the ciphertext that it should attack (i.e., the
ciphertext for which it has to learn partial information). That is, the attack consists
of two stages: In the first stage, the attacker is given the above oracle access, and in
the second stage, the oracle is removed and the attacker is given a “test ciphertext”
(i.e., a test of successful learning).
2. In an a posteriori chosen ciphertext attack, after being given the test ciphertext, the
decryption oracle is not removed, but rather the adversary’s access to this oracle is
restricted in the natural way (i.e., the adversary is allowed to query the oracle on any
ciphertext except for the test ciphertext).
In both cases, the adversary may make queries that do not correspond to a legitimate
ciphertext, and the answer will be accordingly (i.e., a special “failure” symbol). Fur-
thermore, in both cases the adversary may effect the selection of the test ciphertext (by
specifying a distribution from which the corresponding plaintext is to be drawn).
Formal definitions of all these types of attacks are given in the following subsections
(i.e., in Sections 5.4.2, 5.4.3, and 5.4.4, respectively). In addition, in Section 5.4.5,
we consider the related notion of malleability, that is, attacks aimed at generating
encryptions of plaintexts related to the secret plaintext, rather than gaining information
about the latter.

5.4.1.2. Constructions
As in the basic case (i.e., Section 5.3), actively secure private-key encryption schemes
can be constructed based on the existence of one-way functions, whereas actively
secure public-key encryption schemes are based on the existence of (enhanced) trapdoor
permutations. In both cases, withstanding a posteriori chosen ciphertext attacks is harder
than withstanding a priori chosen ciphertext attacks. We will present the following
results.

For Private-Key Schemes. In Section 5.4.4.3, we show that the private-key encryption
scheme based on pseudorandom functions (i.e., Construction 5.3.9) is secure also under
a priori chosen ciphertext attacks, but is not secure under an a posteriori chosen
ciphertext attack. We also show how to transform any passively secure private-key
encryption scheme into a scheme secure under (a posteriori) chosen ciphertext attacks
by using a message-authentication scheme on top of the basic encryption. Thus, the latter
construction relies on message-authentication schemes as defined in Section 6.1. We
423
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

mention that message-authentication schemes can be constructed using pseudorandom

functions; see Section 6.3.

For Public-Key Schemes. Assuming the existence of enhanced trapdoor permutations

(see Section C.1 in Appendix C), we will present constructions of public-key encryption
schemes that are secure against (a priori and a posteriori) chosen ciphertext attacks.
The constructions utilize various forms of non-interactive zero-knowledge proofs (see
Section 4.10 in Volume 1), which can be constructed under the former assumption.
We warn that these constructions, which are presented in Section 5.4.4.4, are rather
complex.
As a corollary to the relation between these strong notions of security and non-
malleable encryption schemes, we will conclude that the schemes withstanding a pos-
teriori chosen ciphertext attacks are non-malleable. For details, see Section 5.4.5.

5.4.1.3. Methodological Comments

As hinted, we do not cover all possible intermediate types of attacks but, rather, focus on
some natural ones. For example, we only consider key-dependent attacks on public-key
encryption schemes (but not on private-key schemes).
The attacks are presented in increasing order of strength; hence, resilience against
such attacks yields increasingly stronger notions of security.22 This fact may be
best verified when considering the indistinguishability variants of these security
definitions.
A uniform-complexity treatment seems more appealing in the current section (i.e.,
more than in the previous sections). However, for the sake of consistency with the
basic definitions (i.e., the previous sections of this chapter), we use non-uniform for-
mulations of the various definitions. In fact, our treatment of the active attacks (i.e.,
in Sections 5.4.3 and 5.4.4) only uses non-uniformity in referring to (non-uniform)
auxiliary inputs, and so non-uniformity can be easily eliminated in that case (i.e., by
just eliminating these auxiliary inputs from all the definitions). (In Section 5.4.2 we
refer to non-uniform families of [polynomial-size] circuits, but also in this case, all re-
sults extend to the uniform-complexity setting [because all the reductions are actually
uniform].)
As mentioned, non-interactive zero-knowledge proofs play a central role in the
construction of public-key encryption schemes that are secure under chosen ciphertext
attacks. Thus, we will assume that the reader is fairly comfortable with the notion
of zero-knowledge proofs. Furthermore, although we recall the relevant definition of
non-interactive zero-knowledge, which will serve as our starting point toward stronger
notions, we recommend that the more basic definitions (and results) regarding non-
interactive zero-knowledge proofs (as presented in Section 4.10) be studied first. In
our constructions of encryption schemes that are secure under a posteriori chosen

22 Indeed, an alternative presentation may start with the strongest notion of security (i.e., corresponding to a-
posteriori chosen ciphertext attacks), and obtain the weaker notions by imposing various restrictions (on the
attacks).

424

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

ciphertext attacks, we will use some results from Chapter 6. In the case of private-key
encryption schemes (treated in Section 5.4.4.3), we will use a message-authentication
scheme, but do so in a self-contained way. In the case of public-key encryption schemes
(treated in Section 5.4.4.4), we will use signature schemes (having an extra property)
in order to construct a certain non-interactive zero-knowledge proof, which we use for
the construction of the encryption scheme. At that point we will refer to a specific result
proved in Chapter 6.

5.4.2. Key-Dependent Passive Attacks

The following discussion, as well as the entire subsection, refers only to public-key
encryption schemes. For sake of simplicity, we present the single-message definitions
of security. We note that, as in the basic case (for public-key encryption schemes), the
single-message definitions of security are equivalent to the multiple-message ones.
In Definitions 5.2.2 and 5.2.4, the plaintext distribution (or pair) is fixed obliviously
of the encryption-key. This suffices for the natural case in which the (high-level) appli-
cation (using the encryption scheme) is oblivious of the encryption-key.23 However, in
some settings, the adversary may have partial control on the application. Furthermore,
in the public-key case, the adversary knows the encryption-key in use, and so (if it may
partially control the application then) it may be able to cause the application to invoke
the encryption scheme on plaintexts that are related to the encryption-key in use. Thus,
for such settings, we need stronger definitions of security that postulate that partial
information about the plaintext remains secret even if the plaintext does depend on the
encryption-key in use. Note that here we merely consider the dependence of the “test”
plaintext (i.e., the one for which the adversary wishes to obtain partial information)
on the encryption-key, and ignore the fact that the foregoing motivation also suggests
that the adversary can obtain the encryptions of additional plaintexts chosen by it (as
discussed in Section 5.4.3). However, it is easy to see that (in the public-key setting
discussed here) these additional encryptions are of no use because the adversary can
generate them by itself (see Section 5.4.3).

5.4.2.1. Definitions
Recall that we seek a definition that guarantees that partial information about the plain-
text remains secret even if the plaintext does depend on the encryption-key in use. That
is, we seek a strengthening of semantic security (as defined in Definition 5.2.2) in which
one allows the plaintext distribution ensemble (denoted {X n }n∈N in Definition 5.2.2)
to depend on the encryption-key in use (i.e., for encryption-key e, we consider the
distribution X e over {0, 1}poly(|e|) ). Furthermore, we also allow the partial information
functions (denoted f and h in Definition 5.2.2) to depend on the encryption-key in use
(i.e., for encryption-key e, we consider the functions f e and h e ). In the actual definition

23 Indeed, it is natural (and even methodologically imperative) that a high-level application that uses encryption as
a tool be oblivious of the keys used by that tool. However, this refers only to a proper operation of the application,
and deviation may be caused (in some settings) by an improper behavior (i.e., an adversary).

425
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

it is important to restrict the scope of the functions {h e }e and the distributions {X e }e

so that their dependency on e is polynomial-time computable (see Exercise 28). This
yields the definition presented in Exercise 29, which is equivalent to the following
formulation.24

Definition 5.4.1 (semantic security under key-dependent passive attacks): The se-
quence {( f e , h e , X e )}e∈{0,1}∗ is admissible for the current definition if

1. The functions f e : {0, 1}∗ → {0, 1}∗ are polynomially bounded; that is, there exists
a polynomial  such that | f e (x)| ≤ (|x| + |e|).
2. There exists a non-uniform family of polynomial-size (h-evaluation) circuits {Hn }n∈N
such that for every e in the range of G 1 (1n ) and every x in the support of X e , it holds
that Hn (e, x) = h e (x).
3. There exists a non-uniform family of (probabilistic) polynomial-size (sampling) cir-
cuits {Sn }n∈N such that for every e in the range of G 1 (1n ) and for some m = poly(|e|),
the random variables Sn (e, Um ) and X e are identically distributed.25

An encryption scheme, (G, E, D), is semantically secure under key-dependent pas-

sive attacks if for every probabilistic polynomial-time algorithm A, there exists a
probabilistic polynomial-time algorithm A such that for every admissible sequence
{( f e , h e , X e )}e∈{0,1}∗ , every positive polynomial p, and all sufficiently large n it holds
that


Pr A(e, E e (X e ), 1|X e | , h e (X e )) = f e (X e )

 1
< Pr A (e, 1|X e | , h e (X e )) = f e (X e ) +
p(n)
where (e, d) ← G(1n ), and the probability is taken over the internal coin tosses of
algorithms G, E, A, and A , as well as over X e .

We stress that the performance of A is measured against the same distribution of

triplets ( f e , h e , X e ) (i.e., e ← G 1 (1n )) as the one considered for algorithm A. Unlike
in other versions of the definition of semantic security, here it is important to let A
have the encryption-key e because the task (i.e., the evaluation of f e (X e )) as well as its
main input (i.e., the value h e (X e )) are related to e. (Indeed, if e were not given to A ,
then no encryption scheme (G, E, D) could have satisfied the revised Definition 5.4.1:
Considering h e (x) = x ⊕ e (for |x| = |e|) and f e (x) = x, note that it is easy for A to
compute x from e and h e (x), which are explicit in (e, E e (x), 1|x| , h e (x)), whereas no
A can compute x from (1n , 1|x| , h e (x)).)
Using Exercise 14.2, one may verify that Definition 5.2.2 is a special case of Def-
inition 5.4.1. An analogous modification (or generalization) of Definition 5.2.4 yields
the following:

24 Recall that without loss of generality, we may assume that the keys generated by G(1n ) have length n. Thus,
there is no point in providing the algorithms with 1n as an auxiliary input (as done in Definition 5.2.2).
25 As usual, S (e, r ) denotes the output of the circuit S on input e and coins r . We stress that for every e, the
n n
length of X e is fixed.

426

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

Definition 5.4.2 (indistinguishability of encryptions under key-dependent passive at-

tacks): The sequence {(x e , ye )}e∈{0,1}∗ is admissible for the current definition if there ex-
ists a non-uniform family of polynomial-size circuits {Pn }n∈N that maps each encryption-
key e ∈ {0, 1}∗ to the corresponding pair of (equal-length) strings (x e , ye ). That is, for
every e in the range of G 1 (1n ), it holds that Pn (e) = (xe , ye ). An encryption scheme,
(G, E, D), has indistinguishable encryptions under key-dependent passive attacks
if for every non-uniform family of polynomial-size circuits {Cn }, every admissible se-
quence {(x e , ye )}e∈{0,1}∗ , every positive polynomial p, and all sufficiently large n it holds
that
1
| Pr [Cn (e, E e (xe )) = 1] − Pr [Cn (e, E e (ye )) = 1] | <
p(n)
where (e, d) ← G(1n ), and the probability is taken over the internal coin tosses of
algorithms G and E.

As in the basic case (i.e., Section 5.2), the two definitions are equivalent.

Theorem 5.4.3 (equivalence of definitions for key-dependent passive attacks): A

public-key encryption scheme (G, E, D) is semantically secure under key-dependent
passive attacks if and only if it has indistinguishable encryptions under key-dependent
passive attacks.

Proof Sketch: In order to show that indistinguishability of encryptions implies semantic

security, we follow the proof of Proposition 5.2.6. Specifically, A is constructed and
analyzed almost as before, with the exception that A gets and uses the encryption-
key e (rather than letting it generate a random encryption-key by itself).26 That is, we
let A (e, 1|x| , h e (x)) = A(e, E e (1|x| ), 1|x| , h e (x)), and show that for all (deterministic)
polynomial-size circuit families {Sn }n∈N and {Hn }n∈N it holds that
 

Pr A(e, E e (Sn (e)), 1|Sn (e)| , Hn (e, Sn (e))) = f e (Sn (e)) (5.11)
  
 1
< Pr A(e, E e (1|Sn (e)| ), 1|Sn (e)| , Hn (e, Sn (e))) = f e (Sn (e)) +
poly(n)
where e ← G 1 (1n ) and µ : N →[0,1] is a negligible function. Once established,
Eq. (5.11) implies that (G, E, D) satisfies Definition 5.4.1.
On how Eq. (5.11) implies Definition 5.4.1: The issue is that Eq. (5.11) refers to
deterministic plaintext-selecting circuits (i.e., the Sn ’s), whereas Definition 5.4.1
refers to probabilistic plaintext-sampling circuits (i.e., the Sn ’s). This small gap
can be bridged by fixing a sequence of coins for the latter probabilistic (sam-
pling) circuits. Specifically, starting with any admissible (for Definition 5.4.1) se-
quence {( f e , h e , X e )}e∈{0,1}∗ , where Hn (e, x) = h e (x) and X e ≡ Sn (e, Upoly(n) ), we
consider some sequence of coins rn (for Sn ) that maximizes the gap between
Pr[A(e, E e (x e ), 1|xe | , Hn (e, x e )) = f e (x e )] and Pr[A (e, 1|xe | , Hn (e, x e )) = f e (x e )],

26 Here we use the convention by which A gets e along with h e (x) (and 1|x| ). This is important because A must
feed a matching pair (e, h e (x)) to A.

427
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

where e is random and x e = Sn (e, rn ). Recalling that A (e, 1 , γ ) =

A(e, Ee (1 ), 1 , γ ) and incorporating the sequence of rn ’s in A, we obtain a contra-
diction to Eq. (5.11) (i.e., by letting Sn (e) = Sn (e, rn ) = xe ).

Assuming (to the contrary of the above claim) that Eq. (5.11) does not hold, we obtain
a sequence of admissible pairs {(xe , ye )}e∈{0,1}∗ for Definition 5.4.2 such that their
encryptions can be distinguished (in contradiction to our hypothesis). Specifically,
def def def
we set x e = Sn (e) and ye = 1|xe | , and let C n (e, α) = A(e, α, 1|xe | , Hn (e, x e )). Thus, we
obtain a (poly(n)-size) circuit Cn such that for some positive polynomial p and infinitely
many n’s
  1
Pr[C  (e, E e (x e )) = f e (xe )] − Pr[C  (e, E e (ye )) = f e (xe )] >
n n
p(n)

where e is distributed according to G 1 (1n ). Using an idea as in the proof of Theo-

rem 5.2.15, we derive a (poly(n)-size) circuit Cn that distinguishes (e, E e (xe )) from
(e, E e (ye )), where e ← G 1 (1n ), in contradiction to our hypothesis.

Details: We refer to the proof of Claim 5.2.15.1 (contained in the proof of The-
orem 5.2.15). Recall that the idea was to proceed in two stages. First, using
only
 e (which also yields xe and ye ), we find  an arbitrary value v such that
Pr[C  (e, E e (x e )) = v] − Pr[C  (e, E e (ye )) = v] is large. In the second stage, we
n n
use this value v in order to distinguish the case in which we are given an encryption
of x e from the case in which we are given an encryption of ye . (We comment if
(e, x) → f e (x) were computable by a poly(n)-size circuit, then converting Cn into a
distinguisher Cn would have been much easier; we further comment that as a corol-
lary to the current proof, one can conclude that the restricted form is equivalent to
the general one.)

This concludes the proof that indistinguishability of encryptions (as per Definition 5.4.2)
implies semantic security (as per Definition 5.4.1), and we now turn to the opposite
direction.
Suppose that (G, E, D) does not have indistinguishable encryptions, and consider an
admissible sequence {(xe , ye )}e∈{0,1}∗ that witnesses this failure. Following the proof of
Proposition 5.2.7, we define a probability ensemble {X e }e∈{0,1}∗ and function ensembles
{h e }e∈{0,1}∗ and { f e }e∈{0,1}∗ in an analogous manner:

r The distribution X e is uniformly distributed over {xe , ye }.

r The function f e satisfies f e (x e ) = 1 and f e (ye ) = 0.
r The function h e is defined such that h e (X e ) equals the description of the circuit Cn that
distinguishes (e, E e (xe )) from (e, E e (ye )), where e ← G 1 (1n ) (and (xe , ye ) = Pn (e)).

Using the admissibility of the sequence {(x e , ye )}e (for Definition 5.4.2), it follows that
{( f e , h e , X e )}e is admissible for Definition 5.4.1. Using the same algorithm A as in the
proof of Proposition 5.2.7 (i.e., A(e, β, Cn ) = Cn (e, β), where β is a ciphertext and
Cn = h e (X e )), and using the same analysis, we derive a contradiction to the hypothesis
that (G, E, D) satisfies Definition 5.4.1.
428

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

Details: Without loss of generality, suppose that

1
Pr [C n (e, E e (x e )) = 1] > Pr [Cn (e, E e (ye )) = 1] +
p(n)
for e ← G 1 (1n ). Then, as shown in Claim 5.2.7.1,
1 1
Pr [A(e, E e (X e ), h e (X e )) = f e (X e )] > +
2 2 p(n)
On the other hand, as shown in Fact 5.2.7.2, for every algorithm A

 1
Pr A (e, 1|X e | , h e (X e )) = f e (X e ) ≤
2
because (e, 1|X e | , h e (X e )) contains no information about the value of f e (X e ) (which
is uniformly distributed in {0, 1}). This violates Definition 5.4.1, and so our initial
contradiction hypothesis (i.e., that one can distinguish encryptions under (G, E, D))
must be false.

The theorem follows.

Multiple-Message Security. Definitions 5.4.1 and 5.4.2 can be easily generalized to

handle the encryption of many messages (as in Section 5.2.4), yielding again two
equivalent definitions. Since we are in the public-key setting, one can show (analo-
gously to Theorem 5.2.11) that the single-message definitions of security are equiv-
alent to the multiple-message ones (i.e., by showing that Definition 5.4.2 implies its
multiple-message generalization). One important observation is that admissibility for
the multiple-message definition enables one to carry out a hybrid argument (as in the
proof of Theorem 5.2.11). For details, see Exercise 31. The bottom-line is that we can
freely use any of the four security definitions for key-dependent passive attacks, and
security under that definition implies security under any of the other definitions.

5.4.2.2. Constructions
All the results presented in Section 5.3.4 extend to security under key-dependent passive
attacks. That is, for each of the constructions presented in Section 5.3.4, the same
assumption used to prove security under key-oblivious passive attacks actually suffices
for proving security under key-dependent passive attacks. Before demonstrating this
fact, we comment that (in general) security under key-oblivious passive attacks does
not necessarily imply security under key-dependent passive attacks; see Exercise 32.

Initial observations. We start by observing that Construction 5.3.7 (i.e., the transfor-
mation of block-ciphers to general encryption schemes) maintains its security in our
context. That is:

Proposition 5.4.4: (extension of Proposition 5.3.8): Let (G, E, D) and (G  , E  , D  )

be as in Construction 5.3.7; that is, let (G  , E  , D  ) be the full-fledged encryption
constructed based on the block-cipher (G, E, D). Then if (G, E, D) is secure under
key-dependent passive attacks, then so is (G  , E  , D  ).
429
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Proof Idea: As in the proof of Proposition 5.3.8, we merely observe that multiple-
message security of (G  , E  , D  ) is equivalent to multiple-message security of
(G, E, D).

We next observe that Construction 5.3.13 (a block-cipher with block-length  ≡ 1)

maintains its security also under a key-dependent passive attack. This is a special case
of the following observation:

Proposition 5.4.5: Let (G, E, D) be a block-cipher with logarithmically bounded

block-length (i.e., (n) = O(log n)). If (G, E, D) is secure under key-oblivious pas-
sive attacks, then it is also secure under key-dependent passive attacks.

Proof Sketch: Here we use the definition of ciphertext-indistinguishability in the single-

message setting. The key observation is that the set of possible messages is relatively
small, and so selecting a message in a key-dependent manner does not give much
advantage over selecting a message at random (i.e., obliviously of the key).
Consider an arbitrary admissible (for Definition 5.4.2) set of pairs, {(xe , ye )}e∈{0,1}∗ ,
where |xe | = |ye | = O(log |e|), and a circuit family {Cn } that tries to distinguish
(e, E e (xe )) from (e, E e (ye )). We shall show that {Cn } necessarily fails by relating its
distinguishing gap to the distinguishing gap of a key-oblivious attack (represented in
x, y
the next paragraph by the Cn ’s).
Let {Pn }n∈N be the circuit family producing the aforementioned admissible set (i.e.,
Pn (e) = (x e , ye )). Fixing some n ∈ N and an arbitrary (x, y) ∈ {0, 1}∗ × {0, 1}∗ , we
x, y
consider a circuit Cn (depending on the circuits Cn and Pn and the pair (x, y)) that,
on input (e, α), operates as follows:
x, y
1. Using the hard-wired circuit Pn and the input (key) e, the circuit Cn checks whether
x, y
(xe , ye ) equals the hard-wired pair (x, y) (i.e., C n checks whether Pn (e) = (x, y)).
x, y
In case the check fails, Cn outputs an arbitrary value (e.g., 1) obliviously of the
ciphertext α.
x, y
2. Otherwise (i.e., Pn (e) = (x, y)), the circuit C n invokes C n on its own input and
answers accordingly (i.e., outputs C n (e, α)).

Since (G, E, D) is secure under key-oblivious passive attacks, it follows that (for every
x, y
(x, y) ∈ {0, 1}m × {0, 1}m , where m ≤ poly(n)) the circuit C n cannot distinguish the
case α = E e (x) from the case α = E e (y). Thus, for some negligible function µ : N →
[0,1] and every pair (x, y) ∈ {0, 1}m × {0, 1}m , the following holds:
 
µ(n) > Pre [Cnx, y (e, E e (x)) = 1] − Pre [Cnx, y (e, E e (y)) = 1]
 
 Cn (e, E e (xe )) = 1 Cn (e, E e (ye )) = 1 
= Pre − Pre 

∧ (xe , ye ) = (x, y) ∧ (xe , ye ) = (x, y)

where e ← G 1 (1n ), and equality holds because in case (xe , ye ) = (x, y), the output of
x, y x, y x, y
C n (e, α) is independent of α (and so in this case Cn (e, E e (x)) = Cn (e, E e (y))).
Since this holds for any pair (x, y) ∈ {0, 1}m × {0, 1}m , and since |xe | = |ye | = (n), it
430

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

follows that
|Pre [Cn (e, E e (xe )) = 1] − Pre [Cn (e, E e (ye )) = 1]|
 
  
≤ Pre C n (e, E e (x e )) = 1 − Pre Cn (e, E e (ye )) = 1 
 ∧ (x e , ye ) = (x, y) ∧ (xe , ye ) = (x, y) 
|x|=|y|=(n)

< 22(n) · µ(n)

and the proposition follows (because (n) = O(log n)).

A Feasibility Result. Combining Theorem 5.3.15 with Propositions 5.4.4 and 5.4.5,
we obtain a feasibility result:

Theorem 5.4.6: If there exist collections of (non-uniformly hard) trapdoor permu-

tations, then there exist public-key encryption schemes that are secure under key-
dependent passive attacks.

More Efficient Schemes. In order to obtain more efficient schemes, we directly analyze
the efficient constructions presented in Section 5.3.4. For example, extending the proof
of Proposition 5.3.19, we obtain:

Proposition 5.4.7: Suppose that b is a (non-uniformly strong) hard-core of the trapdoor

collection { pα }. Furthermore, suppose that this trapdoor collection utilizes a domain
sampling algorithm S so that the statistical difference between S(α) and the uniform
distribution over the domain of pα is negligible in terms of |α|. Then Construction 5.3.18
constitutes a public-key encryption scheme that is secure under key-dependent passive
attacks.

Proof Sketch: Again, we prove single-message ciphertext-indistinguishability. We rely

heavily on the admissibility condition. In analogy to the proof of Proposition 5.3.19,
it suffices to show that for every polynomial-size circuit family {Cn }, the distribu-
tions (α, pα (S(α)), Cn (α) ⊕ G α (S(α))) and (α, pα (S(α)), Cn (α) ⊕ U ) are indistin-
()

guishable, for a randomly generated (encryption-key) α, where  = |Cn (α)| and U is

uniformly distributed (independently of anything else).27 Incorporating {Cn } in the po-
tential distinguisher, it suffices to show that the distributions (α, pα (S(α)), G α (S(α)))
()

and (α, pα (S(α)), U ) are indistinguishable. The latter claim follows as in the proof of
Proposition 5.3.19 (i.e., by a minor extension to Proposition 3.4.6). The proposition
follows.

5.4.3. Chosen Plaintext Attack

So far, we have discussed only passive attacks (in two variants: key-oblivious versus
key-dependent, discussed in Section 5.2 and 5.4.2, respectively). Turning to active

27 Recall that here α serves as an encryption-key and C n (α) is a key-dependent plaintext. Typically, Cn (α) would
be the first or second element in the plaintext pair (xα , yα ) = Pn (α).

431
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

attacks, we start with mild active attacks in which the adversary may obtain (from some
legitimate user) ciphertexts corresponding to plaintexts of the adversary’s choice. Such
attacks will be called chosen plaintext attacks, and they characterize the adversary’s
abilities in some applications. For example, in some settings, the adversary may (directly
or indirectly) control the encrypting module (but not the decrypting module).
Intuitively, a chosen plaintext attack poses additional threat in the case of private-
key encryption schemes (see Exercise 33), but not in the case of public-key encryption
schemes. In fact, we will show that in the case of public-key encryption schemes, a
chosen plaintext attack can be emulated by a passive key-dependent attack.

5.4.3.1. Definitions
We start by rigorously formulating the framework of chosen plaintext attacks. Intu-
itively, such attacks proceed in four stages corresponding to the generation of a key (by
a legitimate party), the adversary’s requests (answered by the legitimate party) to encrypt
plaintexts under this key, the generation of a challenge ciphertext (under this key and
according to a template specified by the adversary), and additional requests to encrypt
plaintexts (under the same key). That is, a chosen plaintext attack proceeds as follows:

1. Key generation: A key-pair (e, d) ← G(1n ) is generated (by a legitimate party). In

the public-key setting the adversary is given (1n , e), whereas in the private-key setting
the adversary is only given 1n . Actually, assuming (without loss of generality) that
|e| = n, we may replace (1n , e) by e in the former case.
2. Encryption requests: Based on the information obtained so far, the adversary may
request (the legitimate party) to encrypt plaintexts of its (i.e., the adversary’s) choice.
A request to encrypt the plaintext x is answered with a value taken from the distribu-
tion E e (x), where e is as determined in Step 1. After making several such requests,
the adversary moves to the next stage.
3. Challenge generation: Based on the information obtained so far, the adversary spec-
ifies a challenge template and is given an actual challenge.
When defining semantic security, the challenge template is a triplet of circuits
(Sm , h m , f m ), where Sm specifies a distribution of m-bit long plaintexts (and
h m , f m : {0, 1}m → {0, 1}∗ ), and the actual challenge is a pair (E e (x), h m (x)) where
x is distributed according to Sm (Upoly(n) ). When defining indistinguishability of en-
cryptions, the challenge template is merely a pair of equal-length strings, and the
actual challenge is an encryption of one of these two strings.
4. Additional encryption requests: Based on the information obtained so far, the ad-
versary may request the encryptions of additional plaintexts of its choice. These
requests are handled as in Step 2. After making several such requests, the adversary
produces an output and halts.
In the actual definition, the adversary’s strategy will be decoupled into two parts cor-
responding to its actions before and after the generation of the actual challenge. Each
part will be represented by a (probabilistic polynomial-time) oracle machine, where
the oracle is an “encryption oracle” (with respect to the key generated in Step 1). The
432

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

first part, denoted A1 , represents the adversary’s behavior during Step 2. It is given
a security parameter (and possibly an encryption-key), and its output is a pair (τ, σ ),
where τ is the template generated in the beginning of Step 3 and σ is state information
passed to the second part of the adversary. The second part of the adversary, denoted
A2 , represents the adversary’s behavior during Step 4. It is given the state σ (of the first
part), as well as the actual challenge (generated Step 3), and produces the actual output
of the adversary.
In accordance with the use of non-uniform formulations, we let each of the two
oracle machines have a (non-uniform) auxiliary input. In fact, it suffices to provide
only the first machine with such a (non-uniform) auxiliary input, because it can pass
auxiliary input to the second machine in the state information σ . (Similarly, in the case
of public-key schemes, it suffices to provide only the first machine with the encryption-
key.) We comment that we provide these machines with probabilistic oracles; that is, in
response to a plaintext query x, the oracle E e returns a random ciphertext E e (x) (i.e.,
the result of a probabilistic process applied to e and x). Thus, in the case of public-key
schemes, the four-step attack process can be written as follows:
(e, d) ← G(1n )
(τ, σ ) ← A1E e (e, z)
def
c = an actual challenge generated according to the template τ
output ← A2E e (σ, c)
where z denotes (non-uniform) auxiliary input given to the adversary. In the case of
private-key schemes, the adversary (i.e., A1 ) is given 1n instead of e.

Semantic Security. Instantiating this framework to derive a definition of semantic

security amounts to specifying the challenge generation and to postulating that the
success probability in such an attack should be met by a corresponding benign process.
As hinted in the preceding discussion, the challenge generation consists of the adversary
specifying a triplet of circuits, denoted (Sm , h m , f m ), and being presented with an
encryption of x ← Sm (Upoly(n) ) ∈ {0, 1}m along with the partial information h m (x).
The adversary’s goal is to guess f m (x), and semantic security amounts to saying that
the adversary’s success probability can be matched by a corresponding algorithm that
is only given h m (x) and 1|x| = 1m . Like the adversary, the corresponding algorithm is
decoupled into two parts; the first is in charge of outputting a challenge template, and
the second is in charge of solving the challenge (without being given a ciphertext),
where state information is passed from the first part to the second part. It is important
to require that the challenge template produced by the corresponding algorithm be
distributed exactly as the challenge template produced by the adversary. (See further
discussion following Definition 5.4.8.)

Definition 5.4.8 (semantic security under chosen plaintext attacks):

For public-key schemes: A public-key encryption scheme, (G, E, D), is said to be se-
mantically secure under chosen plaintext attacks if for every pair of probabilistic

433
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

polynomial-time oracle machines, A1 and A2 , there exists a pair of probabilistic

polynomial-time algorithms, A1 and A2 , such that the following two conditions hold:
1. For every positive polynomial p, and all sufficiently large n and z ∈ {0, 1}poly(n)
it holds that
⎡ ⎤
v = f m (x) where
⎢ (e, d) ← G(1n ) ⎥
⎢ ⎥
⎢
Pr ⎢
Ee
((Sm , h m , f m ), σ ) ← A1 (e, z) ⎥
⎥
⎣ c ← (E e (x), h m (x)) , where x ← Sm (Upoly(n) ) ⎦
v ← A2E e (σ, c)
⎡ ⎤
v = f m (x) where
⎢ ((Sm , h m , f m ), σ ) ← A1 (1n , z) ⎥
< Pr ⎢ ⎥+ 1
⎣ x ← Sm (Upoly(n) ) ⎦ p(n)
 |x|
v ← A2 (σ, 1 , h m (x))
Recall that (Sm , h m , f m ) is a triplet of circuits produced as in Step 3 of the
foregoing description, and that x is a sample from the distribution induced by Sm .
2. For every n and z, the first elements (i.e., the (Sm , h m , f m ) part) in the random
E G (1n )
variables A1 (1n , z) and A1 1 (G 1 (1n ), z) are identically distributed.
For private-key schemes: The definition is identical except that algorithm A1 gets the
security parameter 1n instead of the encryption-key e.

Note that as in almost all other definitions of semantic security (with the exception of
Definition 5.4.1), algorithm A1 does not get a (random) encryption-key as input (but may
rather generate one by itself).28 Since the challenge template is not fixed (or determined
by e) but, rather, is chosen by A and A themselves, it is very important to require
that in both cases, the challenge template be distributed identically (or approximately
so): There is no point in relating the success probability of A and A , unless these
probabilities refer to same distribution of problems (i.e., challenge templates).29 (The
issue arises also in Definition 5.4.1 where it was resolved by forcing A to refer to the
challenge template determined by the public-key e.)30
Definition 5.4.8 implies Definition 5.4.1, but this may not be evident from the def-
initions themselves (most importantly, because here f m is computationally bounded
whereas in Definition 5.4.1 the function is computationally unbounded). Still, the va-
lidity of the claim follows easily from the equivalence of the two definitions to the

28 In fact, A1 is likely to start by generating e ← G 1 (1n ), because it has to generate a challenge template that is
distributed as the one produced by A1 on input e ← G 1 (1n ).
29 Failure to make this requirement would have resulted in a fundamentally bad definition (by which every encryp-

tion scheme is secure). For example, algorithm A1 could have set h m to equal the function f m selected by A1 (in
a corresponding attack). Doing so, the success of A to guess the value of fm (x) from the (insecure) encryption
of x and a (possibly) useless value h m (x) (e.g., for a constant function h m ) would have been met by the success
of A to “guess” the value of f m (x) from f m (x) itself (without being given the encryption of x). An alternative
approach, which follows the formulation of Definition 5.4.1, is presented in Exercise 34.
30 Indeed, an alternative solution could have been the one adopted here and in the sequel; that is, in an alternative

to Definition 5.4.1, one may allow A to select the challenge template by itself, provided that the selection yields
a distribution similar to the one faced by A (as induced by the public-key e). For details, see Exercise 30.

434

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

corresponding notions of indistinguishability of encryptions (and the fact that the im-
plication is evident for the latter formulations).

Indistinguishability of Encryptions. Deriving the corresponding definition of indis-

tinguishability of encryptions (from the previous framework) is considerably simpler.
Here, the challenge generation consists of the adversary specifying two equal-length
strings and the adversary being presented with the encryption of one of them. The
adversary’s goal is to distinguish the two possible cases.

Definition 5.4.9 (indistinguishability of encryptions under chosen plaintext attacks):

For public-key schemes: A public-key encryption scheme, (G, E, D), is said to have
indistinguishable encryptions under chosen plaintext attacks if for every pair
of probabilistic polynomial-time oracle machines, A1 and A2 , for every positive
polynomial p, and for all sufficiently large n and z ∈ {0, 1}poly(n) it holds that
1
| pn,z
(1)
− pn,z
(2)
| <
p(n)
where
⎡ ⎤
v = 1 where
⎢ (e, d) ← G(1n ) ⎥
⎢ ⎥
(i)
pn,z
def ⎢
= Pr ⎢ ((x , x ), σ ) ← A1 (e, z) ⎥
(1) (2) Ee
⎥
⎣ c ← E e (x (i) ) ⎦
Ee
v ← A2 (σ, c)
where |x (1) | = |x (2) |.
For private-key schemes: The definition is identical except that A1 gets the security
parameter 1n instead of the encryption-key e.

Clearly, Definition 5.4.9 implies Definition 5.4.2 as a special case. Furthermore, for
public-key schemes, the two definitions are equivalent (see Proposition 5.4.10), whereas
for private-key schemes, Definition 5.4.9 is strictly stronger (see Exercise 33).

Proposition 5.4.10: Let (G, E, D) be a public-key encryption scheme that has indis-
tinguishable encryptions under key-dependent passive attacks. Then (G, E, D) has
indistinguishable encryptions under chosen plaintext attack.

Proof Sketch: The key observation is that in the public-key model, a chosen plaintext
attack can be emulated by a passive key-dependent attack. Specifically, the (passive)
attacker can emulate access to an encryption oracle by itself (by using the encryption-
key given to it). Thus, we obtain an attacker as in Definition 5.4.9, with the important
exception that it never makes oracle calls (but rather emulates E e by itself). In other
words, we have an attacker as in Definition 5.4.2, with the minor exception that it is
a probabilistic polynomial-time machine with auxiliary input z (rather than being a
polynomial-size circuit) and that it distinguishes a pair of plaintext distributions rather
than a pair of (fixed) plaintexts (which depend on the encryption-key). However, fixing
435
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

the best-possible coins for this attacker (and incorporating them as well as z in an
adequate circuit), we obtain an attacker exactly as in Definition 5.4.2 such that its
distinguishing gap is at least as large as the one of the (initial) chosen plaintext attacker.
(For details, see Exercise 30.)

Equivalence of Semantic Security and Ciphertext-Indistinguishability. As in pre-

vious cases, we show that the two formulations of (chosen plaintext attack) security
(i.e., semantic security and indistinguishability of encryptions) are in fact equivalent.

Theorem 5.4.11 (equivalence of definitions for chosen plaintext attacks): A public-key

(resp., private-key) encryption scheme (G, E, D) is semantically secure under cho-
sen plaintext attacks if and only if it has indistinguishable encryptions under chosen
plaintext attacks.

Proof Sketch: In order to show that indistinguishabity of encryptions implies semantic

security, we follow again the ideas underlying the proof of Proposition 5.2.6. Specif-
ically, for both the private-key and public-key cases, A1 and A2 are constructed as
follows:
def
1. A1 (1n , z) = (τ, σ  ), where (τ, σ  ) is generated as follows:
First, A1 generates an instance of the encryption scheme; that is, A1 lets (e, d) ←
G(1n ). Next, A1 invokes A1 , while emulating the oracle E e , and sets (τ, σ ) ←
def
A1Ee (1n , z). Finally, A1 sets σ  = (e, σ ).
We warn that the generation of the key-pair by A1 should not be confused with the
generation of the key-pair in the probabilistic experiment referring to the combined
algorithm A = (A1 , A2 ). In particular, the generated encryption-key e allows A1 to
emulate the encryption oracle Ee (also in the private-key case). Furthermore, A1
outputs the encryption-key e as part of the state passed by it to A2 , whereas A1 does
not necessarily do so (and, in fact, cannot do so in the case of the private-key model).
This will allow A2 , too, to emulate the encryption oracle E e .
def
2. A2 ((e, σ ), 1m , γ ) = A2E e (σ, (E e (1m ), γ )), where typically γ = h m (x) and m = |x|.

Since A1 merely emulates the generation of a key-pair and the actions of A1 with respect
to such a pair, the equal distribution condition (i.e., Item 2 in Definition 5.4.8) holds.
Using the (corresponding) indistinguishability of encryption hypothesis, we show that
(even in the presence of an encryption oracle E e ) the distributions (σ, (E e (x), h(x)))
and (σ, (E e (1|x| ), h(x))) are indistinguishable, where (e, d) ← G(1n ), ((S, h, f ), σ ) ←
A1Ee (y, z) (with y = e or y = 1n depending on the model), and x ← S(Upoly(n) ).
E
Details: Suppose that given ((S, h, f ), σ ) generated by A1 e (y, z) and oracle
access to E e , where e ← G 1 (1 ), one can distinguish (σ, (E e (x), h(x))) and
n

(σ, (Ee (1|x| ), h(x))), where x ← S(Upoly(n) ). Then we obtain a distinguisher as in

Definition 5.4.9 as follows. The first part of the distinguisher invokes A1 (while
answering its oracle queries by forwarding these queries to its own E e oracle),
and obtains ((S, h, f ), σ ) ← A1Ee (y, z). It sets x (1) ← S(Upoly(n) ) and x (2) = 1|x | ,
(1)

436

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

and outputs ((x (1) , x (2) ), (σ, h(x (1) ))). That is, (x (1) , x (2) ) is the challenge template,
and it is answered with E e (x (i) ), where i is either 1 or 2. The second part of the
new distinguisher gets as input a challenge ciphertext α ← E e (x (i) ) and the state
generated by the first part (σ, h(x (1) )), and invokes the distinguisher of the contra-
diction hypothesis with input (σ, (α, h(x (1) ))), while answering its oracle queries by
forwarding these queries to its own E e oracle. Thus, the new distinguisher violates
the condition in Definition 5.4.9, in contradiction to the hypothesis that (G, E, D)
has indistinguishable encryptions.

It follows that indistinguishability of encryptions (as per Definition 5.4.9) implies se-
mantic security (as per Definition 5.4.8). (Here, this implication is easier to prove than
in previous cases because the function f is computable via a circuit that is generated
as part of the challenge template [and, without loss of generality, is part of σ ].)
We now turn to the opposite direction. Suppose that (G, E, D) does not have in-
distinguishable encryptions, and consider the pairs (x (1) , x (2) ) produced as a challenge
template by the distinguishing adversary. Following the ideas of the proof of Proposi-
tion 5.2.7, we let the semantic-security adversary generate a corresponding challenge
template (S, h, f ) such that
r The circuit S samples uniformly in {x (1) , x (2) }.
r The function f satisfies f (x (1) ) = 1 and f (x (2) ) = 0.
r The function h is defined arbitrarily subject to h(x (1) ) = h(x (2) ).
Note that here we do not need to use h for passing non-uniform information (e.g., a
description of the distinguisher). Instead, non-uniform information (i.e., the auxiliary
input z to the distinguisher) is passed explicitly by other means (i.e., as the auxiliary
input to the semantic-security adversary).
We stress that when the semantic-security adversary invokes the distinguishing adver-
sary, the former uses its own oracle to answer the queries made by the latter. (Likewise,
the former passes its auxiliary input z to the latter.) The reader may easily verify that
the semantic-security adversary has a noticeable advantage in guessing f (S(Upoly(n) ))
(by using the distinguishing gap between E e (x (1) ) and E e (x (2) )), whereas no algorithm
that only gets h(S(Upoly(n) )) can have any advantage in such a guess. We derive a con-
tradiction to the hypothesis that (G, E, D) satisfies Definition 5.4.8, and the theorem
follows.

Multiple-Message Security. Definitions 5.4.8 and 5.4.9 can be easily generalized to

handle challenges in which multiple plaintexts are encrypted. As in previous cases,
the corresponding (multiple-plaintext) definitions are equivalent. Furthermore, the
multiple-plaintext definitions are equivalent to the single-plaintext definition, both for
public-key and private-key schemes. We stress the equivalence for private-key schemes
(which does not hold for the basic definitions presented in Section 5.1; see Proposi-
tion 5.2.12). To see the equivalence, it is best to consider the notion of indistinguisha-
bility of encryptions. In this case, the argument used in the proof of Theorem 5.2.11
(i.e., the public-key case) can be applied here by using the encryption oracle in order
to produce the ciphertexts needed for the hybrid argument (rather than by generating
437
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

these ciphertexts using knowledge of the encryption-key, which is only possible in the
public-key setting).

5.4.3.2. Constructions
In view of Proposition 5.4.10 (and Theorem 5.4.11), we focus on private-key encryption
schemes (because a public-key encryption scheme is secure under chosen plaintext
attacks if and only if it is secure under passive key-dependent attacks). All the results
presented in Section 5.3.3 extend to security under chosen plaintext attacks. Specifically,
we prove that Constructions 5.3.9 and 5.3.12 remain secure also under a chosen plaintext
attack.

Proposition 5.4.12: Let F and (G, E, D) be as in Construction 5.3.9, and suppose

that F is pseudorandom with respect to polynomial-size circuits. Then the private-key
encryption scheme (G, E, D) is secure under chosen plaintext attacks. The same holds
with respect to Construction 5.3.12.

Proof Sketch: We focus on Construction 5.3.9 and follow the technique underlying the
proof of Proposition 5.3.10. That is, we consider an idealized version of the scheme, in
which one uses a uniformly selected function φ : {0, 1}n → {0, 1}n , rather than the pseu-
dorandom function f s . Essentially, all that the adversary obtains by encryption queries
in the ideal version is pairs (r, φ(r )), where the r ’s are uniformly and independently
distributed in {0, 1}n . As to the challenge itself, the plaintext is “masked” by the value
of φ at another uniformly and independently distributed element in {0, 1}n . Thus, unless
the latter element happens to equal one of the r ’s used by the encryption oracle (which
happens with negligible probability), the challenge plaintext is perfectly masked. Thus,
the ideal version is secure under a chosen plaintext attack, and the same holds for the
real scheme (since otherwise one derives a contradiction to the hypothesis that F is
pseudorandom).

Summary. Private-key and public-key encryption schemes that are secure under cho-
sen plaintext attacks exist if and only if corresponding schemes that are secure under
passive (key-dependent) attacks exist.31

5.4.4. Chosen Ciphertext Attack

We now turn to stronger forms of active attacks in which the adversary may obtain
(from some legitimate user) plaintexts corresponding to ciphertexts of its choice. We
consider two types of such attacks, called chosen ciphertext attacks: In the milder
type, called a priori chosen ciphertext attacks, such decryption requests can be made
only before the challenge ciphertext (for which the adversary should gain knowledge)
is presented. In the stronger type, called a posteriori chosen ciphertext attacks, such
decryption requests can also be made after the challenge ciphertext is presented, so
long as one does not request a decryption of this very (challenge) ciphertext.

31 Hint: When establishing the claim for the private-key case, use Exercise 2.

438

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

Both types of attacks address security threats in realistic applications: In some set-
tings, the adversary may experiment with the decryption module, before the actual
ciphertext in which it is interested is sent. Such a setting corresponds to an a priori
chosen ciphertext attack. In other settings, one may invoke the decryption module on
inputs of one’s choice at any time, but all these invocations are recorded, and real
damage is caused only by knowledge gained with respect to a ciphertext for which a
decryption request was not recorded. In such a setting, protection against a posteriori
chosen ciphertext attacks is adequate. Furthermore, in both cases, decryption requests
can also be made with respect to strings that are not valid ciphertexts, in which case
the decryption module returns a special error symbol.
Typically, in settings in which a mild or strong form of a chosen ciphertext attack is
possible, a chosen plaintext attack is possible, too. Thus, we actually consider combined
attacks in which the adversary may ask for encryption and decryption of strings of its
choice. Indeed (analogously to Proposition 5.4.10), in the case of public-key schemes
(but not in the case of private-key schemes), the combined attack is equivalent to a
“pure” chosen ciphertext attack.

Organization. We start by providing security definitions for the two types of attacks
discussed here. In Section 5.4.4.2, we further extend the definitional treatment of se-
curity (and derive a seemingly stronger notion that is in fact equivalent to the notions
in Section 5.4.4.1). In Section 5.4.4.3 (resp., Section 5.4.4.4) we discuss the construc-
tion of private-key (resp., public-key) encryption schemes that are secure under chosen
ciphertext attacks.

5.4.4.1. Definitions for Two Types of Attacks

Following Section 5.4.3.1 and bearing in mind that we wish to define two types of
chosen ciphertext attacks (i.e., a priori and a posteriori ones), we first formulate the
framework of chosen ciphertext attacks. As in the case of chosen plaintext attacks, we
consider attacks that proceed in four stages corresponding to the generation of a pair
of keys (by a legitimate party), the adversary’s requests (answered by the legitimate
party) to encrypt and/or decrypt strings under the corresponding key, the generation
of a challenge ciphertext (under this key and according to a template specified by the
adversary), and additional requests to encrypt and/or decrypt strings. That is, a chosen
ciphertext attack proceeds as follows:

1. Key generation: A key-pair (e, d) ← G(1n ) is generated (by a legitimate party). In

the public-key setting the adversary is given e, whereas in the private-key setting the
adversary is only given 1n .
2. Encryption and decryption requests: Based on the information obtained so far, the
adversary may request (the legitimate party) to encrypt and/or decrypt strings of
its (i.e., the adversary’s) choice. A request to encrypt the plaintext x is answered
with a value taken from the distribution Ee (x), where e is as determined in Step 1.
A request to decrypt a valid (with respect to E e ) ciphertext y is answered with the
value Dd (y), where d is as determined in Step 1. A request to decrypt a string y that
439
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

is not a valid ciphertext (with respect to E e ) is answered with a special error symbol.
After making several such requests, the adversary moves to the next stage.
3. Challenge generation: Based on the information obtained so far, the adversary spec-
ifies a challenge template and is given an actual challenge. This is done as in the
corresponding step in the framework of chosen plaintext attacks.
4. Additional encryption and decryption requests: Based on the information obtained
so far, the adversary may request the encryptions of additional plaintexts of its
choice. In addition, in the case of an a posteriori chosen ciphertext attack (but
not in the case of an a priori chosen ciphertext attack), the adversary may make
additional decryption requests with the only (natural) restriction that it not be allowed
to ask for a decryption of the challenge ciphertext. All requests are handled as in
Step 2. After making several such requests, the adversary produces an output and
halts.
In the actual definition, as in the case of chosen plaintext attacks, the adversary’s
strategy will be decoupled into two parts corresponding to its actions before and after
the generation of the actual challenge. Each part will be represented by a (proba-
bilistic polynomial-time) two-oracle machine, where the first oracle is an “encryp-
tion oracle” and the second is a “decryption oracle” (both with respect to the cor-
responding key generated in Step 1). As in the case of chosen plaintext attacks, the
two parts are denoted A1 and A2 , and A1 passes state information (denoted σ ) to
A2 . Again, in accordance with the use of non-uniform formulations, we provide A1
with a (non-uniform) auxiliary input. Thus, in the case of (a posteriori chosen cipher-
text attacks on) public-key schemes, the four-step attack process can be written as
follows:
(e, d) ← G(1n )
(τ, σ ) ← A1Ee , Dd (e, z)
def
c = an actual challenge generated according to the template τ
output ← A2Ee , Dd (σ, c)
where A2 is not allowed to make a query regarding the ciphertext in c, and z denotes
the (non-uniform) auxiliary input given to the adversary. In the case of private-key
schemes, the adversary (i.e., A1 ) is given 1n instead of e. In the case of a priori chosen
ciphertext attacks, A2 is not allowed to query Dd (or, equivalently, A2 is only given
oracle access to the oracle Ee ).

Semantic Security. As in the case of chosen plaintext attacks, a definition of semantic

security is derived by an adequate specification of the challenge generation and the
meaning of success. As before, the challenge generation consists of the adversary spec-
ifying a triplet of circuits, denoted (S, h, f ), and being presented with an encryption
of x ← S(Upoly(n) ) along with the partial information h(x). The adversary’s goal is to
guess f (x), and semantic security amounts to saying that the adversary’s success prob-
ability can be matched by a corresponding algorithm that is only given h(x) and 1|x| .
Again, the corresponding algorithm is decoupled into two parts; the first is in charge of
440

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

outputting a challenge template, and the second is in charge of solving the challenge,
where state information is passed from the first part to the second part. Furthermore,
it is again important to require that the challenge template produced by the corre-
sponding algorithm be distributed exactly as the challenge template produced by the
adversary.

Definition 5.4.13 (Semantic Security under Chosen Ciphertext Attacks):

For public-key schemes: A public-key encryption scheme, (G, E, D), is said to be se-
mantically secure under a priori chosen ciphertext attacks if for every pair of
probabilistic polynomial-time oracle machines, A1 and A2 , there exists a pair of
probabilistic polynomial-time algorithms, A1 and A2 , such that the following two
conditions hold:

1. For every positive polynomial p, and all sufficiently large n and z ∈ {0, 1}poly(n)
it holds that
⎡ ⎤
v = f (x) where
⎢ (e, d) ← G(1n ) ⎥
⎢ ⎥
Pr ⎢⎢ ((S, h, f ), σ ) ← A1 E e , Dd
(e, z) ⎥
⎥
⎣ c ← (E e (x), h(x)) , where x ← S(Upoly(n) ) ⎦
v ← A2Ee (σ, c)
⎡ ⎤
v = f (x) where
⎢ ((S, h, f ), σ ) ← A1 (1n , z) ⎥
< Pr ⎢ ⎥+ 1
⎣ x ← S(Upoly(n) ) ⎦ p(n)
 |x|
v ← A2 (σ, 1 , h(x))
2. For every n and z, the first elements (i.e., the (S, h, f ) part) in the random variables
E G (1n ) , DG 2 (1n )
A1 (1n , z) and A1 1 (G 1 (1n ), z) are identically distributed.

Semantic security under a posteriori chosen ciphertext attacks is defined anal-

ogously, except that A2 is given oracle access to both E e and Dd with the restriction
that when given the challenge c = (c , c ), machine A2 is not allowed to make the
query c to the oracle Dd .
For private-key schemes: The definition is identical except that algorithm A1 gets the
security parameter 1n instead of the encryption-key e.
Clearly, the a posteriori version of Definition 5.4.13 implies its a priori version, which
in turn implies Definition 5.4.8. Furthermore, these implications are strict (see Exer-
cises 36 and 35, respectively).

Indistinguishability of Encryptions. As in the case of chosen plaintext attacks,

deriving the corresponding definition of indistinguishability of encryptions (from
the previous framework) is considerably simpler: The challenge generation consists
of the adversary specifying two equal-length strings, and the adversary is presented
with the encryption of one of them.
441
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Definition 5.4.14 (indistinguishability of encryptions under chosen ciphertext attacks):

For public-key schemes: A public-key encryption scheme, (G, E, D), is said to have
indistinguishable encryptions under a priori chosen ciphertext attacks if for
every pair of probabilistic polynomial-time oracle machines, A1 and A2 , for every
positive polynomial p, and for all sufficiently large n and z ∈ {0, 1}poly(n) it holds
that
1
| pn,z
(1)
− pn,z
(2)
| <
p(n)

where
⎡ ⎤
v = 1 where
⎢ (e, d) ← G(1n ) ⎥
⎢ ⎥
(i)
pn,z
def ⎢
= Pr ⎢ E e , Dd
((x , x ), σ ) ← A1
(1) (2)
(e, z) ⎥
⎥
⎣ c ← E e (x (i) ) ⎦
Ee
v ← A2 (σ, c)
where |x (1) | = |x (2) |.

Indistinguishability of encryptions under a posteriori chosen ciphertext attacks

is defined analogously, except that A2 is given oracle access to both E e and Dd with
the restriction that when given the challenge c, machine A2 is not allowed to make
the query c to the oracle Dd .
For private-key schemes: The definition is identical, except that A1 gets the security
parameter 1n instead of the encryption-key e.

Clearly, the a posteriori version of Definition 5.4.14 implies its a priori version, which
in turn implies Definition 5.4.9 as a special case. Again, these implications are strict
(see again Exercises 36 and 35, respectively).

Terminology. We use CCA as a shorthand for chosen ciphertext attack.

Equivalence of Semantic Security and Ciphertext-Indistinguishability. Again, we

show that the two formulations of security (i.e., semantic security and indistinguisha-
bility of encryptions) are in fact equivalent.

Theorem 5.4.15 (equivalence of definitions for CCA): A public-key (resp., private-

key) encryption scheme (G, E, D) is semantically secure under a priori CCA if and
only if it has indistinguishable encryptions under a priori CCA. An analogous claim
holds for a posteriori CCA.

Proof Sketch: We adapt the proof of Theorem 5.4.11 to the current setting. The adap-
tation is straightforward, and we focus on the case of a posteriori CCA security (while
commenting on the case of a priori CCA security).
442

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

In order to show that indistinguishability of encryptions implies semantic security,

given an adversary (A1 , A2 ) we construct the following matching algorithm A1 , A2 :
def
1. A1 (1n , z) = (τ, σ  ), where (τ, σ  ) is generated as follows:
First, A1 generates an instance of the encryption scheme; that is, A1 lets (e, d) ←
G(1n ). Next, A1 invokes A1 , while emulating the oracles E e and Dd , and sets
def
(τ, σ ) ← A1E e , Dd (1n , z). Finally, A1 sets σ  = ((e, d), σ ). (In the case of a-priori
def
CCA security, we may set σ  = (e, σ ), as in the proof of Theorem 5.4.11.)
We comment that the generated key-pair (e, d), allows A1 to emulate the encryption
and decryption oracles E e and Dd .
def
2. A2 (((e, d), σ ), 1m , γ ) = A2E e , Dd (σ, (E e (1m ), γ )), where typically γ = h(x) and m =
def
|x|. (In the case of a priori CCA security, we may set A2 ((e, σ ), 1m , γ ) =
A2Ee (σ, (E e (1m ), γ )), as in the proof of Theorem 5.4.11.)

Again, since A1 merely emulates the generation of a key-pair and the actions of A1 with
respect to such a pair, the equal distribution condition (i.e., Item 2 in Definition 5.4.13)
holds. Using the (corresponding) indistinguishability of encryption hypothesis, we show
that (even in the presence of the encryption oracle E e and a restricted decryption oracle
Dd ) the distributions (σ, (E e (x), h(x))) and (σ, (E e (1|x| ), h(x))) are indistinguishable,
where (e, d) ← G(1n ), ((S, h, f ), σ ) ← A1Ee (y, z) (with y = e or y = 1n depending
on the model), and x ← S(Upoly(n) ). The main thing to notice is that the oracle queries
made by a possible distinguisher of these distributions can be handled by a distinguisher
of encryptions (as in Definition 5.4.14), by passing these queries to its own oracles.
It follows that indistinguishability of encryptions (as per Definition 5.4.14) implies
semantic security (as per Definition 5.4.13).
We now turn to the opposite direction. Here, the construction of a challenge template
(as per Definition 5.4.13) is exactly as the corresponding construction in the proof of
Theorem 5.4.11. Again, the thing to notice is that the oracle queries made by a possible
distinguisher of encryptions (as in Definition 5.4.14) can be handled by the semantic-
security adversary, by passing these queries to its own oracles. We derive a contra-
diction to the hypothesis that (G, E, D) satisfies Definition 5.4.13, and the theorem
follows.

Multiple-Message Security. Definitions 5.4.13 and 5.4.14 can be easily generalized

to handle challenges in which multiple plaintexts are encrypted. We stress that in the
case of a posteriori CCA, the adversary is not allowed to make a decryption query
that equals any of the challenge ciphertexts. As in previous cases, the corresponding
(multiple-plaintext) definitions are equivalent. Furthermore, as in the case of chosen
plaintext attacks, the multiple-plaintext definitions are equivalent to the single-plaintext
definitions (both for public-key and private-key schemes). We stress that this notion of
multiple-message CCA security refers to a single challenge-generation step in which a
sequence of messages (rather than a single message) can be specified. A more general
notion of multiple-message CCA security allows multiple challenge-generation steps
that may be interleaved with the query steps. This notion generalizes the notion of
443
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

chosen ciphertext attacks and is discussed next (i.e., in Subsection 5.4.4.2). Actually,
we will focus on this generalization when applied to a posteriori chosen ciphertext
attacks, although a similar generalization can be applied to a priori chosen ciphertext
attacks (and in fact also to chosen plaintext attacks).

5.4.4.2. A Third Equivalent Definition of a posteriori CCA Security

In continuation of the last paragraph, we consider general attacks during which several
challenge templates may be produced (at arbitrary times and possibly interleaved with
encryption and decryption queries).32 Each of these challenge templates will be an-
swered similarly to the way such templates were answered previously (i.e., by selecting
a plaintext from the specified distribution and providing its encryption together with
the specified partial information). Unlike in Section 5.4.4.1, we will even allow attacks
that make decryption queries regarding ciphertexts obtained as (part of) the answer
to previous challenge templates. After such an attack, the adversary will try to obtain
information about the unrevealed plaintexts, and security holds if its success probabil-
ity can be met by a corresponding benign adversary that does not see the ciphertexts.
Indeed, the discussion requires clarification and careful formulation, provided next.
We start with a description of the actual attacks. It will be convenient to change the
formalism and consider the generation of challenge templates as challenge queries
that are answered by a special oracle called the tester, and denoted Te,r , where e is
an encryption-key and r is a random string of adequate length. On query a challenge
template of the form (S, h), where S is a sampling circuit and h is (an evaluation
circuit for) a function, the (randomized) oracle Te,r returns the pair (E e (x), h(x)), where
x = S(r ). (Indeed, we may assume without loss of generality that for all queries (S, h),
it holds that S is a sampling circuit that generates strings of length that fits h’s input.)
We stress that r is not known to the adversary, and that this formalism supports the
generation of dependent challenges as well as of independent ones.33 A multiple-
challenge CCA is allowed queries to Te,r , as well as unrestricted queries to both E e
and the corresponding Dd , including decryption queries referring to previously obtained
challenge ciphertexts. It terminates by outputting a function f and a value v, hoping
that f (x 1 , ..., x t ) = v, where x i = S i (r ) and (S i , h i ) is the i-th challenge query made
by the adversary. Note that the description of f (in terms of an evaluation circuit) may
encode various information gathered by the adversary during its attack (e.g., it may
even encode its entire computation transcript).34

32 Note that in this section we generalize the notion of an a posteriori chosen ciphertext attack. When generalizing
the notion of an a priori chosen ciphertext attack, we disallow decryption queries after the first challenge template
is produced.
33 Independently distributed plaintexts can be obtained by sampling circuits that refer to disjoint parts of the

random string r . On the other hand, we can obtain a pair of plaintexts such that the second plaintext is a function
of the first one by letting the second sampling circuit equal the composition of the first sampling circuit with
the said function. That is, making queries of the form (S, ·) and (C ◦ S, ·), where C is a deterministic circuit,
def
we obtain answers that refer to the plaintexts x = S(r ) and C(x).
34 In general, the description of functions in terms of circuits that are not of minimal size is redundant, and opens

the door for encoding of additional information.

444

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

We now turn to describe the benign adversary (which does not see the ciphertexts).
Such an adversary is given oracle access to a corresponding oracle, Tr , that behaves
as follows. On query a challenge template of the form (S, h), the oracle returns h(x),
where x = S(r ). (Again, r is not known to the adversary.) Like the real adversary, the
benign adversary also terminates by outputting a function f and a value v, hoping that
f (x 1 , ..., x t ) = v, where x i = S i (r ) and (S i , h i ) is the i-th challenge query made by
the adversary.
Security amounts to asserting that the effect of any efficient multiple-challenge CCA
can be simulated by an efficient benign adversary that does not see the ciphertexts. As
in Definition 5.4.13, the simulation has to satisfy two conditions: First, the probability
that f (x 1 , ..., x t ) = v in the CCA must be met by the probability that a corresponding
event holds in the benign model (where the adversary does not see ciphertexts). Second,
the challenge queries, as well as the function f , should be distributed similarly in the
two models. Actually, the second condition should be modified in order to account
for the case that the real CCA adversary makes a decryption query that refers to a
ciphertext that is contained in the answer given to a previous challenge query, denoted
(S, h). Note that such a decryption query (i.e., E e (S(r ))) reveals S(r ) to the attacker,
and that this has nothing to do with the security of the encryption scheme. Thus, it is
only fair to also allow the benign adversary (which sees no ciphertexts) to make the
corresponding query, which is equivalent to the challenge query (S, id), where id is
the identity function. (Indeed, the answer will be id(S(r )) = S(r ).)
In order to obtain the actual definition, we need to define the trace of the execution
of these two types of adversaries. For a multiple-challenge CCA adversary, denoted
A, the trace is defined as the sequence of challenge queries made during the attack,
augmented by fictitious challenge queries such that the (fictitious challenge) query
(S, id) is included if and only if the adversary made a decryption query c such that (c, ·)
is the answer given to a previous challenge query of the form (S, ·). (This convention
is justified by the fact that the answer (E e (S(r )), id(S(r ))) to the fictitious challenge
query (S, id) is efficiently computable from the answer S(r ) to the decryption query
c = E e (S(r )).)35 In fact, for simplicity, we will assume in the following definition that A
(or rather a minor modification of A) actually makes these fictitious challenge queries.
For the benign adversary, denoted B, the trace is defined as the sequence of challenge
queries made during the attack.

Definition 5.4.16 (multiple-challenge CCA security):

For public-key schemes: A public-key encryption scheme, (G, E, D), is said to be se-
cure under multiple-challenge-chosen ciphertext attacks if for every probabilis-
tic polynomial-time oracle machine A there exists a probabilistic polynomial-time
oracle machine B such that the following two conditions hold:
1. For every positive polynomial p, and all sufficiently large n and z ∈ {0, 1}poly(n)

35 Indeed, the value (E e (S(r )), id(S(r ))) is obtained from S(r ) by making an encryption query S(r ).

445
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

it holds that
⎡ ⎤
v = f (x 1 , ..., x t ) where
⎢ (e, d) ← G(1n ) and r ← Upoly(n) ⎥
Pr ⎢
⎣
⎥
⎦
( f, v) ← A Ee , Dd , Te,r (e, z)
x ← S (r ), for i = 1, ..., t.
i i

⎡ ⎤
v = f (x 1 , ..., x t ) where
⎢ r ← Upoly(n) ⎥
< Pr ⎢ ⎥+ 1
⎣ ( f, v) ← B Tr (1n , z) ⎦ p(n)
x ← S (r ), for i = 1, ..., t.
i i

where t is the number of challenge queries made by A (resp., B), and S i is the
first part of the i-th challenge query made by A (resp., B) to Te,r (resp., to Tr ).
2. The following two probability ensembles, indexed by n ∈ N and z ∈ {0, 1}poly(n),
are computationally indistinguishable:
(a) The trace of A E G 1 (1 ) , DG 2 (1 ) , TG 1 (1 ),Upoly(n) (G 1 (1n ), z) augmented by the first ele-
n n n

ment of its output pair (i.e., the function f ).

(b) The trace of B TUpoly(n) (1n , z) augmented by the first element of its output pair.

That is, in both cases, we refer to the corresponding sequence

((S 1 , h 1 ), ..., (S t , h t ), f )
where (S i , h i ) denotes the i-th challenge query.
For private-key schemes: The definition is identical, except that machine A gets the
security parameter 1n instead of the encryption-key e.

To get more comfortable with Definition 5.4.16, consider the special case in which
the real CCA adversary does not make decryption queries to ciphertexts obtained as
part of answers to challenge queries. (In the proof of Theorem 5.4.17, such adver-
saries will be called canonical and will be showed to be as powerful as the general
ones.) The trace of such adversaries equals the sequence of actual challenge queries
made during the attack (without any fictitious challenge queries), which simplifies the
meaning of Condition 2. Furthermore, the special case in which such an adversary
makes a single challenge query is very similar to Definition 5.4.13, with the exception
that here Condition 2 allows computational indistinguishability (rather than requiring
identical distributions). Still, this very restricted case (of Definition 5.4.16) does imply
security under a posteriori CCA (see Exercise 37). More importantly, the following
holds:

Theorem 5.4.17 (a posteriori CCA implies Definition 5.4.16): Let (G, E, D) be a

public-key (resp., private-key) encryption scheme that is secure under a posteriori
CCA. Then (G, E, D) is secure under multiple-challenge CCA.

Proof Sketch: As a bridge between the multiple-challenge CCA and the corresponding
benign adversary that does not see the ciphertext, we consider canonical adversaries that
446

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

can perfectly simulate any multiple-challenge CCA without making decryption queries
to ciphertexts obtained as part of answers to challenge queries. Instead, these canonical
adversaries make corresponding queries of the form (S, id), where id is the identity
function and (S, ·) is the challenge-query that was answered with the said ciphertext.
Specifically, suppose that a multiple-challenge CCA has made the challenge query
(S, h), which was answered by (c, v) where c = Ee (x), v = h(x) and x = S(r ), and at a
later stage makes the decryption query c, which is to be answered by Dd (c) = x. Then,
the corresponding canonical adversary makes the challenge query (S, h) as the original
adversary, receiving the same pair (c, v), but later (instead of making the decryption
query c) the canonical adversary makes the challenge query (S, id), which is answered
by id(S(r )) = x = Dd (c). Note that the trace of the corresponding canonical adversary
is identical to the trace of the original CCA adversary (and the same holds with respect
to their outputs).
Thus, given an a posteriori CCA–secure encryption scheme, it suffices to establish
Definition 5.4.16 when the quantification is restricted to canonical adversaries A. In-
deed, as in previous cases, we construct a benign adversary B in the natural manner:
On input (1n , z), machine B generates (e, d) ← G(1n ), and invokes A on input (y, z),
where y = e if we are in the public-key case and y = 1n otherwise. Next, B emulates
all oracles expected by A, while using its own oracle Tr . Specifically, the oracles E e and
Dd are perfectly emulated by using the corresponding keys (known to B), and the oracle
Te,r is (imperfectly) emulated using the oracle Tr ; that is, the query (S, h) is forwarded
to Tr , and the answer h(S(r )) is augmented with E e (1m ), where m is the number of
output bits in S. Note that the latter emulation (i.e., the answer (E e (1|S(r )| ), h(S(r )))) is
imperfect since the answer of Te,r would have been (E e (S(r )), h(S(r ))), yet (as we shall
show) A cannot tell the difference.
In order to show that B satisfies both conditions of Definition 5.4.16 (with respect
to this A), we will show that the following two ensembles are computationally indis-
tinguishable:

1. The global view in a real attack of A on (G, E, D). That is, we consider the output
of the following experiment:

(a) (e, d) ← G(1n ) and r ← Upoly(n) .

(b) ( f, v) ← A E e , Dd , Te,r (y, z), where y = e if we are in the public-key case and
y = 1n otherwise. Furthermore, we let ((S 1 , h 1 ), ..., (S t , h t )) denote the trace of
the execution A Ee , Dd , Te,r (y, z).
(c) The output is ((S 1 , h 1 ), ..., (S t , h t )), ( f, v), r .

2. The global view in an attack emulated by B . That is, we consider the output of an

experiment as in Item 1, except that A Ee , Dd , Te,r (y, z) is replaced by A E e , Dd , Te,r (y, z),

where on query (S, h) the oracle Te,r replies with (E e (1|S(r )| ), h(S(r ))) rather than
with (Ee (S(r )), h(S(r ))).

Note that computational indistinguishability of these ensembles immediately implies

Condition 2 of Definition 5.4.16, whereas Condition 1 also follows because using r , we
447
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

can determine whether or not f (S 1 (r ), ..., S t (r )) = v holds (for ( f, v) and S 1 , ..., S t that
appear in the ensemble’s output). Also note that these ensembles may be computationally
indistinguishable only in the case where A is canonical (which we have assumed to be
the case).36
The computational indistinguishability of these two ensembles is proven using a
hybrid argument, which in turn relies on the hypothesis that (G, E, D) has indistin-
guishable encryptions under a posteriori CCAs. Specifically, we introduce t + 1 mental
experiments that are hybrids of the two ensembles (which we wish to relate). Each of
these mental experiments is given oracle access to E e and Dd , where (e, d) ← G(1n ) is
selected from the outside. The i-th hybrid experiment uses these two oracles (in addition
to y, which equals e in the public-key case and 1n otherwise) in order to emulate an
i
execution of A Ee , Dd , e,r ( y, z), where r is selected by the experiment itself and ie,r is

a hybrid of Te,r and Te,r . Specifically, ie,r is a history-dependent process that answers

like Te,r on the first i queries and like Te,r on the rest. Thus, for i = 0, ..., t, we define
the i-th hybrid experiment as a process that, given y (which equals e or 1n ) and oracle
access to E e and Dd , where (e, d) ← G(1n ), behaves as follows:

1. The process selects r ← Upoly(n) .

i
2. The process emulates an execution of A E e , Dd , e,r (y, z), where y = e if we are in the
public-key case and y = 1n otherwise, by using the oracles E e and Dd . Specifically,
the answers of ie,r are emulated using the knowledge of r and oracle access to E e :
the j-th query to ie,r , denoted (S j , h j ), is answered by (E e (S j (r )), h j (S j (r ))) if
j ≤ i and is answered by (E e (1|S (r )| ), h j (S j (r ))) otherwise. (The process answers
j

A’s queries to E e and Dd by forwarding them to its own corresponding oracles.)

i
3. As before, ( f, v) denotes the output of A E e , Dd , e,r (y, z), and ((S 1 , h 1 ), ..., (S t , h t ))
denotes its trace. The process outputs ((S 1 , h 1 ), ..., (S t , h t )), ( f, v), r .

We stress that since A is canonical, none of the Dd -queries equals a ciphertext obtained
as part of the answer of a ie,r -query.
Clearly, the distribution of the 0-hybrid is identical to the distribution of the global
view in an attack emulated by B, whereas the distribution of the t-hybrid is identical to
the distribution of the global view in a real attack by A. On the other hand, distinguishing
the i-hybrid from the (i + 1)-hybrid yields a successful a posteriori CCA (in the sense of
distinguishing encryptions). That is, assuming that one can distinguish the i-hybrid from
the (i + 1)-hybrid, we construct an a posteriori CCA adversary (as per Definition 5.4.14)

36 Non-canonical adversaries can easily distinguish the two types of views by distinguishing the oracle Te,r from
 . For example, suppose we make a challenge query with a sampling-circuit S that generates some
oracle Te,r
distribution over {0, 1}m \ {1m }, next make a decryption query on the ciphertext obtained in the challenge
query, and output the answer. Then, in case we query the oracle Te,r , we output Dd (E e (S(r ))) = 1m ; whereas
 , we output D (E (1m )) = 1m . Recall, however, that at this point of the proof,
in case we query the oracle Te,r d e
we are guaranteed that A is canonical (and indeed A might have been derived by perfectly emulating some
non-canonical A ). An alternative way of handling non-canonical adversaries is to let B handle the disallowed
(decryption) queries by making the corresponding challenge query, and returning its answer rather than the
 , can detect which queries are disallowed.)
decryption value. (Note that B, which emulates Tr,e

448

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

as follows. For (e, d) ← G(1n ), given y = e if we are in the public-key case and y = 1n
otherwise, the attacker (having oracle access to E e and Dd ) behaves as follows:

1. The attacker selects r ← Upoly(n) .

j
2. The attacker emulates an execution of A Ee , Dd , e,r (y, z), where j ∈ {i, i + 1} (is un-
known to the attacker), as follows. The queries to E e and Dd are answered by using
the corresponding oracles available to the attacker, and the issue is answering the
j j
queries to e,r . The first i queries to e,r are answered as in both ie,r and i+1 e,r (i.e.,
query (S, h) is answered by (E e (S(r )), h(S(r )))), and the last t − (i + 1) queries are
|S(r )|
also answered as in both ie,r and i+1 e,r (i.e., by (E e (1 ), h(S(r ))), this time).
The i + 1 query, denoted (S , h ), is answered by producing the challenge
st i+1 i+1

template (S i+1 (r ), 1|S (r )| ), which is answered by the challenge ciphertext c (where

i+1

c ∈ {E e (S i+1 (r )), E e (1|S (r )| )}), and replying with (c, h i+1 (S i+1 (r ))).
i+1

= E e (1|S
i+1 (r )|
Note that if c = E e (S i+1 (r )), then we emulate i+1
e,r , whereas if c ) then
we emulate ie,r .
j
3. Again, ( f, v) denotes the output of A E e , Dd , e,r (y, z), and ((S 1 , h 1 ), ..., (S t , h t )) de-
notes its trace. The attacker feeds ((S 1 , h 1 ), ..., (S t , h t )), ( f, v), r to the hybrid dis-
tinguisher (which we have assumed to exist toward the contradiction), and outputs
whatever the latter does.

This is an a posteriori CCA as in Definition 5.4.14: It produces a single challenge

(i.e., the pair of plaintexts (Si+1 (r ), 1|S (r )| )), and distinguishes the case that it is
i+1

given the ciphertext c = E e (S i+1 (r )) from the case that it is given the ciphertext c =
E e (1|S (r )| ), without querying Dd on the challenge ciphertext c. The last assertion
i+1

follows by the hypothesis that A is canonical, and so none of the Dd -queries that A
j
makes equals the ciphertext c obtained as (part of) the answer to the i + 1st e,r -query.
Thus, distinguishing the i + 1 and i-th hybrids implies distinguishing encryptions
st

under an a posteriori CCA, which contradicts our hypothesis regarding (G, E, D). The
theorem follows.

Further Generalization. Recall that we have allowed arbitrary challenge queries of the
form (S, h) that were answered by (E e (S(r )), h(S(r ))). Instead, we may allow queries
of the form (S, h) that are answered by (E e (S(r )), h(r )); that is, h is applied to r itself
rather than to S(r ). Actually, given the “independence” of h from S, one could have
replaced the challenge queries by two types of queries: partial-information (on r )
queries that correspond to the h’s (and are answered by h(r )), and encrypted partial-
information queries that correspond to the S’s (and are answered by E e (S(r ))). As
shown in Exercise 38, all these forms are in fact equivalent.

5.4.4.3. Constructing CCA-Secure Private-Key Schemes

In this section we present simple constructions of CCA-secure private-key encryption
schemes. We start with a priori CCA, and next turn to a posteriori CCA.
449
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Security under a-priori CCA. All the results presented in Section 5.3.3 extend to
security under a priori chosen ciphertext attacks. Specifically, we prove that Construc-
tions 5.3.9 and 5.3.12 remain secure also under an a priori CCA.

Proposition 5.4.18: Let F and (G, E, D) be as in Construction 5.3.9, and suppose

that F is pseudorandom with respect to polynomial-size circuits. Then the private-key
encryption scheme (G, E, D) is secure under a priori chosen ciphertext attacks. The
same holds with respect to Construction 5.3.12.

Proof Sketch: As in the proof of 5.4.12, we focus on Construction 5.3.9, and consider
an idealized version of the scheme in which one uses a uniformly selected function
φ : {0, 1}n → {0, 1}n (rather than the pseudorandom function f s ). Again, all that the ad-
versary obtains by encryption queries in the ideal version is pairs (r, φ(r )), where the
r ’s are uniformly and independently distributed in {0, 1}n . Similarly, decryption queries
provide the adversary with pairs (r, φ(r )), but here the r ’s are selected by the adversary.
Still in an a priori CCA, all decryption queries are made before the challenge is pre-
sented, and so these r ’s are selected (by the adversary) independent of the challenge.
Turning to the challenge itself, we observe that the plaintext is “masked” by the value
of φ at another uniformly and independently distributed element in {0, 1}n , denoted
rC . We stress that rC is independent of all r ’s selected in decryption queries (because
these occur before rC is selected), as well as being independent of all r ’s selected by the
encryption oracle (regardless of whether these queries are made prior or subsequently
to the challenge). Now, unless rC happens to equal one of the r ’s that appear in the
pairs (r, φ(r )) obtained by the adversary (which happens with negligible probability),
the challenge plaintext is perfectly masked. Thus, the ideal version is secure under an a
priori CCA. The same holds for the real scheme, because pseudorandom functions are
indistinguishable from truly random ones (even by machines that adaptively query the
function at arguments of their choice).

Security under a-posteriori CCA. Unfortunately, Constructions 5.3.9 and 5.3.12 are
not secure under a posteriori chosen ciphertext attacks: Given a challenge ciphertext
(r, x ⊕ f s (r )), the adversary may obtain f s (r ) by making the query (r, y  ), for any
y  = x ⊕ f s (r ). This query is allowed and is answered with x  such that y  = x  ⊕ f s (r ).
Thus, the adversary may recover the challenge plaintext x from the challenge ciphertext
def
(r, y), where y = x ⊕ f s (r ), by computing y ⊕ (y  ⊕ x  ). Thus, we should look for new
private-key encryption schemes if we want to obtain one that is secure under a posteriori
CCA. Actually, we show how to transform any private-key encryption scheme that is
secure under chosen plaintext attack (CPA) into one that is secure under a posteriori
CCA.
The idea underlying our transformation (of CPA-secure schemes into CCA-secure
ones) is to eliminate the adversary’s gain from chosen ciphertext attacks by making it
infeasible to produce a legitimate ciphertext (other than the ones given explicitly to the
adversary). Thus, an a posteriori CCA adversary can be emulated by a chosen plaintext
attack (CPA) adversary, while almost preserving the success probability.
450

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

The question is indeed how to make it infeasible for the (a posteriori CCA) adversary
to produce a legitimate ciphertext (other than the ones explicitly given to it). One answer
is to use “Message Authentication Codes” (MACs) as defined in Section 6.1.37 That is,
we augment each ciphertext with a corresponding authentication tag (which is “hard
to forge”), and consider an augmented ciphertext to be valid only if it consists of a
valid (string,tag)-pair. For the sake of self-containment (and concreteness), we will use
a specific implementation of such MACs via pseudorandom functions. Incorporating
this MAC in Construction 5.3.9, we obtain the following:

Construction 5.4.19 (a private-key block-cipher secure against a-posteriori CCA): As

in Construction 5.3.9, let F = {Fn } be an efficiently computable function ensemble
and let I be the function-selection algorithm associated with it; i.e., I (1n ) selects a
function f s with distribution Fn . We define a private-key block-cipher, (G, E, D), with
block-length (n) = n as follows:

Key-generation: G(1n ) = ((k, k  ), (k, k  )), where k and k  are generated by two inde-
pendent invocations of I (1n ).
Encrypting plaintext x ∈ {0, 1}n (using the key (k, k  )):

E k,k  (x) = ((r, f k (r ) ⊕ x), f k  (r, f k (r ) ⊕ x)),

where r is uniformly chosen in {0, 1}n .

Decrypting ciphertext (r, y) (using the key (k, k  )): Dk, k  ((r, y), t) = f k (r ) ⊕ y if
f k  (r, y) = t and Dk,k  ((r, y), t) = ⊥ otherwise.

Proposition 5.4.20: Let F and (G, E, D) be as in Construction 5.4.19, and suppose

that F is pseudorandom with respect to polynomial-size circuits. Then the private-key
encryption scheme (G, E, D) is secure under a posteriori chosen ciphertext attacks.

Proof Sketch: Following the motivation preceding the construction, we emulate any a
posteriori CCA adversary by a CPA adversary. Specifically, we need to show how to
answer decryption queries made by the CCA adversary. Let us denote such a generic
query by ((r, y), t), and consider the following three cases:

1. If ((r, y), t) equals the answer given to some (previous) encryption query x, then we
answer the current query with x.
Clearly, the answer we give is always correct.
2. If ((r, y), t) equals the challenge ciphertext, then this query is not allowed.
3. Otherwise, we answer that ((r, y), t) is not a valid ciphertext.
We need to show that our answer is indeed correct. Recall that in this case, ((r, y), t)
neither appeared before as an answer to an encryption query nor equals the chal-
lenge ciphertext. Since for every (r, y) there is a unique t  such that ((r, y), t  ) is
a valid ciphertext, the case hypothesis implies that one of the following sub-cases
37 In fact, we need to use secure Message Authentication Codes that have unique valid tags (or at least are
super-secure), as discussed in Section 6.5.1 (resp., Section 6.5.2).

451
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

must occur:
Case 1: Some ((r, y), t  ), with t  = t, has appeared before either as an answer to an
encryption query or as the challenge ciphertext. In this case, ((r, y), t) is definitely
not a valid ciphertext, because ((r, y), t  ) is the unique valid ciphertext of the form
((r, y), ·).
Case 2: No triple of the form ((r, y), ·) has appear before (as such an answer to an
encryption query or as the challenge ciphertext). In this sub-case, the ciphertext
is valid if and only if t = f k  (r, y). That is, in order to produce such a valid
ciphertext, the adversary must guess the value of f k  at (r, y), when only seeing
the value of f k  at other arguments. By the pseudorandomness of the function f k  ,
the adversary may succeed in such a guess only with negligible probability, and
hence our answer is wrong only with negligible probability.
Finally, note that the CPA-security of Construction 5.3.9 (see Proposition 5.4.12) implies
the CPA-security of Construction 5.4.19. The proposition follows.

An Alternative Proof of Proposition 5.4.20. Augmenting the proof of Proposi-

tion 5.4.18, we (need to) consider here also decryption queries made after the challenge
ciphertext, denoted ((rC , yC ), tC ), is presented. Let us denote such a generic decryp-
tion query by ((r, y), t). We consider four cases, ignoring the unlikely case that some
encryption query is answered by a pair of the form ((rC , ·), ·):
1. If r = rC then the query ((r, y), t) can be treated as in the proof of Proposition 5.4.18,
because it reveals nothing on f k (rC ). Indeed, such a query is not more dangerous
than a query made during an a priori CCA attack.
2. If r = rC and y = yC then, except with negligible probability, the query ((r, y), t) is
not a valid ciphertext, because it is infeasible to guess the value of f k  (r, y) (which
is the only value of t  such that ((r, y), t  ) is valid). Thus, such queries (which are
almost always answered by ⊥) can be ignored.
3. If (r, y) = (rC , yC ) and t = tC then (surely) the query ((r, y), t) is not a valid cipher-
text, and can be ignored (as in the previous case).
4. If (r, y, t) = (rC , yC , tC ) then the query ((r, y), t) is not allowed.
The proposition follows.

The same construction and analysis can be applied to Construction 5.3.12. Combin-
ing Proposition 5.4.20 with Corollary 3.6.7, we get:

Theorem 5.4.21: If there exist (non-uniformly hard) one-way functions, then there exist
private-key encryption schemes that are secure under a posteriori chosen ciphertext
attacks.

5.4.4.4. Constructing CCA-Secure Public-Key Schemes

In this section we present fairly complicated constructions of CCA-secure public-key
encryption schemes. Again, we start by considering a priori CCA, and then augment
452

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

the constructions in order to handle a posteriori CCA. Specifically, we will show how
to transform any public-key encryption scheme that is secure in the passive (key-
dependent) sense into one that is secure under a posteriori CCA. As in the case of private-
key schemes, the idea underlying the transformation is to eliminate the adversary’s gain
from chosen ciphertext attacks.
Recall that in the case of private-key schemes, the adversary’s gain from a CCA was
eliminated by making it infeasible (for the adversary) to produce legitimate ciphertexts
(other than those explicitly given to it). However, in the context of public-key schemes,
the adversary can easily generate legitimate ciphertexts (by applying the keyed encryp-
tion algorithm to any plaintext of its choice). Thus, in the current context, the adversary’s
gain from a CCA is eliminated by making it infeasible (for the adversary) to produce
legitimate ciphertexts without “knowing” the corresponding plaintext. This, in turn,
will be achieved by augmenting the plaintext with a non-interactive zero-knowledge
“proof of knowledge” of the corresponding plaintext.

NIZK: Preliminaries. Strong forms of Non-Interactive Zero-Knowledge (NIZK)

proofs will play a key role in our transformation, and we will assume that the reader is
familiar with the main notions and results that are presented in Section 4.10. Since the
notion of a proof-of-knowledge is quite complex in general (cf. Section 4.7), and more
so in the non-interactive (zero-knowledge) context (let alone that we will need strength-
enings of it), we will not make explicit use of this notion (i.e., of proof-of-knowledge).
Instead, we will use non-interactive (zero-knowledge) proofs of membership (NIZK)
as defined in Section 4.10. In fact, our starting point is the definition of adaptive NIZK
systems (i.e., Definition 4.10.15), when extended to handle assertions of a priori un-
bounded length (as discussed at the beginning of Section 4.10.3.1). We focus on proof
systems in which the prover is implemented by a probabilistic polynomial-time al-
gorithm that is given a suitable auxiliary-input (e.g., an NP-witness). For the sake of
clarity, let us reproduce the resulting definition.

Definition 5.4.22 (adaptive NIZK): An adaptive non-interactive zero-knowledge

proof system (adaptive NIZK) for a language L ∈ N P, with an NP-relation R L ,
consists of a pair of probabilistic polynomial-time algorithms, denoted (P, V ), that
satisfy the following:
r Syntax: Both machines are given the same uniformly selected reference string
r ∈ {0, 1}m along with an actual input x ∈ {0, 1}∗ such that |x| = poly(m) and
an auxiliary input. Specifically, on input r , x and w (supposedly, (x, w) ∈ R L ), the
prover P outputs an alleged proof π ← P(x, w, r ); whereas on input r , x and π,
the verifier V decides according to V (x, r, π) ∈ {0, 1}.
r Completeness: For every (x, w) ∈ R L with |x| = poly(m), the probability that V
does not accept the input x (based on the proof P(x, w, Um ) and the reference string
Um ) is negligible; that is, Pr[V (x, Um , P(x, w, Um )) = 1] is negligible. (Typically,
the error probability here is zero, in which case we say that the proof has perfect
completeness.)
453
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

r Adaptive Soundness: For every  : {0, 1}m → ({0, 1}poly(m) \ L) and every :
{0, 1} → {0, 1}
m
, the probability that V accepts the input (Um ) (based
poly(m)

on the proof (Um ) and the reference string Um ) is negligible; that is,
Pr[V ((Um ), Um , (Um )) = 1] is negligible.
r Adaptive Zero-Knowledge: There exist two probabilistic polynomial-time algo-
rithms, S1 and S2 , such that for every pair of functions  : {0, 1}m → ({0, 1}poly(m) ∩
L) and W : {0, 1}m → {0, 1}poly(m) such that  and W are both implementable
by polynomial-size circuits and ((r ), W (r )) ∈ R L (∀r ∈ {0, 1}m ), the ensembles
{(Um , (Um ), P((Um ), W (Um ), Um ))}m∈N and {S  (1m )}m∈N are computationally
indistinguishable (by non-uniform families of polynomial-size circuits), where
S  (1m ) denotes the output of the following randomized process:
1. (r, s) ← S1 (1m );
2. x ← (r );
3. π ← S2 (x, s);
4. Output (r, x, π).
Indeed, S is a two-stage simulator that first produces (obliviously of the actual input)
an alleged reference string r (along with the auxiliary information s),38 and then,
given an actual input (which may depend on r ), simulates the actual proof.

Note that it is important that in the zero-knowledge condition, the function  is required
to be implementable by polynomial-size circuits (because otherwise only languages in
BPP can have such proof systems; see Exercise 39). In the rest of this subsection,
whenever we refer to an adaptive NIZK, we mean this definition. Actually, we may
relax the adaptive soundness condition so that it only applies to functions  and that
are implementable by polynomial-size circuits. That is, computational soundness will
actually suffice for the rest of this subsection.

Additional Conventions. Note that (analogously to Proposition 5.4.10) in the case

of public-key schemes, the combined chosen plaintext and ciphertext attack (as in
Definitions 5.4.13 and 5.4.14) is equivalent to a “pure” chosen ciphertext attack. Thus,
in this subsection we consider only attacks of the latter type. Another technical point is
that in our construction we can use any public-key encryption scheme that is secure in
the passive (key-dependent) sense, provided that for all but a negligible measure of the
key-pairs that it generates, there is no decryption error. For simplicity of presentation,
we will assume that the basic encryption scheme has no decryption error at all (i.e., on
all key-pairs).

The General Framework. The following schema (for constructing CCA-secure

public-key encryption schemes) uses a passively secure public-key encryption

38 The auxiliary information s may explicitly contain r . Alternatively, s may just equal the coins used by S1 . In the
constructions that follow, we do not follow either of these conventions, but rather let s equal the very information
about r that S2 needs.

454

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

scheme, denoted (G, E, D), and an adaptive NIZK, denoted (P, V ), for a related
NP-set.

Construction 5.4.23 (CCA-security construction framework): Let E e (x, s) denote the

ciphertext produced by E when given the encryption-key e, the plaintext x, and the
coins s; that is, E e (x) ← E e (x, s), where s is selected uniformly among the set of
poly(|e|, |x|)-long bit strings. We use an adaptive NIZK (P, V ) for the language L R
defined by the following NP-relation:
def
R = {((e1 , e2 , y1 , y2 ), (x, s1 , s2 )) : y1 = E e1 (x, s1 ) & y2 = E e2 (x, s2 )} (5.12)

That is, (e1 , e2 , y1 , y2 ) ∈ L R if y1 and y2 are encryptions of the same plaintext, produced
using the encryption-keys e1 and e2 , respectively.
def
Key-generation: G  (1n ) = ((e1 , e2 , r ), (d1 , d2 , r )), where (e1 , d1 ) and (e2 , d2 ) are se-
lected at random by invoking G(1n ) twice, and r is uniformly distributed in
{0, 1}n .
Encrypting plaintext x ∈ {0, 1}∗ (using the key e = (e1 , e2 , r )):
def
E e (x) = (y1 , y2 , π), where s1 , s2 are uniformly selected poly(n)-long bit strings,
y1 = E e1 (x, s1 ), y2 = E e2 (x, s2 ), and π ← P((e1 , e2 , y1 , y2 ), (x, s1 , s2 ), r ).
Decrypting ciphertext (y1 , y2 , π) (using the key d = (d1 , d2 , r )):
If V ((e1 , e2 , y1 , y2 ), r, π) = 1, then return Dd1 (y1 ) or else return an error symbol
indicating that the ciphertext is not valid.

Indeed, our choice to decrypt according to y1 (in case π is a valid proof) is immaterial,
and we could as well decrypt according to y2 . Another alternative could be to decrypt
according to both y1 and y2 and return a result only if both outcomes are identical (and
π is a valid proof). We stress that, here as well as in the following analysis, we rely
on the hypothesis that decryption is error-free, which implies that Dd (E e (x)) = x for
every (e, d) in the range of G. Thus, Dd1 (y1 ) = Dd2 (y2 ), for any (e1 , e2 , y1 , y2 ) ∈ L R ,
where the (ei , di )’s are in the range of G.
Clearly, Construction 5.4.23 constitutes a public-key encryption scheme; that is,
Dd (E e (x)) = x, provided that the NIZK proof generated during the encryption stage


was accepted during the decryption stage. Indeed, if the NIZK system enjoys perfect
completeness (which is typically the case), then the decryption error is zero. By the zero-
knowledge property, the passive security of the original encryption scheme (G, E, D)
is preserved by Construction 5.4.23. Intuitively, creating a valid ciphertext seems to
imply “knowledge” of the corresponding plaintext, but this appealing claim should be
examined with more care (and in fact is not always valid). Furthermore, as stated pre-
viously, our actual proof will not refer to the notion of “knowledge.” Instead, the actual
proof will proceed by showing how a chosen ciphertext attack on Construction 5.4.23
can be transformed into a (key-dependent) passive attack on (G, E, D). In fact, we will
need to augment the notion of (adaptive) NIZK in order to present such a transfor-
mation. We will do so in two steps. The first augmentation will be used to deal with
a priori CCA, and further augmentation will be used to deal with a posteriori CCA.
455
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Step I: a-priori CCA

Let us start by considering an a priori CCA. Given such an adversary A, we construct
a passive adversary B that attacks (G, E, D) by emulating the attack of A on Con-
struction 5.4.23. One important observation is that the latter encryption scheme uses
two keys of the original scheme. Thus, given an encryption-key of the original scheme,
B generates another encryption-key (while storing the corresponding decryption-key)
and invokes A, giving it the pair of encryption-keys (along with a reference string to
be generated as discussed in the next paragraph). When A makes a decryption query,
B may answer the query by using the stored decryption-key (generated by B before).
This works provided that the query ciphertext contains a pair of ciphertexts of the same
plaintext according to the two keys, which is the reason we augmented the ciphertext
pairs by a proof of consistency. Thus, actually, B should examine the latter proof and
act analogously to the decryption process of Construction 5.4.23.
The next problem arises when A asks to be given a challenge. Algorithm B forwards
the request as its own challenge template, but the challenge given to B is a single
ciphertext of the original scheme, and so B needs to augment it into something that
looks like a ciphertext of Construction 5.4.23. Here is where we rely on the zero-
knowledge property of the proof of consistency (for producing the required proof that
relates to a plaintext we do not know), but in order to do so, the reference string needs
to be generated by the simulator (rather than be uniformly distributed). But this leads to
the following problem: When referring (in the previous paragraph) to the soundness of
the proofs of consistency, we assumed that the reference string is uniformly distributed
(since soundness was stated for that case), and it is not clear whether soundness holds
when the reference string is generated by the simulator (who must use a different39
distribution). This issue is addressed by the notion of (weak) simulationsoundness.

Defining and Constructing Adaptive NIZKs with a Weak Simulation-Soundness

Property. This discussion leads to the following definition:

Definition 5.4.24 (weak simulation-soundness): Let (P, V ) be an adaptive NIZK for

a language L, and (S1 , S2 ) be a corresponding two-stage simulator. We say that weak
simulation-soundness holds if for all polynomial-size implementable functions  and
, it holds that

Pr [(r ) ∈ L and V ((r ), r, (r )) = 1, where (r, s) ← S1 (1n )] < µ(n)

where µ : N →[0,1] is a negligible function.

Note that the computational limitation on is essential to the viability of the definition
(see Exercise 40). It is tempting to conjecture that every adaptive NIZK (or rather its
simulator) satisfies weak simulation-soundness; however, this is not true (for further
discussion see Exercise 41). Nevertheless, adaptive NIZK (for N P) with a simulator

39 Indeed, prove that the distribution produced by the simulator must be far away from uniform. See related
Exercises 39 and 40.

456

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

satisfying weak simulation-soundness can be constructed given any adaptive NIZK (for
N P).

Construction 5.4.25 (from adaptive NIZK to weak simulation-soundness): Let (P, V )

be an adaptive NIZK for some language L, and let (S1 , S2 ) be the corresponding two-
stage simulator. We construct the following adaptive NIZK that works with reference
string ((r10 , r11 ), ..., (rn0 , rn1 )), where riσ ∈ {0, 1}n .

Prover P  : On common input x and auxiliary-input w (s.t., (x, w) ∈ R L ), (and ref-

erence string ((r10 , r11 ), ..., (rn0 , rn1 ))), uniformly select b1 , ..., bn ∈ {0, 1}, compute
def
πi ← P(x, w, ribi ) for i = 1, ..., n, and output π = (b1 , ..., bn , π1 , ..., πn ).
Verifier V  : On common input x (and reference string ((r10 , r11 ), ..., (rn0 , rn1 ))), given an
alleged proof π = (b1 , ..., bn , π1 , ..., πn ), accept if and only if V (x, ribi , πi ) = 1 for
each i ∈ {1, ..., n}.
Simulator’s first stage S1 : On input 1n , select uniformly c1 , ..., cn ∈ {0, 1}, generate
(rici , si ) ← S1 (1n ) for i = 1, ..., n, select uniformly r11−c1 , ..., rn1−cn ∈ {0, 1}n , and
def def
output (r , s), where r = ((r10 , r11 ), ..., (rn0 , rn1 )) and s = (c1 , ..., cn , s1 , ..., sn ).
Simulator’s second stage S2 : On input (s, x), where s = (c1 , ..., cn , s1 , ..., sn ), compute
πi ← S2 (x, si ) for i = 1, ..., n, and output (c1 , ..., cn , π1 , ..., πn ).

It is easy to see that Construction 5.4.25 preserves the adaptive NIZK features of
(P, V, S1 , S2 ). Furthermore, as will be shown, Construction 5.4.25 is weak simulation-
sound.

Proposition 5.4.26: Construction 5.4.25 is an adaptive NIZK for L, and weak

simulation-soundness holds with respect to the prescribed simulator.

Proof Sketch: Completeness and soundness follow by the corresponding properties

of (P, V ). To see that the simulation is indistinguishable from the real execution of
(P  , V  ), note that the two probability ensembles differ in two aspects: First, the simu-
lation uses rici ’s generated by S1 (1n ), whereas in the real execution, the rici ’s are uniformly
distributed; and second, the simulation uses simulated proofs produced by S2 (x, si ),
rather than real proofs produced by P(x, w, ribi ). Still, the indistinguishability of the
output of the original simulator from the real execution of (P, V ) can be used to prove
that the current ensembles are indistinguishable, too. Specifically, we consider a hybrid
distribution in which all rib ’s are generated by S1 (1n ) but the individual proofs (i.e.,
πi ’s) are produced by P(x, w, ribi ). Using the fact that indistinguishability (by small
circuits) is preserved under repeated sampling, we show that this hybrid ensemble is
indistinguishable from each of the two original ensembles (i.e., the real execution of
(P  , V  ) and the simulation by (S1 , S2 )).
To establish the weak simulation-soundness property, we consider an arbitrary cheat-
ing prover C = (, ) that is implementable by a family of small circuits. We say that
C(r ) = ((r ), (r )) succeeds if it holds that (r ) ∈ L and V  ((r ), r , (r )) = 1.
We are interested in the probability that C(r ) succeeds when (r , s) ← S1 (1n ). Recall
457
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

that s = (c1 , ..., cn , s1 , ..., sn ), where the ci ’s are selected uniformly in {0, 1}, whereas
(r ) has the form (b1 , ..., bn , π1 , ..., πn ). Let us denote the latter sequence of bi ’s by
B(r ); that is, (r ) = (B(r ),  (r )). We distinguish two cases according to whether or
def
not B(r ) = c = (c1 , ..., cn ):

Pr[C(r ) = ((r ), (B(r ), 

(r ))) succeeds, when (r , s) ← S1 (1n )]
= Pr[C(r ) succeeds and B(r ) = c, when (r , (c, s  )) ← S1 (1n )] (5.13)
+ Pr[C(r ) succeeds and B(r ) = c, when (r , (c, s  )) ← S1 (1n )] (5.14)

Eq. (5.13), which corresponds to the first case, must be negligible because the corre-
sponding probability that refers to a uniformly selected reference string (as appearing
in the real proof) is negligible, and the indistinguishability of a simulated reference
string from a uniformly distributed one was established previously.
Details: For a uniformly distributed reference string r , we have Pr[B(r ) = c] = 2−n
by information-theoretic considerations (i.e., r is statistically independent of c). On
the other hand, for a simulated reference string r and a corresponding c, the quantity
def
q = Pr[B(r ) = c] is lower-bounded by Eq. (5.13). The quality of the simulator’s
output (established in the first paragraph of the proof ) implies that the simulated
reference string is computationally indistinguishable from a uniformly distributed
reference string, which in turn implies that q − 2−n is negligible. It follows that
Eq. (5.13) is negligible.

Eq. (5.14) must be negligible because in this case, at least one of the alleged proofs (to
a false assertion) is with respect to a uniformly distributed reference string.
Details: By the case hypothesis (i.e., B(r ) = c), there exists an i such that the i-th bit
of B(r ) is different from ci (i.e., bi = ci ). Thus, the i-th alleged proof (i.e., πi ) is with
respect to a uniformly distributed reference string, that is, with respect to ribi = ri1−ci ,
where ri1−ci is selected uniformly in {0, 1}n . By the (adaptive) soundness of (P, V ),
this proof for a false assertion can be valid only with negligible probability, which
in turn implies that Eq. (5.14) is negligible.

Having established that both Eq. (5.13) and Eq. (5.14) are negligible, the proposition
follows.

Using Adaptive NIZKs with Weak Simulation-Soundness. Following the foregoing

motivating discussion, we show that if the adaptive NIZK used in Construction 5.4.23
has the weak simulation-soundness property, then the resulting encryption scheme
(G  , E  , D  ) is secure under a priori CCA.

Theorem 5.4.27: Suppose that the adaptive NIZK (P, V ) used in Construction 5.4.23
has the weak simulation-soundness property and that the public-key encryption scheme
(G, E, D) is passively secure in the key-dependent sense. Further suppose that the
probability that G(1n ) produces a pair (e, d) such that Pr[Dd (E e (x)) = x] < 1, for
some x ∈ {0, 1}poly(n) , is negligible. Then Construction 5.4.23 constitutes a public-key
encryption scheme that is secure under a priori CCA.
458

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

Combining the above with Theorem 4.10.16 and Proposition 5.4.26, it follows that
public-key encryption schemes that are secure under a priori CCA exist, provided that
enhanced40 trapdoor permutations exists.

Proof Sketch: Assuming toward the contradiction that the scheme (G  , E  , D  ) is not
secure under a priori CCA, we show that the scheme (G, E, D) is not secure under
a (key-dependent) passive attack. Specifically, we refer to the definitions of security
in the sense of indistinguishability of encryptions (as in Definitions 5.4.14 and 5.4.2,
respectively). To streamline the proof, we reformulate Definition 5.4.2, incorporating
both circuits (i.e., the one selecting message pairs and the one trying to distinguish their
encryptions) into one circuit and allow this circuit to be probabilistic. (Certainly, this
model of a key-dependent passive attack is equivalent to the one in Definition 5.4.2.)
Let (A1 , A2 ) be an a priori CCA adversary attacking the scheme (G  , E  , D  ) (as per
Definition 5.4.14), and (S1 , S2 ) be the two-stage simulator for (P, V ). We construct a
(key-dependent) passive adversary A (attacking (G, E, D)) that, given an encryption-
key e (in the range of G 1 (1n )), behaves as follows:

1. Initialization: A generates (e1 , d1 ) ← G(1n ), (r, s) ← S1 (n), and sets e = (e1 , e, r ).

def
(We assume that (e, d) ← G(1n ), and let (e2 , d2 ) = (e, d), so e = (e1 , e2 , r ).)
2. Emulation of A1 Dd (e): A invokes A1 on input e, and answers its (decryption) queries
as follows. When asked to decrypt the alleged ciphertext (q1 , q2 , q3 ), adversary A
checks if q3 is a valid proof of consistency of q1 and q2 (with respect to the reference
string r ). If the proof is valid, then A answers with Dd1 (q1 ) or else A returns the
error symbol.
(Note that the emulation of the oracle Dd by A is perfect, although A only knows
part of the corresponding decryption-key d. Also note that A emulates A1 on an
input and oracle access that are computationally indistringuishable from the input
and oracle access given to A1 in a real attack.)
3. Using A2 for the final decision: Let ((x (1) , x (2) ), σ ) denote the challenge template out-
put by A1 . Then, A outputs (x (1) , x (2) ) as its own challenge pair. Next, given a cipher-
text y = E e (x), where x ∈ {x (1) , x (2) }, adversary A forms a corresponding (almost
certainly illegal) ciphertext under E  , denoted (y1 , y, π), by letting y1 ← E e1 (0|x | )
(1)

and π ← S2 (s, (e1 , e, y1 , y)). Finally, A invokes A2 on input (σ, (y1 , y, π)), and out-
puts whatever the latter does. Recall that here (in the case of a priori CCA), A2 is an
ordinary machine (rather than an oracle machine).
(Note that A emulates A2 on an input that is computationally indistringuishable from
the input given to A2 in a real attack. In particular, A typically invokes A2 with an
illegal ciphertext, whereas in a real attack, A2 is always given a legal ciphertext.)

In order to analyze the performance of A, we introduce the following hybrid process,

denoted H , as a mental experiment. The hybrid process behaves as A, with the only
exception that (in Step 3) y1 ← E e1 (x) (rather than y1 ← E e1 (0|x| )). Thus, unlike A,

40 See Section C.1 in Appendix C

459
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

the hybrid process invokes A2 with a legal ciphertext. (The question of how the hybrid
process “knows” or gets this y1 is out of place; we merely define a mental experiment.)
( j) ( j) ( j) ( j)
Let p A = p A (n) (resp., p H = p H (n)) denote the probability that A (resp., the hybrid
process H ) outputs 1 when x = x ( j) , where the probability is taken over the choices of
(e, d) ← G(1n ) and the internal coin tosses of A (resp., H ).
( j) ( j)
Claim 5.4.27.1: For both j’s, the absolute difference between p A (n) and p H (n) is a
negligible function in n.

Proof: Define an auxiliary hybrid process that behaves as the hybrid process, except
that when emulating Dd , the auxiliary process answers according to Dd2 (rather than
( j)
according to Dd1 ). (Again, this is a mental experiment.) Let p H H denote the probability
that this auxiliary process outputs 1 when x = x ( j) . Similarly, define another mental
experiment that behaves as A, except that when emulating Dd , this process answers
( j)
according to Dd2 (rather than according to Dd1 ), and let p A A denote the probability that
the latter process outputs 1 when x = x ( j) . We stress that in Step 3, the latter mental
experiment behaves exactly like A; the only aspect in which this mental experiment
differs from A is in its decryption operations at Step 2. The various processes are
tabulated next.

answers the challenge nature of process

dec-queries ciphertext for A
A by using Dd1 (E e1 (0|x| ), E e (x), ·) a real (passive) attack
on (G, E, D) (w.r.t. key e)
H by using Dd1 (E e1 (x), E e (x), ·) a mental experiment
HH by using Dd2 (E e1 (x), E e (x), ·) a mental experiment
AA by using Dd2 (E e1 (0|x| ), E e (x), ·) a mental experiment

We establish the following facts regarding these processes:

( j) ( j)
Fact 1. For both j’s, the absolute difference between p H and p H H is negligible.
The reason is that the two processes differ only in the way they answer the decryption
queries: In the first process the decryption is according to Dd1 , and in the second
it is according to Dd2 . However, by weak simulation-soundness, it is infeasible to
produce triples (q1 , q2 , q3 ) such that (e1 , e, q1 , q2 ) ∈ L R and yet q3 is a valid proof
(with respect to r , for the false assertion that (e1 , e, q1 , q2 ) is in L R ). Thus, except
with negligible probability, either Dd1 (q1 ) = Dd2 (q2 ) or q3 is not valid, and so it does
not matter whether one decrypts according to Dd1 or to Dd2 .41
( j) ( j)
Fact 2. Similarly, for both j’s, the absolute difference between p A and p A A is
negligible.
( j) ( j)
Fact 3. Finally, for both j’s, the absolute difference between p H H and p A A is negligible.

41 Here, we rely on the hypothesis that except with negligible probability over the key-generation process, the
decryption is error-less (i.e., always yields the original plaintext).

460

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

The reason is that the experiments A A and H H differ only in the in-
put (σ, (y1 , y, π)) that they feed to A2 ; whereas A A forms y1 ← E e1 (0|x| )
(and π ← S2 (s, (e1 , e, y1 , y))), the process H H forms y1 ← E e1 (x) (and π ←
S2 (s, (e1 , e, y1 , y))). However, A2 cannot distinguish the two cases because this
would have violated the security of E e1 .
That is, to establish Fact 3, we construct a passive attack, denoted B, that behaves
similarly to A except that it switches its reference to the two basic keys (i.e., the
first two components of the encryption-key e) and acts very differently in Step 3
(e.g., B produces a different challenge template). Specifically, given an attacked
encryption-key e, adversary B generates (e2 , d2 ) ← G(1n ), sets e = (e, e2 , ·), and
emulates A1 Dd (e) using the decryption-key d2 to answer queries. For a fixed j, when
obtaining (from A1 ) the challenge template ((x (1) , x (2) ), σ ), adversary B produces the
challenge template ((0|x | , x ( j) ), σ ), and invokes A2 on input (σ, (y, y2 , π)), where
( j)

y = E e (x) (x ∈ {0|x | , x ( j) }) is the challenge ciphertext given to B, and B computes

( j)

y2 ← E e2 (x ( j) ) and π ← S2 (s, (e, e2 , y, y2 )). (Finally, B outputs the output obtained

from A2 .) Note that when given the challenge ciphertext E e (x ( j) ), the adversary
B effectively behaves as experiment H H (for the same j), whereas when given
E e (0|x | ), it effectively behaves as experiment A A (for the same j). Thus, if p H H
( j) ( j)
( j)
and p A A differ in a non-negligible manner, then B violates the passive security of
the encryption scheme (G, E, D).

Combining these three facts, the current claim follows.

Let us denote by pcca (n) the probability that the CCA adversary ( A1 , A2 ) outputs 1
( j)

when given a ciphertext corresponding to the j th plaintext in its challenge template (see
(1) (2)
Definitions 5.4.14). Recall that by the contradiction hypothesis, | pcca (n) − pcca (n)| is
not negligible.
( j) ( j)
Claim 5.4.27.2: For both j’s, the absolute difference between pcca (n) and p H (n) is a
negligible function in n.

Proof: The only difference between the output in a real attack of ( A1 , A2 ) and the output
of the hybrid process is that in the hybrid process, a “simulated reference string” and
a “simulated proof” are used instead of a uniformly distributed reference string and a
real NIZK proof. However, this difference is indistinguishable.42

Combining Claims 5.4.27.1 and 5.4.27.2, we obtain that for some negligible function
µ it holds that
(1) (2) (1) (2)
| p A (n) − p A (n)| > | p H (n) − p H (n)| − µ(n)
(1) (2)
> | pcca (n) − pcca (n)| − 2µ(n)
We conclude that (the passive attack) A violates the passive security of (G, E, D). This
contradicts the hypothesis (regarding (G, E, D)), and so the theorem follows.

42 We stress that the current claim relies only on the fact that the simulated reference-string and proof are indis-
tinguishable from the corresponding real objects.

461
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Step II: a-posteriori CCA

In order to use Construction 5.4.23 in the context of a posteriori CCA security, we
need to further strengthen the NIZK proof in use. The reason is that in an a posteriori
CCA, the adversary may try to generate proofs of false claims (as part of the ciphertext
queries in the second stage) after being given a (single) simulated proof (as part of the
challenge ciphertext). Specifically, when trying to extend the proof of Theorem 5.4.27,
we need to argue that, given a simulated proof (to either a false or a true statement), it is
infeasible to generate a proof to a false statement (so long as one does not just copy the
given simulated proof [in case it is to a false statement]). The notion of weak simulation-
soundness does not suffice to bound the probability of success in such attempts, because
the former notion refers to what one can do when only given the simulated reference
string (without a corresponding simulated proof). The following definition addresses
the situation in which one is given a single simulated proof (along with the simulated
reference string). (We comment that a more general notion that refers to a situation in
which one is given many simulated proofs is not necessary for the current application.)

Definition 5.4.28 (1-proof simulation-soundness): Let (P, V ) be an adaptive NIZK

for a language L, and (S1 , S2 ) be a corresponding two-stage simulator. We say that
1-proof simulation-soundness holds if for every triplet of polynomial-size circuit
families (1 , 2 , 2 ), the probability of the following event is negligible:
The event: For r and (x 1 , π 1 , x 2 , π 2 ) generated as described next, the following three
conditions hold: x 2 ∈ L, (x 2 , π 2 ) = (x 1 , π 1 ), and V (x 2 , r, π 2 ) = 1.
The generation process: First (r, s) ← S1 (1n ), then x 1 ← 1 (r ), next π 1 ← S2 (s, x 1 ),
and finally (x 2 , π 2 ) ← (2 (r, π 1 ), 2 (r, π 1 )).
That is, the adversary is represented by three circuits, and the process considered is as
follows. Given a simulated reference string r , the adversary selects an input x 1 , gets a
corresponding simulated proof π 1 , and tries to form a (valid with respect to r ) proof
π 2 for some no-instance x 2 . Note that x 1 is not required to be a yes-instance. In case
x 2 = x 1 , we consider only π 2 = π 1 (and in case x 2 = x 1 , we also consider π 2 = π 1 ).
Definition 5.4.28 requires that the success probability of any such feasible adversary
be negligible. Note that weak simulation-soundness is obtained as a special case of
Definition 5.4.28 (by setting (r ) = 2 (r, λ) and (r ) = 2 (r, λ), where λ denotes
the empty string).

Theorem 5.4.29: Suppose that the adaptive NIZK (P, V ) used in Construction 5.4.23
has the 1-proof simulation-soundness property and that the encryption scheme
(G, E, D) is as in Theorem 5.4.27. Then Construction 5.4.23 constitutes a public-key
encryption scheme that is secure under a posteriori CCA.

Proof Sketch: The proof follows the structure of the proof of Theorem 5.4.27. Specif-
ically, given an a posteriori CCA adversary ( A1 , A2 ) (attacking (G  , E  , D  )), we first
construct a passive adversary A (attacking (G, E, D)). The construction is as in the proof
of Theorem 5.4.27, with the exception that in Step 3 we need to emulate the decryption
oracle (for A2 ). This emulation is performed exactly as the one performed in Step 2
462

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

(for A1 ). Next, we analyze this passive adversary as in the proof of Theorem 5.4.27,
while referring to an A2 that may make decryption queries.43 The analysis of the hand-
ling of these (additional) queries relies on the 1-proof simulation-soundness property.
In particular, when proving a claim analogous to Claim 5.4.27.1, we have to establish
two facts (corresponding to Facts 1 and 2) that refer to the difference in the process’s
output when decrypting according to Dd1 and Dd2 , respectively. Both facts follow from
the fact (established next) that, except with negligible probability, neither A1 nor A2 can
produce a query (q1 , q2 , q3 ) such that q3 is a valid proof that q1 and q2 are consistent
and yet Dd1 (q1 ) = Dd2 (q2 ). (We stress that in the current context we refer also to
A2 , which may try to produce such a query based on the challenge ciphertext given
to it.)
Fact 5.4.29.1: The probability that A1 produces a query (q1 , q2 , q3 ) such that q3 is a
valid proof (with respect to reference string r ) that (supposedly) there exists x, s1 , s2
such that qi = E ei (x, si ) (for i = 1, 2), and yet Dd1 (q1 ) = Dd2 (q2 ) is negligible. The
same holds for A2 so long as the query is different from the challenge ciphertext given
to it. This holds regardless of whether the challenge ciphertext (given to A2 ) is produced
as in A (i.e., y1 = E e1 (0m )) or as in the hybrid process H (i.e., y1 = E e1 (x)).

Proof: Recall that one of our hypotheses is that the encryption (G, E, D) is error-free
(except for a negligible measure of the key-pairs). Thus, the current fact refers to a
situation that either A1 or A2 produces a valid proof for a false statement. The first part
(i.e., referring to A1 ) follows from the weak simulation-soundness of the NIZK, which
in turn follows from its 1-proof simulation-soundness property. We focus on the second
part, which refers to A2 .
Let (y1 , y2 , π ) denote the challenge ciphertext given to A2 ; that is, y2 = y is the
challenge ciphertext given to A(e) (or to H (e)), which augments it with y1 and
π ← S2 (s, (e1 , e2 , y1 , y2 )). Recall that (r, s) ← S1 (1n ) and that e2 = e. Suppose that
A2 produces a query (q1 , q2 , q3 ) as in the claim; that is, (q1 , q2 , q3 ) = (y1 , y2 , π), the
encryptions q1 and q2 are not consistent (with respect to e1 and e2 , respectively), and
def
yet V ((e1 , e2 , q1 , q2 ), r, q3 ) = 1. Specifically, it holds that x 2 = (e1 , e2 , q1 , q2 ) ∈ L R ,
where L R is as in Construction 5.4.23 (see Eq. (5.12)), and yet V (x 2 , r, q3 ) = 1 (i.e.,
def
π 2 = q3 is a valid proof of the false statement regarding x 2 ). Since (y1 , y2 , π) is
def
produced by letting π ← S2 (s, (e1 , e2 , y1 , y2 )), it follows that π 1 = π is a simu-
lated proof (with respect to the reference string r ) for the alleged membership of
def
x 1 = (e1 , e2 , y1 , y2 ) in L R , where (r, s) ← S1 (1n ). Furthermore, given such a proof
(along with the reference string r ), A2 produces a query (q1 , q2 , q3 ) that yields a pair
(x 2 , π 2 ), where π 2 = q3 , such that x 2 = (e1 , e2 , q1 , q2 ) ∈ L R and yet V (x 2 , r, π 2 ) = 1
and (x 2 , π 2 ) = (x 1 , π 1 ). Thus, using A1 and A2 (along with (G, E, D)), we obtain cir-
cuits 1 , 2 , 2 that violate the hypothesis that (S1 , S2 ) is 1-proof simulation-sound.
Details: On input a (simulated) reference string r , the circuit 1 selects (e1 , d1 )
and (e2 , d2 ) in the range of G(1n ), and emulates the execution of A1 Dd (e), where
e = (e1 , e2 , r ) and d = (d1 , d2 , r ). (Indeed, we fix the best possible choice of

43 Indeed, in the proof of Theorem 5.4.27, where ( A1 , A2 ) is an a priori CCA, A2 makes no such queries.

463
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

(e1 , d1 ) and (e2 , d2 ), rather than selecting both at random, and emulate the or-
acle Dd using d that is known to the circuit.) When A1 outputs a challenge
template, 1 emulates the selection of the challenge x, sets y1 ← E e1 (0|x| ) (or
y1 ← E e1 (x) when we argue about the hybrid process H ), y2 ← E e2 (x), and out-
def
puts x 1 = (e1 , e2 , y1 , y2 ). (Again, we may fix the best choice of x 1 , y1 , and y2 ,
rather than generating them at random.) The challenge ciphertext is formed by
augmenting y1 , y2 with π 1 ← S2 (s, x 1 ), where s is the auxiliary information gen-
erated by S(1n ) (i.e., (r, s) ← S(1n )). Next, we describe the circuits 2 and 2 ,
which obtain x 1 = (e1 , e2 , y1 , y2 ) (as produced by 1 ) along with a simulated proof
π 1 = S2 (s, x 1 ). On input a reference string r and x 1 , π 1 (as just discussed), these
circuits emulate A2 Dd (σ, (y1 , y2 , π 1 )), where σ is the state information generated
by A1 . For some i (fixed as the best choice), we consider the i-th decryption query
made during the emulation (i.e., we emulate the answers to previous queries by
emulating Dd ). Denoting this (i.e., i-th) query by (q1 , q2 , q3 ), the circuit 2 out-
def def
puts x 2 = (e1 , e2 , q1 , q2 ) and 2 outputs π 2 = q3 . Since (q1 , q2 , q3 ) = (y1 , y2 , π 1 ),
it follows that (x 2 , π 2 ) = ((e1 , e2 , q1 , q2 ), π 2 ) = ((e1 , e2 , y1 , y2 ), π 1 ) = (x 1 , π 1 ).
The event stated in the claim refers to the case that x 2 ∈ L R and yet π 2 is ac-
cepted as a proof (with respect to the reference string r ). But this event and the
current process are exactly as in the definition of 1-proof simulation soundness. We
stress that the argument applies to the process defined by the actual attack, as well
as to the process defined by the hybrid H . In the first case x 1 ∈ L R , whereas in the
second case x 1 ∈ L R , but 1-proof simulation soundness applies to both cases.

It follows that a query (q1 , q2 , q3 ) as in the claim can be produced only with negligible
probability.

Fact 5.4.29.1 implies (an adequate extension of) the first two facts in the proof of a
claim analogous to Claim 5.4.27.1. The third fact in that proof, as well as the proof of
the analogue of Claim 5.4.27.2, do not refer to the soundness of the NIZK-proofs, and
are established here exactly as in the proof of Theorem 5.4.27. The current theorem
follows.

Constructing Adaptive NIZK with 1-Proof Simulation-Soundness Property. We

construct the desired NIZK by using a standard (adaptive) NIZK proof, a weak form
of a signature scheme, and a specific commitment scheme. Since all ingredients can be
implemented using enhanced trapdoor permutations (see Definition C.1.1 in Appendix
C), we obtain:

Theorem 5.4.30: If there exist collections of (non-uniformly hard) enhanced trap-

door permutations, then every language in N P has an adaptive NIZK with 1-proof
simulation-soundness property.

Proof Sketch: Let L ∈ N P. We construct a suitable NIZK for L using the following
three ingredients:

1. An adaptive Non-Interactive Witness-Indistinguishable (NIWI ) proof, denoted

(P WI , V WI ), for a suitable language in N P. We stress that we mean a proof system
464

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

that operates with a reference string of length n and can be applied to prove (adap-
tively chosen) statements of length poly(n), where the adaptivity refers both to the
soundness and witness-indistinguishability requirements.
As shown in Section 4.10.3.2,44 the existence of enhanced trapdoor permutations
implies that every language in N P has an adaptive NIZK that operates with a
reference string of length n and can be applied to prove statements of length poly(n).
Indeed, in analogy to discussions in Section 4.6, any NIZK is a NIWI.
2. A super-secure one-time signature scheme, denoted (G OT , S OT , V OT ). Specifically,
one-time security (see Section 6.4.1) means that we consider only attacks in which
the adversary may obtain a signature to a single document of its choice (rather
than signatures to polynomially many documents of its choice). On the other hand,
super-security (see Section 6.5.2) means that the adversary should fail to produce a
valid document-signature that is different from the query-answer pair that appeared
in the attack. (We stress that unlike in ordinary security, the adversary is deemed
successful even if it produces a different signature to the same document for which
it has obtained a signature during the attack.)
By Theorem 6.5.2, super-secure one-time signature schemes can be constructed
on the basis of any one-way function. (If we were willing to assume the existence
of collision-free hashing functions, then we could have used instead the easier-to-
establish Theorem 6.5.1.)
3. A perfectly-binding commitment scheme, denoted C, as defined in Section 4.4.1,
with the following two additional properties: The first additional property is that
the commitment strings are pseudorandom; that is, the ensembles {C(x)}x∈{0,1}∗ and
{U|C(x)| }x∈{0,1}∗ are computationally indistinguishable. The second property is that
the support of C(Un ) is a negligible portion of {0, 1}|C(Un )| .
Using any collection of one-way permutations (e.g., the one in the hypothesis),
we may obtain the desired commitment scheme. Specifically, Construction 4.4.2
constitutes a commitment scheme that satisfies the pseudorandomness property (but
not the “negligible portion” property). To obtain the additional “negligible portion”
property, we merely let C(x) equal a pair of two independent commitments to x
(and it follows that the support of C(Un ) is at most a 2n · (2−n )2 = 2−n fraction of
{0, 1}|C(Un )| ).45 We denote by C(x, r ) the commitment to value x produced using
coins r ; that is, C(x) = C(x, r ), where r is uniformly chosen in {0, 1}(|x|) , for some
polynomial .

Given these ingredients, we construct an adaptive (1-proof simulation-sound) NIZK

for L (with witness relation R) as follows. The NIZK proof uses a reference string of
def def
the form r = (r1 , r2 ), where n = |r2 | and m = |r1 | = poly(n). (The length of r1 is set
to equal the length of C(v), where (s, v) ← G OT (1n ).)
44 See Theorem 4.10.16 and comment following it, as well as Sections C.1 and C.4 in Appendix C.
45 This presupposes that in the original commitment scheme, the support of C(x) is at most a 2−|x| fraction
of {0, 1}|C(x)|, which does hold for Construction 4.4.2. An alternative construction of a commitment scheme
satisfying both additional properties can be obtained using any one-way function. Specifically, Construction 4.4.4
will do, except that it uses two messages. However, since the first message (i.e., sent by the receiver) is a random
string, we may incorporate it in the reference string (of the NIZK scheme presented next).

465
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Prover P: On common input x ∈ {0, 1}poly(n) and auxiliary-input w (and reference

string r = (r1 , r2 )), where supposedly (x, w) ∈ R, the prover behaves as follows:

1. Generates a key-pair for the one-time signature scheme; that is, (s, v) ← G OT (1n ).
2. Computes a pre-proof p ← P WI ((x, r1 , v), w, r2 ), where (P WI , V WI ) is a proof
system (using r2 as reference string) for the following NP-language L  :

L  = {(x, y, v) : (x ∈ L) ∨ (∃w  y = C(v, w  ))}

def
(5.15)

The corresponding NP-relation is

R  = {((x, y, v), w  ) : ((x, w  ) ∈ R) ∨ (y = C(v, w  ))}

def
(5.16)

Note that P indeed feeds P WI with an adequate NP-witness (i.e., ((x, r1 , v), w)
∈ R  since (x, w) ∈ R). The first part of the reference string of P is part of the
statement fed to P WI , whereas the second part of P’s reference string
serves as a reference string for P WI . The behavior of V (with respect
to V WI ) will be analogous.
3. The prover computes a signature σ to (x, p) relative to the signing-key s (generated
in Step 1). That is, P computes σ ← SsOT (x, p).

The prover outputs the triplet (v, p, σ ).

Verifier V : On common input x and an alleged proof (v, p, σ ) (and reference string
r = (r1 , r2 )), the verifier accepts if and only if the following two conditions hold:

1. σ is a valid signature, with respect to the verification-key v, of the pair (x, p).
That is, VvOT ((x, p), σ ) = 1.
2. p is a valid proof, with respect to the reference string r2 , of the statement
(x, r1 , v) ∈ L  . That is, V WI ((x, r1 , v), r2 , p) = 1.

Simulator’s first stage S1 : On input 1m+n (from which S1 determines n and m), the
first stage produces a reference string and auxiliary information as follows:

1. Like the real prover, S1 (1m+n ) starts by generating a key-pair for the one-time
signature scheme; that is, (s, v) ← G OT (1n ).
2. Unlike in the real setting, S1 (1m+n ) selects s1 uniformly in {0, 1}(|v|) , and sets r1 =
C(v, s1 ). (Note that in the real setting, r1 is uniformly distributed independently
of v, and thus in the real setting, r1 is unlikely to be in the support of C, let alone
in that of C(v).)
3. Like in the real setting, S1 (1m+n ) selects r2 uniformly in {0, 1}n .

S1 (1m+n ) outputs the pair (r , s), where r = (r1 , r2 ) is a simulated reference string
and s = (v, s, s1 , r2 ) is auxiliary information to be passed to S2 .
Simulator’s second stage S2 : On input a statement x and auxiliary input s =
(v, s, s1 , r2 ) (as generated by S1 ), S2 proceeds as follows:
466

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

1. Using (the NP-witness) s1 , the simulator computes a pre-proof p ←

P WI ((x, C(v, s1 ), v), s1 , r2 ). Note that indeed, ((x, C(v, s1 ), v), s1 ) ∈ R  .
2. Using (the signing-key) s, the simulator computes a signature σ to (x, p) relative
to s, where p is as computed in the first step. That is, σ ← SsOT (x, p).
S2 (s, x) outputs (v, p, σ ) as a simulated proof (with respect to r ) for membership of
x in L.
As we will show in Claim 5.4.30.2, the above (two-stage) simulator produces output that
is indistinguishable from the output of the real execution. Intuitively, the first stage of
the simulator enables cheating by entities (such as the second stage of the simulator) that
can produce signatures with respect to the verification-key committed to in the string r1
(which is part of the reference string generated by S1 ). This allows the simulation (which
gets the signing-key) to cheat, but does not allow cheating by an adversary that sees
only the verification-key as well as a single valid signature (which are both part of the
single proof given to the adversary in the definition of 1-proof simulation-soundness).
We now turn to the actual proof of these properties.
Claim 5.4.30.1: (P, V ) satisfies completeness and adaptive soundness.

Proof: Completeness follows by combining the syntactic properties of the one-time

signature scheme, the completeness property of the proof system (P WI , V WI ), and the
definition of R  . Adaptive soundness follows by combining the (adaptive) soundness of
(P WI , V WI ) with the fact that r1 is unlikely to be a commitment to any string. Specif-
n
ically, using the additional property by which C(G OT 2 (1 )) covers a negligible portion
of {0, 1}m , it follows that for a uniformly selected r1 ∈ {0, 1}m , there exist no v such
that r1 is in the support of C(v). Thus, except with negligible probability (over the
random choice of r1 ), if (x, r1 , v) ∈ L  holds for some v, then x ∈ L. On the other
hand, using the (adaptive) soundness of (P WI , V WI ), except with negligible probabil-
ity (over the random choice of r2 ), the existence of a valid proof (v, p, σ ) for some
x ∈ {0, 1}poly(n) implies that (x, r1 , v) ∈ L  . Thus, for a uniformly distributed refer-
ence string r = (r1 , r2 ) ∈ {0, 1}m+n , except with negligible probability, there exists no
x ∈ {0, 1}poly(n) \ L and π such that V (x, r , π) = 1. The claim follows.
Claim 5.4.30.2 (adaptive zero-knowledge): For every efficient way of selecting in-
puts , the output produced by the two-stage simulator (S1 , S2 ) is indistinguish-
def
able from the one produced by P. That is, the ensembles {S  (1m+n )} and R , W =
{(Um+n , (Um+n ), P((Um+n ), W (Um+n ), Um+n ))} are computationally indistinguish-
able, where S  is defined as in Definition 5.4.22.

Proof: Consider a hybrid distribution H  (1m+n ), in which everything except the pre-
proof is produced as by S  (1m+n ), and the pre-proof is computed as by the real prover.
That is, (r , s) ← S1 (1m+n ) (where r = (r1 , r2 ) and s = (v, s, s1 , r2 )) is produced as by
S  , but then for (x, w) = ((r ), W (r )), the pre-proof is computed using the witness
w; that is, p ← P WI ((x, r1 , v), w, r2 ), rather than p ← P WI ((x, r1 , v), s1 , r2 ). The final
proof π = (v, p, σ ) is obtained (as in both cases) by letting σ ← SsOT (x, p). We now
relate the hybrid ensemble to each of the two ensembles referred to in the claim.
467
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

1. By the (adaptive) witness-indistinguishability of P WI , the ensembles H  and S  are

computationally indistinguishable. (Recall that these ensembles differ only in the
way the pre-proof is produced; specifically, they differ only in the NP-witness used
by P WI to prove the very same claim.)
2. By the pseudorandomness of the commitments produced for any fixed value, H 
and R , W are computationally indistinguishable. (Recall that these ensembles differ
only in the way the first part of the reference string (i.e., r1 ) is produced.)

The claim follows.

Claim 5.4.30.3 (1-proof simulation-soundness): For every triplet of polynomial-size
circuit families (1 , 2 , 2 ), consider the following process: First (r , s) ← S1 (1m+n ),
then x 1 ← 1 (r ), next π 1 ← S2 (s, x 1 ), and finally (x 2 , π 2 ) ← (2 (r , π 1 ), 2 (r , π 1 )).
Then, the probability that the following three conditions hold simultaneously is negli-
gible: (1) x 2 ∈ L, (2) (x 2 , π 2 ) = (x 1 , π 1 ), and (3) V (x 2 , r , π 2 ) = 1.

Proof: Recall that r = (r1 , r2 ) and s = (v, s, s1 , r2 ), where (s, v) ← G OT (1n ) and r1 =
C(v, s1 ) for a uniformly chosen s1 ∈ {0, 1}(|v|) (and r2 is selected uniformly in {0, 1}n ).
Also recall that π 1 = (v 1 , p 1 , σ 1 ), where v 1 = v, p 1 ← P WI ((x, C(v, s1 ), v), s1 , r2 )
def
and σ 1 ← SsOT (x 1 , p 1 ). Let us denote (v 2 , p 2 , σ 2 ) = π 2 . We need to upper-bound the
following:


Pr (x 2 ∈ L) ∧ ((x 2 , π 2 ) = (x 1 , π 1 )) ∧ (V (x 2 , r , π 2 ) = 1)
⎡ 2 ⎤
(x ∈ L) ∧ ((x 2 , π 2 ) = (x 1 , π 1 ))
⎢ ⎥
= Pr ⎣ ∧ (VvOT 2 ((x , p ), σ ) = 1) ⎦
2 2 2
(5.17)
∧ (V ((x , r1 , v ), r2 , p ) = 1)
WI 2 2 2

where the equality is due to the definition of V . We consider two cases (in which the
event in Eq. (5.17) may hold):
v2 = v 1 : In this case, either (x 2 , p2 ) = (x 1 , p 1 ) or σ 2 = σ 1 must hold (because other-
wise (x 2 , π 2 ) = (x 2 , (v 2 , p2 , σ 2 )) = (x 1 , (v 1 , p 1 , σ 1 )) = (x 1 , π 1 ) follows). But this
means that (2 , 2 ), given a single valid signature σ 1 (to the document (x 1 , p 1 )) with
respect to a randomly generated verification-key v = v 1 = v 2 , is able to produce a
valid document-signature pair ((x 2 , p 2 ), σ 2 ) (with respect to the same verification-
key) such that ((x 2 , p 2 ), σ 2 ) = ((x 1 , p1 ), σ 1 ), in contradiction to the super-security
of the one-time signature scheme.
Details: It suffices to upper-bound

(v 2 = v 1 ) ∧ ((x 2 , π 2 ) = (x 1 , π 1 ))
Pr (5.18)
∧ (VvOT
2 ((x , p ), σ ) = 1)
2 2 2

As explained in the previous paragraph, the first two conditions in Eq. (5.18)
imply that ((x 2 , p2 ), σ 2 ) = ((x 1 , p1 ), σ 1 ). Using (S1 , S2 ) and (1 , 2 , 2 ), we
derive an attacker, A, that violates the super-security of the (one-time) signa-
ture scheme. The attacker just emulates the process described in the claim’s
hypothesis, except that it obtains v as input (rather than generating the pair (s, v)

468

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

by invoking G OT ) and uses oracle access to SsOT (rather than s itself ) in order
to produce the signature σ 1 . Specifically, on input v, the attacker A first se-
lects s1 ∈ {0, 1} and r2 ∈ {0, 1}n uniformly, sets r1 = C(v, s1 ) and r = (r1 , r2 ),
and obtains x 1 ← 1 (r ). Next, A computes p 1 ← P WI ((x 1 , r1 , v), s1 , r2 ) and
queries SsOT on (x 1 , p1 ), obtaining the answer σ 1 ← SsOT (x 1 , p1 ) and setting
π 1 = (v, p 1 , σ 1 ). (Indeed, π 1 so produced is distributed exactly as S2 (s, x 1 ),
where s = (v, s, s1 , r2 ), although A does not know s; the argument relies on
the fact that S2 (s, x 1 ) can be implemented without knowledge of s and while
making a single query to the signing oracle SsOT .) Finally, A sets (x 2 , π 2 ) ←
(2 (r , π 1 ), 2 (r , π 1 )), and outputs ((x 2 , p2 ), σ 2 ), where π 2 = (v 2 , p2 , σ 2 ).
Note that A queries its signing oracle only once. (Recall that A queries SsOT
on (x 1 , p1 ), obtains the answer σ 1 , and produces the output pair ((x 2 , p2 ), σ 2 ).)
On the other hand, the probability that A produces a valid document-signature
pair (with respect to the verification-key v) that is different from the (single)
query-answer pair it makes equals Eq. (5.18). Thus, the super-security of the
one-time signature scheme implies that Eq. (5.18) is negligible.

v2 =
 v 1 : Since r1 = C(v 1 , s1 ), it follows (by the perfect binding property of C) that r1
is not in the support of C(v 2 ) (i.e., for every w  , r1 = C(v 2 , w  )). Thus, if x 2 ∈ L, then
(x 2 , r1 , v 2 ) ∈ L  . Now, by the adaptive soundness of (P WI , V WI ) and the fact that r2
was selected uniformly in {0, 1}n , it follows that, except with negligible probability,
p2 is not a valid proof (with respect to the reference string r2 ) of the false statement
“(x 2 , r1 , v 2 ) ∈ L  .”

Details: It suffices to upper-bound

(v 2 = v 1 ) ∧ (x 2 ∈ L)
Pr (5.19)
∧ (V WI ((x 2 , r1 , v 2 ), r2 , p2 ) = 1)

As explained in the previous paragraph, the first two conditions in Eq. (5.19)
imply (x 2 , r1 , v 2 ) ∈ L  . The key observation is that r2 (generated by S1 ) is uni-
formly distributed in {0, 1}n , and thus the adaptive soundness of the NIWI system
applies. We conclude that Eq. (5.19) is upper-bounded by the (negligible) sound-
ness error of the NIWI system, and the claim follows also in this case.

Combining both cases, the claim follows.

Combining Claims 5.4.30.1–5.4.30.3, the current theorem follows.

Conclusion. Combining Theorems 5.4.6, 5.4.30 and 5.4.29, we get:

Theorem 5.4.31: If there exist collections of (non-uniformly hard) enhanced trapdoor

permutations, then there exist public-key encryption schemes that are secure under a
posteriori chosen ciphertext attacks.

(See Section C.1 in Appendix C for a discussion of the notion of enhanced trapdoor
permutations.)
469
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

5.4.5. Non-Malleable Encryption Schemes

So far, our treatment has referred to an adversary that, when given a ciphertext, tries to
gain explicit information about the plaintext. A less explicit gain, captured by the so-
called notion of malleability, is the ability to generate an encryption of a related plaintext
(possibly without learning anything about the original plaintext). Loosely speaking, an
encryption scheme is called non-malleable if, given a ciphertext, it is infeasible (for an
adversary) to produce a (different) valid ciphertext for a related plaintext. For example,
given a ciphertext of a plaintext of the form 1x, for an unknown x, it should be infeasible
to produce a ciphertext to the plaintext 0x.
Non-malleability may relate to any of the types of attacks considered earlier (e.g.,
passive attacks, chosen ciphertext attacks, etc). Thus, we have a “matrix” of adversaries,
with one dimension (parameter) being the type of attack and the second being its
purpose. So far, we have discussed the first dimension (i.e., the type of the attack) when
focusing on a particular purpose (i.e., of violating the secrecy of the plaintext). We
now turn to the second dimension (i.e., the purpose of the attack) and consider also
the purpose of malleability. That is, we make a distinction between the following two
notions (or purposes of attack):

1. Standard security: the infeasibility of obtaining information regarding the plaintext.

As defined in Section 5.2, such information is captured by a function of the bare
plaintext,46 and it may not depend on the encryption-key (or decryption-key).
2. In contrast, the notion of non-malleability refers to the generating of a string depend-
ing on both the plaintext and the current encryption-key. Specifically, one requires
that it be infeasible for an adversary, given a ciphertext, to produce a valid ciphertext
(under the same encryption-key) for a related plaintext.

We shall show that with the exception of passive attacks on private-key schemes,
non-malleability always implies security against attempts to obtain information on the
plaintext. We shall also show that security and non-malleability are equivalent under
a posteriori chosen ciphertext attack. Thus, the results of the previous sections imply
that non-malleable (under a posteriori chosen ciphertext attack) encryption schemes
can be constructed based on the same assumptions used to construct passively secure
encryption schemes.

5.4.5.1. Definitions
For the sake of brevity, we present only a couple of definitions. Specifically, focusing
on the public-key model, we consider only the simplest and strongest types of attacks;
that is, we first consider (key-oblivious) passive attacks, and then we turn to chosen
ciphertext attacks. The definitions refer to an adversary that is given a ciphertext and
tries to generate a (different) ciphertext to a plaintext related to the original one. That
is, given E e (x), the adversary tries to output E e (y) such that (x, y) ∈ R with respect to

46 Note that considering a randomized process applied to the plaintext does not make the definition stronger.

470

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

some (efficiently recognizable)47 relation R. Loosely speaking, the adversary’s success

probability in such an attempt is compared to the success probability of generating
such E e (y) when given e but not E e (x). In fact, we prefer an equivalent formulation in
which the latter algorithm is required to output the plaintext y itself.48 As in the case
of semantic security, we strengthen the definition by considering all possible partial
information functions h.

Definition 5.4.32 (passive non-malleability): A public-key encryption scheme

(G, E, D) is said to be non-malleable under passive attacks if for every probabilistic
polynomial-time algorithm A there exists a probabilistic polynomial-time algorithm A
such that for every ensemble {X n }n∈N , with |X n | = poly(n), every polynomially bounded
h : {0, 1}∗ → {0, 1}∗, every polynomially bounded relation R that is recognizable by a
(non-uniform) family of polynomial-size circuits, every positive polynomial p, and all
sufficiently large n, it holds that
⎡ ⎤
(x, y) ∈ R where
⎢ (e, d) ← G(1n ) and x ← X n ⎥
Pr ⎢ ⎥
⎣  |x|
c ← E e (x) and c ← A(e, c, 1 , h(x)) ⎦
  |x|
y ← Dd (c ) if c = c and y ← 0 otherwise
⎡ ⎤
(x, y) ∈ R where
< Pr ⎣ x ← Xn ⎦+ 1
p(n)
y ← A (1n , 1|x| , h(x))

We stress that the definition effectively prevents the adversary A from just outputting
the ciphertext given to it (because in this case, its output is treated as if it were E e (0|x| )).
This provision is important because otherwise no encryption scheme could have satis-
fied the definition (see Exercise 42). A more subtle issue, which was hand-waved in the
definition, is how to handle the case in which A produces an illegal ciphertext (i.e., is
y defined in such a case to be a standard string [e.g., 1|d| ] or a special error symbol).49
The rest of our text holds under both conventions. Note that A can certainly produce
plaintexts, but its information regarding X n is restricted to h(X n ) (and 1|X n | ). Thus, if
when given h(X n ) and 1|X n | it is infeasible to generate y such that (X n , y) ∈ R, then A
as in Definition 5.4.32 may produce such a y only with negligible probability. Conse-
quently, Definition 5.4.32 implies that in this case, given E e (X n ) (and e, h(X n ), 1|X n | ),
it is infeasible to produce E e (y) such that (X n , y) ∈ R.

47 The computational restriction on R is essential here; see Exercise 16, which refers to a related definition of
semantic security.
48 Potentially, this can only make the definition stronger, because the ability to produce plaintexts implies the

ability to produce corresponding ciphertexts (with respect to a given or a randomly chosen encryption-key).
49 It is interesting to note that in the case of passive attacks, the two possible conventions seem to yield non-

equivalent definitions. The issue is whether the adversary can correlate the generation of an illegal ciphertext
to the encrypted plaintext handed to it. The question of whether this issue is important or not seems to depend
on the type of application. (In contrust, in the case of a posteriori CCA, the two conventions yield equivalent
definitions, because without loss of generality, the attacker may check whether the ciphertext produced by it is
legal.)

471
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Definition 5.4.32 cannot be satisfied by encryption schemes in which one can modify
bits in the ciphertext without changing the corresponding plaintext (i.e., consider the
identity relation). We stress that such encryption schemes may be semantically secure
under passive attacks (e.g., given a semantically secure encryption scheme (G, E, D),
consider E e (x) = E e (x)σ , for randomly chosen σ ∈ {0, 1}). However, such encryption
schemes may not be (semantically) secure under a posteriori CCA.
Turning to the definition of non-malleability under chosen ciphertext attacks, we
adopt the definitional framework of Section 5.4.4.1. Specifically, analogous to Defini-
tion 5.4.13, the challenge template produced by A1 (and A1 ) is a triplet of circuits rep-
resenting a distribution S (represented by a sampling circuit), a function h (represented
by an evaluation circuit), and a relation R (represented by a membership recognition
circuit). The goal of A2 (and A2 ) will be to produce a ciphertext of a plaintext that is
R-related to the challenge plaintext S(Upoly(n) ).

Definition 5.4.33 (non-malleability under chosen ciphertext attacks): A public-key en-

cryption scheme is said to be non-malleable under a priori chosen ciphertext attacks
if for every pair of probabilistic polynomial-time oracle machines, A1 and A2 , there
exists a pair of probabilistic polynomial-time algorithms, A1 and A2 , such that the
following two conditions hold:

1. For every positive polynomial p and all sufficiently large n and z ∈ {0, 1}poly(n) :
⎡ ⎤
(x, y) ∈ R where
⎢ (e, d) ← G(1n ) ⎥
⎢ ⎥
⎢ ((S, h, R), σ ) ← A E e , Dd
(e, z) ⎥
⎢
Pr ⎢ 1 ⎥
(c, v) ← (E (x), h(x)) , where x ← S(U ) ⎥
⎢ e poly(n) ⎥
⎣  Ee
c ← A2 (σ, c, v) ⎦
  |x|
y ← Dd (c ) if c = c and y ← 0 otherwise.
⎡ ⎤
(x, y) ∈ R where
⎢ ((S, h, R), σ ) ← A1 (1n , z) ⎥ ⎥+ 1
< Pr ⎢
⎣ ⎦ p(n)
x ← S(Upoly(n) )
 |x|
y ← A2 (σ, 1 , h(x))
2. For every n and z, the first element (i.e., the (S, h, R) part) in the random variables
E G (1n )
A1 (1n , z) and A1 1 (G 1 (1n ), z) are identically distributed.

Non-malleability under a posteriori chosen ciphertext attacks is defined analogously,

except that A2 is given oracle access to both E e and Dd , with the restriction that when
given the challenge (c, v), machine A2 is not allowed to make the query c to the
oracle Dd .

We comment that the definitional treatment can be extended to multiple-message non-

malleability, but we refrain from doing so here.50

50 We warn that even in the case of public-key schemes, (single-message) non-malleability (under some type of
attacks) does not necessarily imply the corresponding notion of multiple-message non-malleability.

472

www.Ebook777.com
 5.4* BEYOND EAVESDROPPING SECURITY

5.4.5.2. Relation to Semantic Security

With the exception of passive attacks on private-key schemes, for each type of attack
considered in this chapter (and for both private-key and public-key schemes), non-
malleability under this type of attack implies semantic security under the same type.
For example, we show the following:

Proposition 5.4.34: Let (G, E, D) be a public-key encryption scheme that is non-

malleable under passive attacks (resp., under a posteriori chosen ciphertext attacks).
Then, (G, E, D) is semantically secure under passive attacks (resp., under a posteriori
chosen ciphertext attacks).

Proof Sketch: For clarity, the reader may consider the case of passive attacks, but the
same argument holds also for a posteriori chosen ciphertext attacks. Furthermore, the
argument only relies on the hypothesis that (G, E, D) is “non-malleable with respect
to a single (simple) relation.”51
Suppose (toward the contradiction) that (G, E, D) is not semantically secure (under
the relevant type of attack). Using the equivalence to indistinguishability of encryptions,
it follows that under such attacks, one can distinguish encryption to x n from encryption
to yn . Consider the relation R = {(x, x̄) : x ∈ {0, 1}∗ }, where x̄ is the complement of
x, and the uniform distribution Z n on {x n , yn }. We construct an algorithm that, given
a ciphertext (as well as an encryption-key e), runs the said distinguisher and produces
E e (x̄n ) in case the distinguisher “votes” for x n (and produces E e ( ȳn ) otherwise). Indeed,
given E e (Z n ), our algorithm outputs E e ( Z̄ n ) (and thus “hits” R) with probability that is
non-negligibly higher than 1/2. This performance cannot be met by any algorithm that
is not given E e (Z n ). Thus, we derive a contradiction to the hypothesis that (G, E, D)
is non-malleable.

We stress that this argument relies only on the fact that in the public-key model, we
can produce the encryption of any string, since we are explicitly given the encryption-
key. In fact, it suffices to have access to an encryption oracle, and thus the argument
extends also to active attacks in the private-key model (in which the attacker is allowed
encryption queries). On the other hand, under most types of attacks considered here,
non-malleability is strictly stronger than semantic security. Still, in the special case of
a posteriori chosen ciphertext attacks, the two notions are equivalent. Specifically, we
prove that in the case of a posteriori CCA, semantic security implies non-malleability.

Proposition 5.4.35: Let (G, E, D) be a public-key encryption scheme that is seman-

tically secure under a posteriori chosen ciphertext attacks. Then, (G, E, D) is non-
malleable under a posteriori chosen ciphertext attacks. The same holds for private-key
encryption schemes.

Proof Sketch: Suppose toward the contradiction that (G, E, D) is not non-malleable
under a posteriori chosen ciphertext attacks, and let A = (A1 , A2 ) be an adversary
demonstrating this. We construct a semantic-security (a posteriori CCA) adversary
51 In order to avoid certain objections, we refrain from using the simpler relation R = {(x, x) : x ∈ {0, 1}∗ }.

473
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

B = (B1 , B2 ) that emulates A (while using its own oracles) and produces its own
output by querying its own decryption oracle on the ciphertext output by A, which is
assumed (without loss of generality) to be different from the challenge ciphertext given
to A. The key point is that B can make this extra query because it is an a posteriori
CCA adversary, and thus the difference between outputting a ciphertext and outputting
the corresponding plaintext disappears. Intuitively, B violates semantic security (with
respect to relations and a posteriori CCA, as can be defined analogously to Exercise 16).
Details follow.
Given an encryption-key e, algorithm B1 invokes A1 (e), while answering A1 ’s queries
by querying its own oracles, and obtains the challenge template (S, h, R) (and state σ ),
which it outputs as its own challenge template. Algorithm B2 is given a ciphertext
c (along with the adequate auxiliary information) and invokes A2 on the very same
input, while answering A2 ’s queries by querying its own oracles. When A2 halts with
output c = c, algorithm B2 forwards c to its decryption oracle and outputs the answer.
Thus, for every relation R, the plaintext output by B “hits” the relation R with the
same probability that the decryption of A’s output “hits” R. We have to show that this
hitting probability cannot be met by a corresponding benign algorithm that does not
get the ciphertext; but this follows from the hypothesis regarding A (and the fact that
in both cases, the corresponding benign algorithm [i.e., A or B  ] outputs a plaintext
[rather than a ciphertext]). Finally, we have to establish, analogously to Exercise 16,
that semantic security with respect to relations holds (in our current context of chosen
ciphertext attacks) if and only if semantic security (with respect to functions) holds.
The latter claim follows as in Exercise 16 by relying on the fact that in the current
context, the relevant relations have polynomial-size circuits. (A similar argument holds
for private-key schemes.)

Conclusion. Combining Theorem 5.4.31 and Proposition 5.4.35 we get:

Theorem 5.4.36: If there exist collections of (non-uniformly hard) enhanced trapdoor

permutations, then there exist public-key encryption schemes that are non-malleable
under a posteriori chosen ciphertext attacks.

Analogously, using Theorem 5.4.21, we get:

Theorem 5.4.37: If there exist (non-uniformly hard) one-way functions, then there
exist private-key encryption schemes that are non-malleable under a posteriori chosen
ciphertext attacks.

5.5. Miscellaneous

5.5.1. On Using Encryption Schemes

Once defined and constructed, encryption schemes may be (and actually are) used as
building blocks toward various goals that are different from the original motivation.
474

www.Ebook777.com
 5.5 MISCELLANEOUS

Still, the original motivation (i.e., secret communication of information) is of great

importance, and in this section we discuss several issues regarding the use of encryption
schemes toward achieving this goal.

Using Private-Key Schemes: The Key-Exchange Problem. As discussed in Sec-

tion 5.1.1, using a private-key encryption scheme requires the communicating parties
to share a secret key. This key can be generated by one party and secretly communicated
to the other party by an alternative (expensive) secure channel. Often, a preferable solu-
tion consists of employing a key-exchange (or rather key-generation) protocol, which
is executed over the standard (insecure) communication channel. An important distinc-
tion refers to the question of whether the insecure communication channel, connecting
the legitimate parties, is tapped by a passive adversary or may even be subject to active
attacks in which an adversary may modify the messages sent over the channel (and even
delete and insert such messages). Protocols that are secure against passive (resp., ac-
tive) adversaries are often referred to by the term authenticated key-exchange (resp.,
unauthenticated key-exchange), because in the passive case, one refers to the mes-
sages received over the channel as being authentic (rather than possibly modified by
the adversary).
A simple (generic) authenticated key-exchange protocol consists of using a public-
key encryption scheme in order to secretly communicate a key (for the private-key en-
cryption scheme, which is used in the actual communication).52 Specifically, one party
generates a random instance of a public-key encryption scheme, sends the encryption-
key to the other party, which generates a random key (for the private-key encryption
scheme), and sends an encryption (using the received encryption-key) of the newly
generated key to the first party. A famous alternative is the so-called Diffie-Hellman
Key-Exchange [75]: For a (large) prime P and primitive element g, which are universal
or generated on the fly (by one party that openly communicates them to the other), the
first (resp., second) party uniformly selects x ∈ ZP (resp., y ∈ ZP ) and sends g x mod P
(resp., g y mod P) to the other party, and both parties use g x y mod P as their common
key, relying on the fact that g x y ≡ (g x mod P) y ≡ (g y mod P)x (mod P). (The secu-
rity of this protocol relies on the assumption that given a prime P, a primitive element g,
and the triplet (P, g, (g x mod P), (g y mod P), (g z mod P)), it is infeasible to decide
whether or not z ≡ x y (mod P − 1), for x, y, z ∈ ZP .) The construction of unau-
thenticated key-exchange protocols is far more complex, and the interested reader is
referred to [29, 30, 15].

Using State-Dependent Private-Key Schemes. In many communication settings, it

is reasonable to assume that the encryption device may maintain (and modify) a state
(e.g., a counter). In such a case, the stream-ciphers discussed in Section 5.3.1 become
relevant. Furthermore, using a stream-cipher is particularly appealing in applications
where decryption is performed in the same order as encryption (e.g., in fifo commu-
nication). In such applications, the stream-cipher of Construction 5.3.3 is preferable to

52 One reason not to use the public-key encryption scheme itself for the actual (encrypted) communication is that
private-key encryption schemes tend to be much faster.

475
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

the (pseudorandom function-based) encryption scheme of Construction 5.3.9 for a cou-

ple of reasons. First, applying an on-line pseudorandom generator is likely to be more
efficient than applying a pseudorandom function. Second, for an -bit long counter (or

random value), Construction 5.3.3 allows for securely encrypting √2 messages (or bits),
whereas Construction 5.3.9 definitely becomes insecure when 2 messages (or bits)
are encrypted. For small values of  (e.g.,  = 64), this difference is crucial.

Using Public-Key Schemes: Public-Key Infrastructure. As in the case of private-

key schemes, an important distinction refers to the question of whether the insecure
communication channel between the legitimate parties is tapped by a passive adversary
or may even be subject to active attacks. In typical applications of public-key encryption
schemes, the parties communicate through a communication network (and not via a
point-to-point channel), in which case active attacks are very realistic (e.g., it is easy to
send mail over the Internet pretending to be somebody else). Thus, the standard use of
public-key encryption schemes in real-life communication requires a mechanism for
providing the sender with the receiver’s authentic encryption-key (rather than trusting an
“unauthenticated” incoming message to specify an encryption-key). In small systems,
one may assume that each user holds a local record of the encryption-keys of all
other users. However, this is not realistic in large-scale systems, and so the sender
must obtain the relevant encryption-key on the fly in a “reliable” way (i.e., typically,
certified by some trusted authority). In most theoretical work, one assumes that the
encryption-keys are posted and can be retrieved from a public-file that is maintained
by a trusted party (which makes sure that each user can post only encryption-keys
bearing its own identity). Alternatively, such a trusted party may provide each user with
a (signed) certificate stating the authenticity of the user’s encryption-key. In practice,
maintaining such a public-file (and/or handling such certificates) is a major problem, and
mechanisms that implement this abstraction are typically referred to by the generic term
“public-key infrastructure” (PKI). For a discussion of the practical problems regarding
PKI deployment see, e.g., [149, Chap. 13].

5.5.2. On Information-Theoretic Security

In contrast to the bulk of our treatment, which focuses on computationally bounded ad-
versaries, in this section we consider computationally unbounded adversaries. We stress
that also in this case, the length (and number) of the plaintexts is bounded. The result-
ing notion of security is the one suggested by Shannon: A (private-key or public-key)
encryption scheme is called perfectly secure (or information-theoretically secure) if
the ciphertext yields no information regarding the plaintext. That is, perfect-security
is derived from Definitions 5.2.1 and 5.2.2 by allowing computationally unbounded
algorithms (in the roles of A and A ).
It is easy to see that no public-key encryption scheme may be perfectly secure:
A computationally unbounded adversary that is given a encryption-key can find a
corresponding decryption-key, which allows it to decrypt any ciphertext.
In contrast, restricted types of private-key encryption schemes may be perfectly se-
cure. Specifically, the traditional “one-time pad” yields such a (private-key) scheme,
476

www.Ebook777.com
 5.5 MISCELLANEOUS

which can be used to securely communicate an a priori bounded number of bits. Fur-
thermore, multiple messages may be handled provided that their total length is a priori
bounded and that we use a state (as in Construction 5.3.3). We stress that this state-based
private-key perfectly secure encryption scheme uses a key of length equal to the total
length of plaintexts to be encrypted. Indeed, the key must be at least that long (to allow
perfect-security), and a state is essential for allowing several plaintexts to be securely
encrypted.

Partial Information Models. Note that in the case of private-key encryption schemes,
the limitations of perfect-security hold only if the adversary has full information of the
communication over the channel. On the other hand, perfectly secure private channels
can be implemented on top of channels to which the adversary has limited access.
We mention three types of channels of the latter type, which have received a lot of
attention.
r The bounded-storage model, where the adversary can freely tap the communication
channel(s) but is restricted in the amount of data it can store (cf., [148, 48, 187]).53
r The noisy channel model (which generalizes the wiretap channel of [189]), where
both the communication between the legitimate parties and the tapping channel of
the adversary are subjected to noise (cf., [148, 69] and the references therein).
r Quantum channels, where an adversary is (supposedly) prevented from obtaining
full information by the (currently believed) laws of quantum mechanics (cf., [45]
and the references therein).

Following are the author’s subjective opinions regarding these models (as a possible
basis for actual secure communication). The bounded-storage model is very appealing,
because it clearly states its reasonable assumptions regarding the abilities of the ad-
versary. In contrast, making absolute assumptions about the noise level at any point in
time seems (overly) optimistic, and thus not adequate in the context of cryptography.
Basing cryptography on quantum mechanics sounds like a very appealing idea, but at-
tempts to implement this idea have often stumbled over unjustified hidden assumptions
(which are to be expected, given the confusing nature of quantum mechanics and the
discrepancy between its scientific culture and cryptography).

5.5.3. On Some Popular Schemes

The reader may note that we have avoided the presentation of several popular encryption
schemes. We regret to say that most of these schemes are proposed without any reference
to a satisfactory notion of security.54 Thus, it is not surprising that we have nothing to
say about the contents of such proposals. In contrast, we highlight a few things that we

53 Typically, this model postulates the existence of an auxiliary (uni-directional) public channel on which a trusted
party (called a beacon) transmits a huge amount of random bits.
54 Typically, these schemes are not (semantically) secure. Furthermore, these proposals fail to suggest a weaker

definition of security that is supposedly satisfied by the proposed schemes.

477
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

have said about other popular schemes and common practices:

r The common practice of using “pseudorandom generators” as a basis for private-
key stream-ciphers (i.e., Construction 5.3.3) is sound, provided that one actually
uses pseudorandom generators (rather than programs that are called “pseudorandom
generators” but actually produce sequences that are easy to predict).55
r Whereas the plain RSA public-key encryption scheme (which employs a determin-
istic encryption algorithm) is not secure, the randomized RSA encryption scheme
(i.e., Construction 5.3.16) is secure, provided that the large hard-core conjecture
holds (see Section 5.3.4.1). Some support for the latter (clearly stated) conjecture
may be derived from the fact that a related function (i.e., much fewer least-significant
bits) constitutes a hard-core of the RSA.
We comment that the common practice of randomly padding messages before en-
crypting them (by applying the RSA function) is secure under a seemingly stronger
conjecture; see footnote 19 (in Section 5.3.4.1).
r Assuming the intractability of factoring, there exists a secure public-key encryption
scheme with efficiency comparable to that of plain RSA: We refer to the Blum-
Goldwasser public-key encryption scheme (i.e., Construction 5.3.20).
Finally, we warn that encryption schemes proved to be secure in the random oracle
model are not necessarily secure (in the standard sense). For further discussion of the
Random Oracle Methodology, we refer the reader to Section 6.6.3.

5.5.4. Historical Notes

The notion of private-key encryption scheme seems almost as ancient as the alphabet it-
self. Furthermore, it seems that the development of encryption methods went along with
the development of communication media. As the amounts of communication grew,
more efficient and sophisticated encryption methods were required. Computational
complexity considerations were explicitly introduced into the arena by Shannon [185]:
In his seminal work, Shannon considered the classical setting where no computational
considerations are present. He showed that in this information-theoretic setting, secure
communication of information is possible only so long as its entropy is lower than the
entropy of the key. He thus concluded that if one wishes to have an encryption scheme
that is capable of handling messages with total entropy exceeding the length of the key,
then one must settle for a computational relaxation of the secrecy condition. That is,
rather than requiring that the ciphertext yield no information on the plaintext, one has
to settle for the requirement that such information cannot be efficiently computed from
the ciphertext. The latter requirement indeed coincides with the definition of semantic
security.
The notion of a public-key encryption scheme was introduced by Diffie and
Hellman [75]. The first concrete candidates were suggested by Rivest, Shamir, and

55 The linear congruential generator is easy to predict [43]. The same holds for some modifications of it that
output a constant fraction of the bits of each resulting number [94]. We warn that sequences having large
linear-complexity (LFSR-complexity) are not necessarily hard to predict.

478

www.Ebook777.com
 5.5 MISCELLANEOUS

Adleman [176] and by Merkle and Hellman [154]. The abstract notion, as well as
the concrete candidate implementations (especially the RSA scheme of [176]), have
been the driving force behind the theoretical study of encryption schemes. However, the
aforementioned pioneering works did not provide a definition of security. Such satisfac-
tory definitions were provided (only a few years later) by Goldwasser and Micali [123].
The two definitions presented in Section 5.2 originate in [123], where it was shown that
ciphertext-indistinguishability implies semantic security. The converse direction is due
to [156].
Regarding the seminal paper of Goldwasser and Micali [123], a few additional com-
ments are in place. Arguably, this paper is the basis of the entire rigorous approach
to cryptography (presented in the current work): It introduced general notions such
as computational indistinguishability, definitional approaches such as the simulation
paradigm, and techniques such as the hybrid argument. Its title (“Probabilistic Encryp-
tion”) is due to the authors’ realization that public-key encryption schemes in which
the encryption algorithm is deterministic cannot be secure in the sense defined in their
paper. Indeed, this led the authors to (explicitly) introduce and justify the paradigm
of “randomizing the plaintext” as part of the encryption process. Technically speak-
ing, the paper only presents security definitions for public-key encryption schemes,
and furthermore, some of these definitions are syntactically different from the ones
we have presented here (yet all these definitions are equivalent). Finally, the term
“ciphertext-indistinguishability” used here replaces the (generic) term “polynomial-
security” used in [123]. Many of our modifications (to the definitions in [123]) are
due to Goldreich [104], which is also the main source of our uniform-complexity
treatment.56
The first construction of a secure public-key encryption scheme based on a sim-
ple complexity assumption was given by Goldwasser and Micali [123].57 Specifically,
they constructed a public-key encryption scheme assuming that deciding Quadratic
Residiousity modulo composite numbers is intractable. The condition was weakened
by Yao [190], who showed that any trapdoor permutation will do. The efficient public-
key encryption scheme of Construction 5.3.20 is due to Blum and Goldwasser [41].
The security is based on the fact that the least-significant bit of the modular squaring
function is a hard-core predicate, provided that factoring is intractable, a result mostly
due to [1].
For decades, it has been common practice to use “pseudorandom generators” in the
design of stream-ciphers. As pointed out by Blum and Micali [42], this practice is sound
provided that one uses pseudorandom generators (as defined in Chapter 3 of this work).
The construction of private-key encryption schemes based on pseudorandom functions
is due to [111].
We comment that it is indeed peculiar that the rigorous study of (the security of)
private-key encryption schemes has lagged behind the corresponding study of public-
key encryption schemes. This historical fact may be explained by the very thing that

56 Section 5.2.5.5 was added during the copyediting stage, following discussions with Johan Håstad.
57 Recall that plain RSA is not secure, whereas Randomized RSA is based on the Large Hard-Core Conjecture for
RSA (which is less appealing that the standard conjecture referring to the intractability of inverting RSA).

479
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

makes it peculiar; that is, private-key encryption schemes are less complex than public-
key ones, and hence, the problematics of their security (when applied to popular can-
didates) is less obvious. In particular, the need for a rigorous study of (the security of)
public-key encryption schemes arose from observations regarding some of their con-
crete applications (e.g., doubts raised by Lipton concerning the security of the “mental
poker” protocol of [184], which used “plain RSA” as an encryption scheme). In con-
trast, the need for a rigorous study of (the security of) private-key encryption schemes
arose later and by analogy to the public-key case.

Credits for the Advanced Section (i.e., Section 5.4)

Definitional Issues. The original definitional treatment of Goldwasser and Micali [123]
actually refers to key-dependent passive attacks (rather than to key-oblivious passive
attacks). Chosen ciphertext attacks (of the a priori and a posteriori type) were first
considered in [164] (and [174], respectively). However, these papers focused on the
formulation in terms of indistinguishability of encryptions, and formulations in terms
of semantic security have not appeared before. Section 5.4.4.2 is based on [116]. The
study of the non-malleability of the encryption schemes was initiated by Dolev, Dwork,
and Naor [77].

Constructions. The framework for constructing public-key encryption schemes that

withstand Chosen Ciphertext Attacks (i.e., Construction 5.4.23) is due to Naor and
Yung [164], who used it to construct public-key schemes that withstand a priori CCA
(under suitable assumptions). This framework was applied to the setting of a posteriori
CCA by Sahai [179, 180], who followed and improved the ideas of Dolev, Dwork, and
Noar [77] (which were the first to construct public-key schemes that withstand a poste-
riori CCA and prove Theorem 5.4.31). Our presentation of the proof of Theorem 5.4.31
follows subsequent simplification due to [142]. The key role of non-interactive zero-
knowledge proofs in this context was suggested by Blum, Feldman, and Micali [40].
The fact that security and non-malleability are equivalent under a posteriori chosen
ciphertext attack was proven in [77, 16].

5.5.5. Suggestions for Further Reading

For discussion of Non-Malleable Cryptography, which actually transcends the domain
of encryption, see [77]. Specifically, we wish to highlight the notion of non-malleable
commitment schemes, which is arguably the most appealing instantiation of the “non-
malleability paradigm”: It is infeasible for a party that is given a non-malleable commit-
ment to produce a commitment to a related string. Note that ability to produce related
commitments may endanger some applications (see, e.g., [115]), even if this ability is
not decoupled from the ability to properly decommit (to the produced commitment)
once a decommitment to the original commitment is obtained.
Recall that there is a gap between the assumptions currently required for the con-
struction of private-key and public-key encryption schemes: Whereas the former can
be constructed based on any one-way functions, the latter seem to require a trapdoor
480

www.Ebook777.com
 5.5 MISCELLANEOUS

permutation (or, actually, a “trapdoor predicate” [123]). A partial explanation to this

gap was provided by Impagliazzo and Rudich, who showed that generic (black-box)
constructions of public-key encryption schemes cannot rely on one-way functions [133]
(or even on one-way permutations [135]). This may explain the gap in our current state
of knowledge, but it does not indicate that this gap is inherent; that is, it is possible that
non-black-box constructions of public-key encryption schemes based on one-way func-
tions do exist. Indeed, Barak’s recent demonstrations of the power of non-block-box
proofs of security [5, 6] are a good lesson.
For a detailed discussion of the relationship among the various notions of secure
private-key and public-key encryption schemes, the reader is referred to [136] and [16],
respectively.

5.5.6. Open Problems

Secure public-key encryption schemes exist if there exist collections of (non-uniformly
hard) trapdoor permutations (cf. Theorem 5.3.15). It is not known whether the converse
holds (although secure public-key encryption schemes easily imply one-way func-
tions). Note that trapdoor permutations differ from general one-way functions in both
the 1-to-1 and trapdoor properties, and the former property should not be discarded
(see [23]).
Randomized RSA (i.e., Construction 5.3.16) is commonly believed to be a secure
public-key encryption scheme. It would be of great practical importance to gain addi-
tional support for this belief. As shown in Proposition 5.3.17, the security of Random-
ized RSA follows from the Large Hard-Core Conjecture for RSA, but the latter is not
known to follow from a more standard assumption, such as that RSA is hard to invert.
This is indeed the third place in the current work where we suggest the establishment
of the latter implication as an important open problem.
The constructions of public-key encryption schemes (secure against chosen cipher-
text attacks) that are presented in Section 5.4 should be considered plausibility re-
sults (which also offer some useful construction paradigms). Presenting “reasonably-
efficient” public-key encryption schemes that are secure against (a posteriori) chosen
ciphertext attacks, under general widely believed assumptions, is an important open
problem.58

5.5.7. Exercises
Exercise 1: Secure encryption schemes imply secure communication protocols: A
secure communication protocol is a two-party protocol that allows the parties
to communicate in secrecy (i.e., as in Definition 5.2.1). We stress that the sender

58 We comment that the “reasonably-efficient” scheme of [68] is based on a strong assumption regarding a specific
computational problem related to the Diffie-Hellman Key Exchange. Specifically, it is assumed that for a prime P
and primitive element g, given (P, g, (g x mod P), (g y mod P), (g z mod P)), it is infeasible to decide whether
z ≡ x y (mod P − 1).

481
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

enters such a protocol with input that equals the message to be delivered, and the
receiver enters with no input (or with input that equals the security parameter).
1. Show that any secure public-key encryption scheme yields a (two-message) secure
communication protocol.
2. Define secure communication protocol with initial set-up, and show that any se-
cure private-key encryption scheme yields such a (one-message) protocol. (Here,
the communicating parties obtain an [equal] auxiliary input that is generated at
random according to some pre-determined process.)
Advanced: Show that a secure communication protocol (even with initial set-up but
with a priori unbounded messages) implies the existence of one-way functions.
Guideline (advanced part): See guideline for Exercise 2.

Exercise 2: Secure encryption schemes imply one-way function [132]: Show that the
existence of a secure private-key encryption scheme (i.e., as in Definition 5.2.1)
implies the existence of one-way functions.
Guideline: Recall that, by Exercise 11 of Chapter 3 in Volume 1, it suffices to prove
that the former implies the existence of a pair of polynomial-time constructible
probability ensembles that are statistically far apart and still are computationally in-
distinguishable. To prove the existence of such ensembles, consider the encryption
of (n + 1)-bit plaintexts relative to a random n-bit long key, denoted K n . Specif-
ically, let the first ensemble be {(Un+1 , E(Un+1 ))}n∈N , where E(x) = E K n (x), and
(1) (2) (1) (2)
the second ensemble be {(Un+1 , E(Un+1 ))}n∈N , where Un+1 and Un+1 are inde-
pendently distributed. It is easy to show that these ensembles are computationally
indistinguishable and are both polynomial-time constructible. The more interesting
part is to show that these ensembles are statistically far apart. Note that the cor-
rect decryption condition implies that (K n , E K n (Un+1 )) contains n + 1 − o(1) bits
of information about Un+1 . On the other hand, if these ensembles are statistically
close, then E K n (Un+1 ) contains o(1) bits of information about Un+1 . Contradiction
follows, because K n may contain at most n bits of information.

Exercise 3: Encryption schemes with unbounded-length plaintexts: Suppose that the

definition of semantic security is modified so that no bound is placed on the length
of plaintexts. Prove that in such a case there exists no semantically secure encryption
scheme.
Guideline: A plaintext of length exponential in the security parameter allows the
adversary, which runs in time polynomial in its input, to find the decryption-key
by exhaustive search. In the case of public-key schemes, we merely search for a
choice of coins that make the key-generator algorithm output a key-pair with an
encryption-key that fits the one given to us. In the case of private-key schemes,
we assume that we are given all but the first bit of the plaintext (i.e., we refer to
h(1n , σ x) = x where σ ∈ {0, 1}), and search for an adequate key as well as the
value of σ .

Exercise 4: Encryption schemes must leak information about the length of the plain-
text: Suppose that the definition of semantic security is modified so that the
482

www.Ebook777.com
 5.5 MISCELLANEOUS

algorithms are not given the length of the plaintext. Prove that in such a case there
exists no semantically secure encryption scheme.
Guideline: First show that for some polynomial p, |E(1n )| < p(n) (always holds),
whereas for some x ∈ {0, 1} p(n) it must hold that Pr[|E(x)| < p(n)] < 1/2.

Exercise 5: Hiding partial information about the length of the plaintext: Using an ar-
bitrary secure encryption scheme, construct a correspondingly secure encryption
scheme that hides the exact length of the plaintext. In particular, construct an en-
cryption scheme that reveals only the following function h  of the length of the
plaintext:

1. h  (m) = m/n · n, where n is the security parameter.

2. h  (m) = 2 log2 m .

(Hint: Just use an adequate padding convention, making sure that it always allows
correct decryption.)

Exercise 6: Length parameters: Assuming the existence of a secure public-key (resp.,

private-key) encryption scheme, prove the existence of such a scheme in which the
length of the keys equal the security parameter. Furthermore, show that (without
loss of generality) the length of ciphertexts may be a fixed polynomial in the length
of the plaintext and the security parameter.
Exercise 7: On the distribution of public-keys: Let (G, E, D) be a secure public-key
encryption scheme. Prove that for every positive polynomial p, and all sufficiently
large n, it holds that maxe {Pr[G 1 (1n ) = e]} < 1/ p(n).
Guideline: Show that for any encryption-key e in the range of G 1 (1n ), one can find
a corresponding decryption-key in expected time 1/Pr[G 1 (1n ) = e].

Exercise 8: Deterministic encryption schemes: Prove that a semantically secure public-

key encryption scheme must employ a probabilistic encryption algorithm.
Guideline: For any public-key encryption scheme having a deterministic encryp-
tion algorithm, given the encryption-key, one can distinguish the encryptions of two
candidate plaintexts by computing the unique ciphertext corresponding to each of
them.

Exercise 9: An alternative formulation of Definition 5.2.1: Prove that the following

definition, in which we use non-uniform families of polynomial-size circuits (rather
than probabilistic polynomial-time algorithms) is equivalent to Definition 5.2.1.
There exists a probabilistic polynomial-time transformation T such that for
every polynomial-size circuit family {Cn }n∈N , and for every {X n }n∈N , f, h :
{0, 1}∗ → {0, 1}∗ , p and n as in Definition 5.2.1


Pr Cn (E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )

 1
< Pr C n (1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)
483
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

where C n ← T (C n ) and the probability is also taken over the internal coin
tosses of T .

Formulate and show an analogous result for public-key encryption.

Guideline: The alternative view of non-uniformity, discussed in Section 1.3 of

Volume 1, is useful here. That is, we can view a circuit family as a sequence
of advices given to a universal machine. Thus, the alternative formulation of the
definition states that advices for a machine that gets the ciphertext can be efficiently
transformed into advices for a machine that does not get the ciphertext. However, we
can incorporate the (probabilistic) transformation program into the second universal
algorithm (which then become probabilistic). Consequently, the advices are identical
for both machines (and can be incorporated in the auxiliary input h(1n , X n ) used
in Definition 5.2.1). Viewed this way, the alternative formulation is equivalent to
asserting that for some (universal) deterministic polynomial-time algorithm U , there
exists a probabilistic polynomial-time algorithm U  such that for every {X n }n∈N ,
f, h : {0, 1}∗ → {0, 1}∗ , p, and n as in Definition 5.2.1


Pr U (1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )

 1
< Pr U  (1n , 1|X n | , h(1n , X n )) = f (1n , X n ) +
p(n)

Still, a gap remains between Definition 5.2.1 and this definition: The last refers
only to one possible deterministic algorithm U , whereas Definition 5.2.1 refers
to all probabilistic polynomial-time algorithms. To close the gap, we first observe
that (by Propositions 5.2.7 and 5.2.6), Definition 5.2.1 is equivalent to a form
in which one only quantifies over deterministic polynomial-time algorithms A.
We conclude by observing that one can code any algorithm A (and polynomial
time-bound) referred to by Definition 5.2.1 in the auxiliary input (i.e., h(1n , X n ))
given to U .

Exercise 10: In continuation of Exercise 9, consider a definition in which the transfor-

mation T (of the circuit family {Cn }n∈N to the circuit family {C n }n∈N ) is not even
required to be computable.59 Clearly, the new definition is not stronger than the one
in Exercise 9. Show that the two definitions are in fact equivalent.

Guideline: Use the furthermore-clause of Proposition 5.2.7 to show that the new
definition implies indistinguishability of encryptions, and conclude by applying
Proposition 5.2.6 and invoking Exercise 9.

Exercise 11: An alternative formulation of Definition 5.2.3: Prove that Definition 5.2.3
remains unchanged when supplying the circuit with auxiliary input. That is, an
encryption scheme satisfies the modified Definition 5.2.3 if and only if

59 Equivalently, one may require that for any polynomial-size circuit family {Cn }n∈N there exists a polynomial-size
circuit family {Cn }n∈N satisfying the relevant inequality.

484

www.Ebook777.com
 5.5 MISCELLANEOUS

for every polynomial-size circuit family {Cn }, every positive polynomial p,

all sufficiently large n, and every x, y ∈ {0, 1}poly(n) (i.e., |x| = |y|) and z ∈
{0, 1}poly(n) ,


 1
| Pr Cn (z, E G 1 (1n ) (x)) = 1 − Pr C n (z, E G 1 (1n ) (y)) = 1 | <
p(n)
(Hint: Incorporate z in the circuit C n .)

Exercise 12: Equivalence of the security definitions in the public-key model: Prove
that a public-key encryption scheme is semantically secure if and only if it has
indistinguishable encryptions.
Exercise 13: The technical contents of semantic security: The following explains the
lack of computational requirements regarding the function f , in Definition 5.2.1.
Prove that an encryption scheme, (G, E, D), is (semantically) secure (in the private-
key model) if and only if the following holds:
There exists a probabilistic polynomial-time algorithm A such that for ev-
ery {X n }n∈N and h as in Definition 5.2.1, the following two ensembles are
computationally indistinguishable:

1. {E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )}n∈N .

2. {A (1n , 1|X n | , h(1n , X n ))}n∈N .

Formulate and prove an analogous claim for the public-key model.

Guideline: We care mainly about the fact that the latter formulation implies se-
mantic security. The other direction can be proven analogously to the proof of
Proposition 5.2.7.

Exercise 14: Equivalent formulations of semantic security:

1. Prove that Definition 5.2.1 remains unchanged if we restrict the function h to
depend only on the length of its input or, alternatively, h(1n , x) = h  (n) for some
h  : N →{0,1}∗ .
2. Prove that Definition 5.2.1 remains unchanged if we may restrict the function
h and the probability ensemble {X n }n∈N such that they are computable (resp.,
sampleable) by polynomial-size circuits.

Guideline (Part 1): Prove that this special case (i.e., obtained by the restriction on
h) is equivalent to the general one. This follows by combining Propositions 5.2.7
and 5.2.6. Alternatively, this follows by considering all possible probability ensem-
bles {X n }n∈N obtained from {X n }n∈N by conditioning that h(1n , X n ) = an (for every
possible sequence of an ’s).

Guideline (Part 2): The claim regarding h follows from Part 1. To establish the
claim regarding X n , observe that (by Propositions 5.2.7 and 5.2.6) we may consider
the case in which X n ranges over two strings.

485
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Exercise 15: A variant on Exercises 13 and 14.1: Prove that an encryption scheme,
(G, E, D), is (semantically) secure (in the private-key model) if and only if the
following holds:

For every probabilistic polynomial-time algorithm A there exists a probabilistic

polynomial-time algorithm A such that for every ensemble {X n }n∈N , with
|X n | = poly(n), and polynomially-bounded h  , the following two ensembles
are computationally indistinguishable.

1. {A(1n , E G 1 (1n ) (X n ), 1|X n | , h  (1n ))}n∈N .

2. {A (1n , 1|X n | , h  (1n ))}n∈N .

An equivalent form is obtained by replacing h  (1n ) with a poly(n)-bit long string vn .

Formulate and prove an analogous claim for the public-key model.

Guideline: Again, we care mainly about the fact that this variant implies seman-
tic security. The easiest proof of this direction is by applying Propositions 5.2.7
and 5.2.6. A more interesting proof is obtained by using Exercise 13: Indeed, the
current formulation is a special case of the formulation in Exercise 13, and so
we need to prove that it implies the general case. The latter is proven by observ-
ing that otherwise – using an averaging argument – we derive a contradiction in
one of the residual probability spaces defined by conditioning on h(1n , X n ) (i.e.,
(X n |h(1n , X n ) = v) for some v).

Exercise 16: Semantic security with respect to relations: The formulation of seman-
tic security in Definition 5.2.1 refers to computing a function (i.e., f ) of the
plaintext. Here we present a (related) definition that refers to finding strings that
are in a certain relation to the plaintext. Note that, unlike in Definition 5.2.1,
here we consider only efficiently recognizable relations. Specifically, we require the
following:

For every probabilistic polynomial-time algorithm A there exists a proba-

bilistic polynomial-time algorithm A such that for every ensemble {X n }n∈N ,
with |X n | = poly(n), every polynomially bounded function h, every polyno-
mially bounded relation R that is recognizable by a (non-uniform) family
of polynomial-size circuits, every positive polynomial p, and all sufficiently
large n


Pr (X n , A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n ))) ∈ R

 1
< Pr (X n , A (1n , 1|X n | , h(1n , X n ))) ∈ R +
p(n)

1. Prove that this definition is in fact equivalent to the standard definition of semantic
security.
2. Show that if the computational restriction on the relation R is removed, then no
encryption scheme can satisfy the resulting definition.
486

www.Ebook777.com
 5.5 MISCELLANEOUS

Formulate and prove analogous claims for the public-key model.

Guideline (for Part 1): Show that the new definition is equivalent to indistinguisha-
bility of encryptions. Specifically, follow the proofs of Propositions 5.2.6 and 5.2.7,
using the circuits guaranteed for R in the first proof, and noting that the second
proof holds intact.
Guideline (for Part 2): Consider the relation R = {(x, E e (x)) : |x| = 2|e|}, and the
distribution X n = U2n . (Note that if the encryption scheme is semantically secure,
then this R is not recognizable by small circuits.)

Exercise 17: Semantic security with a randomized h: The following syntactic strength-
ening of semantic security is important in some applications. Its essence is in consid-
ering information related to the plaintext, in the form of a related random variable,
rather than partial information about the plaintext (in the form of a function of
it). Prove that an encryption scheme, (G, E, D), is (semantically) secure (in the
private-key model) if and only if the following holds:
For every probabilistic polynomial-time algorithm A there exists a proba-
bilistic polynomial-time algorithm A such that for every {(X n , Z n )}n∈N , with
|(X n , Z n )| = poly(n), where Z n may depend arbitrarily on X n , and f , p, and
n as in Definition 5.2.1


Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , Z n ) = f (1n , X n )

 1
< Pr A (1n , 1|X n | , Z n ) = f (1n , X n ) +
p(n)
That is, the auxiliary input h(1n , X n ) of Definition 5.2.1 is replaced by the random
variable Z n . Formulate and prove an analogous claim for the public-key model.
Guideline: Definition 5.2.1 is clearly a special case of the latter formulation. On
the other hand, the proof of Proposition 5.2.6 extends easily to this (seemingly
stronger) formulation of semantic security.

Exercise 18: Semantic Security with respect to Oracles (suggested by Boaz Barak):
Consider an extended definition of semantic security in which, in addition to the
regular inputs, the algorithms have oracle access to a function H1n , x : {0, 1}∗ →
{0, 1}∗ (instead of being given the value h(1n , x)). The H1n , x ’s have to be restricted to
have polynomial (in n + |x|) size circuits. That is, an encryption scheme, (G, E, D),
is extended-semantically secure (in the private-key model) if the following holds:
For every probabilistic polynomial-time algorithm A there exists a proba-
bilistic polynomial-time algorithm B such that for every ensemble {X n }n∈N ,
with |X n | = poly(n), every polynomially bounded function f , every family of
polynomial-sized circuits {H1n , x }n∈N, x∈{0,1}∗ ,every positive polynomial p, and
all sufficiently large n


Pr A H1n , X n (1n , E G 1 (1n ) (X n ), 1|X n | ) = f (1n , X n )

 1
< Pr B H1n , X n (1n , 1|X n | ) = f (1n , X n ) +
p(n)
487
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

The definition of public-key security is analogous.

1. Show that if (G, E, D) has indistinguishable encryptions, then it is extended-
semantically secure.
2. Show that if no restrictions are placed on the H1n , x ’s, then no scheme can be
extended-semantically secure (in this unrestricted sense).

Guideline (for Part 1): The proof is almost identical to the proof of Proposi-
tion 5.2.6: The algorithm B forms an encryption of 1|X n | , and invokes A on it. In-
distinguishability of encryptions is used in order to establish that B H1n , X n (1n , 1|X n | )
performs essentially as well as A H1n , X n (1n , 1|X n | , E(X n )). Otherwise, we obtain a
distinguisher of E(xn ) from E(1|xn | ), for some infinite sequence of xn ’s. In particu-
lar, the oracle H1n , xn (being implementable by a small circuit) can be incorporated
into a distinguisher.

Guideline (for Part 2): In such a case, H1n , x may be defined such that, when queried
about a ciphertext, it reveals the decryption-key in use.60 Such an oracle allows A
(which is given a ciphertext) to recover the corresponding plaintext, but does not
help A (which is only given 1n , 1|X n | ) to find any information about the value of
Xn.

Exercise 19: Another equivalent definition of security: The following exercise is inter-
esting mainly for historical reasons. In the definition of semantic security appearing
in [123], the term maxu,v {Pr[ f (1n , X n ) = v|h(1n , X n ) = u]} appears instead of the
term Pr[A (1n , 1|X n | , h(1n , X n )) = f (1n , X n )]. That is, it is required that the follow-
ing holds:
For every probabilistic polynomial-time algorithm A, every ensemble {X n }n∈N ,
with |X n | = poly(n), every pair of polynomially bounded functions f, h :
{0, 1}∗ → {0, 1}∗ , every positive polynomial p, and all sufficiently large n


Pr A(1n , E G 1 (1n ) (X n ), 1|X n | , h(1n , X n )) = f (1n , X n )
  1
< max Pr [ f (1n , X n ) = v|h(1n , X n ) = u] +
u,v p(n)
Prove that this formulation is in fact equivalent to Definition 5.2.1.
Guideline: First, note that this definition is implied by Definition 5.2.1 (be-
cause maxu,v {Pr[ f (1n , X n ) = v|h(1n , X n ) = u]} ≥ Pr[A (1n , 1|X n | , h(1n , X n )) =
f (1n , X n )], for every algorithm A ). Next note that in the special case, in which X n
satisfies Pr[ f (1n , X n ) = 0|h(1n , X n ) = u] = Pr[ f (1n , X n ) = 1|h(1n , X n ) = u] = 12 ,
for all u’s, the previous terms are equal (because A can easily achieve success prob-
ability 1/2 by simply always outputting 1). Finally, combining Propositions 5.2.7
and 5.2.6, infer that it suffices to consider only the latter special case.

60 This refers to the private-key case, whereas in the public-key case, H1n , x may be defined such that, when queried
about an encryption-key, it reveals the decryption-key in use.

488

www.Ebook777.com
 5.5 MISCELLANEOUS

Exercise 20: Multiple messages of varying lengths: In continuation of Section 5.2.4,

generalize the treatment to the encryption of multiple messages of varying lengths.
That is, provide adequate definitions and analogous results.
Guideline: For example, a generalization of the first item of Definition 5.2.8
postulates that for every probabilistic polynomial-time algorithm A, there ex-
ists a probabilistic polynomial-time algorithm A such that for every ensemble
(1) (t(n)) (i)
{X n = (X n , ..., X n )}n∈N , with t(n) ≤ poly(n) and |X n | ≤ poly(n), every pair
of polynomially bounded functions f, h : {0, 1}∗ → {0, 1}∗ , every positive polyno-
mial p, and all sufficiently large n
 (1) (t(n))

Pr A(1n , E G 1 (1n ) (X n ), (1|X n | , ..., 1|X n |
), h(1n , X n )) = f (1n , X n )
 (1) (t(n))
 1
< Pr A (1n , (1|X n | , ..., 1|X n | ), h(1n , X n )) = f (1n , X n ) +
p(n)
Exercise 21: Private-key encryption secure with respect to exactly t messages. In con-
tinuation of Proposition 5.2.12, show that if secure private-key encryption schemes
exist, then for every t there are such schemes that are secure with respect to the
encryption of t messages but not with respect to the encryption of t + 1 messages.
Guideline: Given an arbitrary private-key encryption scheme (G, E, D), consider
the following private-key encryption scheme (G  , E  , D  ):
r G  (1n ) = (k, k), where k = (k0 , k1 , ..., kt ) such that (k0 , k0 ) ← G(1n ) and
k1 , ..., kt are uniformly and independently selected in {0, 1}n (without loss of
generality, n = |k0 |);
r t
E (k 0 ,k1 ,...,kt ) (x) = (E k0 (x), r, i=0 ki r i ), where r is uniformly selected in {0, 1}n ,
and the arithmetics is of the field G F(2n );
r and D(k 
(y, r, v) = Dk0 (y).
0 ,k 1 ,...,k t )

Essentially, the original scheme is augmented with a (t + 1)-out-of-2n secret sharing

scheme (see Definition 7.5.34), such that a share of the original key is revealed by
each encryption.

Exercise 22: Known plaintext attacks: Loosely speaking, in a known plaintext attack
on a private-key (resp., public-key) encryption scheme, the adversary is given some
plaintext/ciphertext pairs in addition to some extra ciphertexts (without correspond-
ing plaintexts). Semantic security in this setting means that whatever can be effi-
ciently computed about the missing plaintexts can also be efficiently computed given
only the length of these plaintexts.

1. Provide formal definitions of security under known plaintext attacks, treating both
the private-key and public-key models and referring to both the single-message
and multiple-message settings.
2. Prove that any secure public-key encryption scheme is also secure in the presence
of known plaintext attacks.
3. Prove that any private-key encryption scheme that is secure in the multiple-
message setting is also secure in the presence of known plaintext attacks.

489
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Guideline (for Part 3): Consider a function h in the multiple-message setting that
reveals some of the plaintexts.

Exercise 23: A variant on the uniform-complexity treatment (suggested by Johan

Håstad): The original motivation for the following variant of semantic security was
to allow equivalence to indistinguishability of encryptions also in the single-message
case. Intuitively, the definition asserts that whatever can be efficiently inferred from
the encryption of one piece of partial information and a second piece of partial in-
formation can be efficiently inferred only from the latter. (This should be contrasted
with Definition 5.2.13, in which the encryption is applied to the entire informa-
tion.) That is, as a variant of Definition 5.2.13, we say that an encryption scheme,
(G, E, D), is uniformly semantically secure in the public-key model if
for every probabilistic polynomial-time algorithm A there exists a proba-
bilistic polynomial-time algorithm A such that for every polynomial , ev-
ery polynomial-time computable functions h 1 , h 2 : {0, 1}∗ → {0, 1}∗ , every
f : {0, 1}∗ → {0, 1}∗ , every positive polynomial p, and all sufficiently large
n’s


Pr A(1n , G 1 (1n ), E G 1 (1n ) (h 1 (U(n) )), 1|h 1 (U(n) )| , h 2 (U(n) )) = f (U(n) )

 1
< Pr A (1n , 1|h 1 (U(n) )| , h 2 (U(n) )) = f (U(n) ) +
p(n)
where, for simplicity of notation, we have omitted the argument 1n from all
functions.
Show that this definition is equivalent to the single-message version of Defini-
tion 5.2.14 (i.e., its restriction to the case of t ≡ 1). Show that the non-uniform
variant of this definition (i.e., allowing h 1 and h 2 to be any polynomially bounded
functions) is equivalent to Definition 5.2.1. (Provide two alternative proofs to the
latter statement, with and without invoking Theorem 5.2.5.)

Exercise 24: Alterntaive formulation of state-based ciphers: For E = (E  , E  ) and

D = (D  , D  ), consider the following reformulation of Item 2 of Definition 5.3.1:
For every pair (e(0) , d (0) ) in the range of G(1n ), every sequence of plaintexts α (i) ’s, and
every i, it holds that D  (d (i−1) , E  (e(i−1) , α (i) )) = α (i) , where e( j) = E  (e( j−1) , 1|α | )
( j)

(
 ( j−1) |α |j)
and d ( j) = D  (d ( j−1) , 1|E (e ,1 )| ) for j = 1, ..., i − 1. Prove the equivalence of
the two formulations.
Exercise 25: On the standard notion of block-cipher: A standard block-cipher is a
triple, (G, E, D), of probabilistic polynomial-time algorithms that satisfies Defini-
tion 5.3.5 as well as |E e (α)| = (n) for every pair (e, d) in the range of G(1n ) and
every α ∈ {0, 1}(n) .

1. Prove that a standard block-cipher cannot be semantically secure (in the multiple-
message private-key model). Furthermore, show that any semantically secure
encryption scheme must employ ciphertexts that are longer than the corresponding
plaintexts.
490

www.Ebook777.com
 5.5 MISCELLANEOUS

2. Present a state-based version of the definition of a (secure) standard (private-key)

block-cipher, and note that Construction 5.3.3 satisfies it.

Guideline (for Part 1): Consider the encryption of a pair of two identical messages
versus the encryption of a pair of two different messages, and use the fact that
E e must be a permutation of {0, 1}(n) . Extend the argument to any encryption
scheme in which plaintexts of length (n) are encrypted by ciphertexts of length
(n) + O(log n), observing that in this case most plaintexts have only poly(n)-many
ciphertexts under E e .

Exercise 26: A secure private-key encryption scheme: Assuming that F is pseudo-

random with respect to polynomial-size circuits, prove that Construction 5.3.12
constitutes a secure private-key encryption scheme.
Guideline: Adapt the proof of Proposition 5.3.10. When referring to the security of
t = poly(n) messages, each of length  = poly(n), the adaptation requires bounding
the probability that for t uniformly selected r ( j) ’s there exists j1 , j2 ∈ {1, ..., t} and
i 1 , i 2 ∈ {1, ..., /n} such that r ( j1 ) + i 1 ≡ r ( j2 ) + i 2 (mod 2n ).

Exercise 27: The Blum-Goldwasser public-key encryption scheme was presented in

Construction 5.3.20 as a block-cipher (with arbitrary block-length). Provide an al-
ternative presentation of this scheme as a full-fledged encryption scheme (rather
than a block-cipher), and prove its security (under the factoring assumption).
Guideline: In the alternative presentation, the values of d P and d Q cannot be deter-
mined at key-generation time, but are rather computed by the decryption process.
(This means that decryption requires two additional modular exponentiations.)

Exercise 28: On the importance of restricting the ensembles {h e }e∈{0,1}∗ and {X e }e∈{0,1}∗
in Definition 5.4.1:

1. Show that if one allows arbitrary function ensembles {h e }e∈{0,1}∗ in Defini-

tion 5.4.1, then no encryption scheme can satisfy it.
2. Show that if one allows arbitrary probability ensembles {X e }e∈{0,1}∗ in Defini-
tion 5.4.1, then no encryption scheme can satisfy it, even if one uses only a single
function h that is polynomial-time computable.

Guideline: For Part 1, consider the functions h e (x) = d, where d is a decryption-

key corresponding to the encryption-key e. For Part 2, consider the random variable
X e = (d, U|e| ), where d is as before, and the function h(x  , x  ) = x  .

Exercise 29: An alternative formulation of Definition 5.4.1: Show that the following
formulation of the definition of admissible ensembles {h e }e and {X e }e is equivalent
to the one in Definition 5.4.1:
r There is a non-uniform family of polynomial-size circuits {Tn } that transform
encryption-keys (i.e., e in G 1 (1n )) into circuits that compute the corresponding
functions (i.e., h e ). That is, on input e ← G 1 (1n ), the circuit Tn outputs a circuit Ce
such that Ce (x) = h e (x) holdsfor all strings of adequate length (i.e., ≤ poly(|e|)).
491
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

r There is a non-uniform family of polynomial-size circuits {Tn } that trans-

form encryption-keys (i.e., e in G 1 (1n )) into circuits that sample the cor-
responding distributions (i.e., X e ). That is, on input e ← G 1 (1n ), the cir-
cuit Tn outputs a circuit Se such that Se (Um ) is distributed identically to X e ,
where Um denotes the uniform distribution over the set of strings of length
m = m(e).

Note that this formulation is in greater agreement with the motivating discussion pre-
ceding Definition 5.4.1. The formulation in Definition 5.4.1 was preferred because
of its relative simplicity.

Guideline: Consider, for example, the condition regarding {h e }. The formulation in

Definition 5.4.1 is shown to imply the one in this exercise by considering the circuit
family {Tn } such that on input e (in the range of G 1 (1n )), the circuit Tn outputs
def
the circuit Ce (·) = Hn (e, ·), where Hn is the circuit guaranteed by Definition 5.4.1.
That is, Tn has the description of Hn hard-wired, and outputs the description of
the circuit obtained from Hn by fixing its first input to be e. On the other hand,
given a circuit family {Tn } that transforms e → C e as here, we obtain a circuit Hn
as required in the formulation of Definition 5.4.1 as follows. The circuit Hn has
Tn hard wired, and so, on input (e, x), the circuit Hn first reconstructs the circuit
Ce ← Tn (e), and then emulates the computation of the value Ce (x).

Exercise 30: Alternative formulations of Definitions 5.4.1 and 5.4.2: Following the
framework of Section 5.4.3, present alternative definitions of security for key-
dependent passive attacks (by replacing the oracle machines A1 and A2 in Def-
initions 5.4.8 and 5.4.9 with ordinary machines). Show that these definitions are
equivalent to Definitions 5.4.1 and 5.4.2.

Guideline: For example, show how to derive circuits Pn and C n (as in Defini-
tion 5.4.2) from the machines A1 , A2 and the auxiliary input z (of Definition 5.4.9).

Exercise 31: Multiple-message security in the context of key-dependent passive at-

tacks on public-key schemes: Formulate multiple-message generalizations of Defi-
nitions 5.4.1 and 5.4.2, and prove that both are equivalent (in the public-key model)
to the single-message definitions.

Guideline: Note that admissibility for the multiple-message generalization of Def-

inition 5.4.2 means that, given an encryption-key e, one can compute (via a
polynomial-size circuit that depends only on |e|) a corresponding pair of sequences
(1) (t(|e|)) (1) (t(|e|))
((x e , ..., xe ), (ye , ..., ye )). Thus, ability to distinguish corresponding se-
(i)
quences of encryptions yields ability to distinguish, for some i, the encryption of xe
(i)
from the encryption of ye , where the latter distinguisher generates the correspond-
ing x-y hybrid (by using the circuit guaranteed by the admissibility condition and
the input encryption-key e), and invokes the former distinguisher on the resulting
sequence of encryptions.

Exercise 32: Key-oblivious versus key-dependent passive attacks: Assuming the ex-
istence of secure public-key encryption schemes, show that there exists one that
492

www.Ebook777.com
 5.5 MISCELLANEOUS

satisfies the basic definition (i.e., as in Definition 5.2.2) but is insecure under
key-dependent passive attacks (i.e., as in Definition 5.4.1).
Guideline: Given a scheme (G, E, D), define (G, E  , D  ) such that E e (x) =
(1, E e (x)) if x = e and E e (x) = (0, x) otherwise (i.e., for x = e). Using Exercise 7
(which establishes that each encryption-key is generated with negligible proba-
bility), show that (G, E  , D  ) satisfies Definition 5.2.2. Alternatively, use G  (1n ) =
((r, G 1 (1n )), G 2 (1n )), where r is uniformly distributed in {0, 1}n , which immediately
implies that each encryption-key is generated with negligible probability.

Exercise 33: Passive attacks versus Chosen Plaintext Attacks: Assuming the existence
of secure private-key encryption schemes, show that there exists one that is secure
in the standard (multi-message) sense (i.e., as in Definition 5.2.8) but is insecure
under a chosen plaintext attack (i.e., as in Definition 5.4.8).
Guideline: Given a scheme (G, E, D), define (G  , E  , D  ) such that
1. G  (1n ) = ((k, r ), (k, r )), where (k, k) ← G(1n ) and r is selected uniformly in
{0, 1}n .
 
2. E (k,r ) (x) = (1, r, E k (x)) if x = r and E (k,r ) (x) = (0, k, x) otherwise (i.e., for
x = r ).
Show that (G  , E  , D  ) is secure in the standard sense, and present a (simple but
very “harmful”) chosen plaintext attack on it.

Exercise 34: Alternative formulations of semantic security for CPA and CCA: Consider
an alternative form of Definition 5.4.8 (resp., Definition 5.4.13) in which A1 (1, z) is
replaced by A1E e (e, z) (resp., A1Ee , Dd (e, z)), where (e, d) ← G(1n ) and Condition 2
is omitted. Show that the current form is equivalent to the one presented in the main
text.
Guideline: The alternative forms presented here restrict the choice of A1 (to a
canonical one), and thus the corresponding definitions are at least as strong as the
ones in the main text. However, since Theorem 5.4.11 (resp., Theorem 5.4.15) is
established using the canonical A1 , it follows that the current definitions are actually
equivalent to the ones in the main text. We comment that we consider the formulation
in the main text to be more natural, alas more cumbersome.

Exercise 35: Chosen Plaintext Attacks versus Chosen Ciphertext Attacks: Assuming
the existence of private-key (resp., public-key) encryption schemes that are secure
under a chosen plaintext attack, show that there exists one that is secure in the
former sense but is not secure under a chosen ciphertext attack (not even in the a
priori sense).
Guideline: Given a scheme (G, E, D), define (G  , E  , D  ) such that G  = G and
1. E e (x) = (1, E e (x)) with probability 1 − 2−|e| and E e (x) = (0, x) otherwise.
2. Dd (1, y) = Dd (y) and Dd (0, y) = (d, y).
Recall that decryption is allowed to fail with negligible probability, and note that the
construction is adequate for both public-key and private-key schemes. Alternatively,
to obtain error-free decryption, define Ee (x) = (1, E e (x)), Dd (1, y) = Dd (y) and

493
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Dd (0, y) = (d, y). In the case of private-key schemes, we may define E k (k) =
(0, 1|k| ) and E k (x) = (1, E k (x)) for x = k.

Exercise 36: Chosen Ciphertext Attacks: a priori versus a posteriori: Assuming the
existence of private-key (resp., public-key) encryption schemes that are secure under
an a priori chosen plaintext attack, show that there exists one that is secure in the
former sense but is not secure under an a posteriori chosen ciphertext attack.
Guideline: Given a scheme (G, E, D), define (G  , E  , D  ) such that G  = G and
def
1. E e (x) = (b, E e (x)), where b is uniformly selected in {0, 1}.
def
2. Dd (b, y) = Dd (y).

Exercise 37: Multiple-challenge CCA security implies a posteriori CCA security.

Show that Definition 5.4.16 implies security under a posteriori CCA.

Guideline: It is tempting to claim that Definition 5.4.13 is a special case of Defi-

nition 5.4.16 (obtained when allowing only one challenge query). However, things
are not so simple: In Definition 5.4.13 the challenges are required to be identically
distributed (in the two cases), whereas in Definition 5.4.16 only computational in-
distinguishability is required. Instead, we suggest showing that Definition 5.4.14
(which is equivalent to Definition 5.4.13) is implied by the (very) restricted case of
Definition 5.4.16 discussed following the definition (i.e., a canonical adversary that
makes a single challenge query).61

Exercise 38: Equivalent forms of multiple-challenge CCA security:

1. Consider a modification of Definition 5.4.16 in which challenge queries of the

form (S, h) are answered by (E e (S(r )), h(r )), rather than by (E e (S(r )), h(S(r ))).
Prove that the original definition is equivalent to the modified one.
2. Consider a modification of Definition 5.4.16 in which the challenge queries of the
form (S, h) are replaced by two types of queries: partial-information queries of
the form (leak, h) that are answered by h(r ), and partial-encryption queries of
the form (enc, S) that are answered by E e (S(r )). Prove that the original definition
is equivalent to the modified one.

Guideline: Show how the modified model of Part 1 can emulate the original model
(that’s easy), and how the original model can emulate the modified model of Part 1
(e.g., replace the query (S, h) by the pair of queries (S, 0) and (id, h)). Next relate
the models in Parts 1 and 2.

Exercise 39: On the computational restriction on the choice of input in the definition
of adaptive NIZK: Show that if Definition 5.4.22 is strengthened by waiving the
computational bounds on , then only trivial NIZKs (i.e., languages in BPP) can
satisfy it.

61 Furthermore, we may even restrict this challenge query to be of the form (S, 0), where 0 is the all-zero function
(which yields no information).

494

www.Ebook777.com
 5.5 MISCELLANEOUS

Guideline: Show that allowing a computationally unbounded  forces the simula-

tor to generate a reference string that is statistically close to the uniform distribution.
Thus, soundness implies weak simulation-soundness in the strong sense of Exer-
cise 40 (i.e., with respect to a computationally unbounded as in Definition 5.4.22),
and by applying Exercise 40 we are done.

Exercise 40: Weak simulation-soundness can hold only with respect to computationally
bounded cheating provers. Show that if Definition 5.4.24 is strengthened by waiving
the computational bounds on , then only trivial NIZKs (i.e., for languages in BPP)
can satisfy it.
Guideline: Show that otherwise the two-stage simulation procedure, S = (S1 , S2 ),
can be used to distinguish inputs in the language L from inputs outside the language,
because in the first case it produces a valid proof whereas in the second case it cannot
do so. The latter fact is proved by showing that if S2 (which also gets an auxiliary
input s produced by S1 along with the reference string) produces a valid proof for
some x ∈ L, then a computationally unbounded prover may do the same by first
generating s according to the conditional distribution induced by the reference string
(and then invoking S2 ).

Exercise 41: Does weak simulation-soundness hold for all adaptive NIZKs?

1. Detect the flaw in the following argument toward an affirmative answer: If weak
simulation-soundness does not hold, then we can distinguish a uniformly selected
reference string (for which soundness holds) from a reference string generated
by S1 (for which soundness does not hold).
2. Assuming the existence of one-way permutations (and adaptive NIZKs), show
an adaptive NIZK with a suitable simulator such that weak simulation-soundness
does not hold.
3. (Suggested by Boaz Barak and Yehuda Lindell): Consider languages containing
pairs (α, x) such that one can generate α’s along with suitable trapdoors t(α)’s that
allow for determining whether or not inputs of the form (α, ·) are in the language.
For such languages, define a weaker notion of simulation-soundness that refers
to the setting in which a random α is generated and then one attempts to produce
valid proofs for a no-instance of the form (α, ·) with respect to a reference-string
generated by S1 . (The weaker notion asserts that in this setting it is infeasible to
produce a valid proof for such a no-instance.) Provide a clear definition, prove
that it is satisfied by any adaptive NIZK for the corresponding language, and show
that this definition suffices for proving Theorem 5.4.27.

Guideline (Part 1): The existence of an efficient C = (, ) that violates weak
simulation-soundness only means that for a reference string generated by S1 , the
cheating generates a valid proof for a no-instance selected by . When C is given
a uniformly selected reference string, it either may fail to produce a valid proof or
may produce a valid proof for a yes-instance. However, we cannot necessarily
distinguish no-instances from yes-instances (see, for example, Part 2). This gap is
eliminated in Part 3.

495
 Free ebooks ==> www.Ebook777.com
ENCRYPTION SCHEMES

Guideline (Part 2): Given a one-way permutation f with a corresponding hard-core

def
predicate b, consider the pseudorandom generator G(s) = (G  (s), f 2|s| (s)), where
def
G  (s) = b(s)b( f (s)) · · · b( f 2|s|−1 (s)) (see proof of Proposition 5.3.19). Let L de-
note the set of strings that are not images of G, and note that L is in N P (because
L = {(α, β) : ∃s s.t. β = f 2|s| (s) ∧ α = G  (s)}). Given any adaptive NIZK for L,
denoted (P, V ), consider the modification (P  , V  ) such that P  (x, w, (r1 , r2 )) =
P(x, w, r1 ) and V  (x, (r1 , r2 ), π) = 1 if either V (x, π, r1 ) = 1 or x = r2 . The mod-
def
ified simulator is derived by S1 (1n ) = ((r1 , r2 ), s), where (r1 , s) ← S1 (1n ) and
def
r2 ← G(Un ) (and S2 (x, s) = S2 (x, s)). Verify that the modified algorithms sat-
isfy the definition of an adaptive NIZK, and note that weak simulation-soundness
is easily violated by (r1 , r2 ) = r2 ∈ L (and any ).

Exercise 42: On defining non-malleability: Show that when defining non-malleability

(i.e., in Definitions 5.4.32 and 5.4.33), it is essential to prevent A from outputting
the ciphertext that is given to it.
Guideline: Consider the identity relation, a constant function h, and let X n be
uniform over {0, 1}n . Note that A gets (e, E e (X n ), 1n ), whereas A only gets 1n .

496

www.Ebook777.com
 CHAPTER SIX

Digital Signatures and Message

Authentication

Message authentication and (digital) signatures were the first tasks that joined en-
cryption to form modern cryptography. Both message authentication and digital sig-
natures are concerned with the “authenticity” of data, and the difference between
them is analogous to the difference between private-key and public-key encryption
schemes.
In this chapter, we define message authentication and digital signatures, and the se-
curity notions associated with them. We show how to construct message-authentication
schemes using pseudorandom functions, and how to construct signature schemes using
one-way permutations. We stress that the latter construction employs arbitrary one-way
permutations, which do not necessarily have a trapdoor.

Organization. The basic definitions are presented in Section 6.1. Constructions of

message-authentication schemes and signature schemes are presented in Sections 6.3
and 6.4, respectively. Toward presenting these constructions, we discuss restricted types
of message authentication and signature schemes, which are of independent interest,
such as length-restricted schemes (see Section 6.2) and one-time signature schemes
(see Section 6.4.1). Additional issues are discussed in Sections 6.5 and 6.6.

Teaching Tip. In contrast to the case of encryption schemes (cf. Chapter 5), the def-
initional treatment of signatures (and message authentication) is quite simple. The
treatment of length-restricted schemes (see Section 6.2) plays an important role in
the construction of standard schemes, and thus we strongly recommend highlighting
this treatment. We suggest focusing on the presentation of the simplest construction of
message-authentication schemes (provided in Section 6.3.1) and on the (not-so-simple)
construction of signature schemes that is provided in Sections 6.4.1 and 6.4.2. As in
Chapter 5, we assume that the reader is familiar with the material in Chapters 2 and 3 of
Volume 1 (and specifically with Sections 2.2, 2.4, and 3.6). This familiarity is important
not only because we use some of the notions and results presented in these sections but
also because we use similar proof techniques (and do so while assuming that this is not
the reader’s first encounter with these techniques).
497
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

6.1. The Setting and Definitional Issues

Both signature schemes and message-authentication schemes are methods for “validat-
ing” data, that is, verifying that the data was approved by a certain party (or set of parties).
The difference between signature schemes and message-authentication schemes is that
“signatures” should be “universally verifiable,” whereas “authentication tags” are only
required to be verifiable by parties that are also able to generate them. It is customary
to discuss each of these two types of schemes separately, and we start by providing a
brief overview of such a nature. We then turn to our actual treatment, which applies to
both types of schemes in a unified manner.

6.1.1. The Two Types of Schemes: A Brief Overview

The need to discuss “digital signatures” has arisen with the introduction of computer
communication to the business environment (in which parties need to commit them-
selves to proposals and/or declarations that they make). Discussions of “unforgeable
signatures” also took place in previous centuries, but the objects of discussion were
handwritten signatures (and not digital ones), and the discussion was not perceived
as related to “cryptography.” Loosely speaking, a scheme for unforgeable signatures
should satisfy the following:

r Each user can efficiently produce his/her own signature on documents of his/her
choice;
r every user can efficiently verify whether a given string is a signature of another
(specific) user on a specific document; but
r it is infeasible to produce signatures of other users to documents that they did not
sign.
We note that the formulation of unforgeable digital signatures also provides a clear
statement of the essential ingredients of handwritten signatures. The ingredients are
each person’s ability to sign for him/herself, a universally agreed-upon verification
procedure, and the belief (or assertion) that it is infeasible (or at least hard) to
forge signatures in a manner that passes the verification procedure. It is not clear
to what extent handwritten signatures do meet these requirements. In contrast, our
treatment of digital-signature schemes provides precise statements concerning the
extend to which digital signatures meet these requirements. Furthermore, unforge-
able digital signature schemes can be constructed based on the existence of one-way
functions.
Message authentication is a task related to the setting considered for encryp-
tion schemes; that is, communication over an insecure channel. This time, we con-
sider an active adversary that is monitoring the channel and may alter the mes-
sages sent on it. The parties communicating through this insecure channel wish
to authenticate the messages they send so that their counterpart can tell an orig-
inal message (sent by the sender) from a modified one (i.e., modified by the

498

www.Ebook777.com
 6.1 THE SETTING AND DEFINITIONAL ISSUES

adversary). Loosely speaking, a scheme for message authentication should satisfy the
following:
r Each of the communicating parties can efficiently produce an authentication tag to
any message of his/her choice;
r each of the communicating parties can efficiently verify whether a given string is an
authentication tag of a given message; but
r it is infeasible for an external adversary (i.e., a party other than the communicating
parties) to produce authentication tags to messages not sent by the communicating
parties.
Note that in contrast to the specification of signature schemes, we do not require uni-
versal verification: Only the designated receiver is required to be able to verify the
authentication tags. Furthermore, we do not require that the receiver be unable to pro-
duce authentication tags by itself (i.e., we only require that external parties not be able
to do so). Thus, message-authentication schemes cannot convince a third party that the
sender has indeed sent the information (rather than the receiver having generated it by
itself ). In contrast, signatures can be used to convince third parties. In fact, a signature
to a document is typically sent to a second party so that in the future, this party may
(by merely presenting the signed document) convince third parties that the document
was indeed generated (or sent or approved) by the signer.

6.1.2. Introduction to the Unified Treatment

Loosely speaking, message-authentication and signature schemes are supposed to en-
able reliable transmission of data between parties. That is, the basic setting consists of
a sender and a receiver, where the receiver may be either predetermined or determined
only after the data was sent. Loosely speaking, the receiver wishes to be guaranteed
that the data received was actually sent by the sender, rather than modified (or even
concocted) by somebody else (i.e., an adversary). The receiver may be a party that
shares an explicit (unreliable) point-to-point communication line with the sender; this
is indeed the typical setting in which message authentication is employed. However,
in other cases (typically when signature schemes are employed), the receiver may be
any party that obtains the data in the future and wishes to verify that it was indeed
sent by the declared sender. In both cases, the reliability (or authenticity) of the data is
established by an authentication process that consists of two main procedures:
1. A signing procedure that is employed by the alleged sender in order to produce
signatures to data of its choice.
2. A verification procedure that is employed by the receiver in order to determine the
authenticity of the data using the provided signature.
As in case of encryption schemes, the authentication process presupposes also a third
procedure called key-generation that allows the sender to generate a signing-key (to be
used in the signing procedure), along with a verification-key (to be used in the verifica-
tion procedure). The key-generation procedure is typically invoked by the sender, and

499
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

the possession of the signing-key constitutes the sender’s advantage over the adversary
(see analogous discussion in Chapter 5). That is, without the signing-key, it is infeasible
to generate valid signatures (with respect to the corresponding verification-key). Fur-
thermore, even after receiving signatures to messages of its choice, an adversary (lacking
the signing-key) cannot generate a valid signature to any other message.
As previously stated, the ability to produce valid signatures is linked to the knowl-
edge of the signing-key. Loosely speaking, “security” (or “unforgeability”) means the
infeasibility of producing valid signatures without knowledge of the signing-key, where
validity means passing verification with respect to the corresponding verification-key.
The difference between message-authentication and signature schemes amounts to the
question of whether “security” also holds when the verification-key is publicly known:
In the case of message-authentication schemes, the verification-key is assumed to be
kept secret (and so these schemes are of the “private-key” type), whereas in the case
of signature schemes, the verification-key may be made public (and so these schemes
are of the “public-key” type). Thus, the difference between message-authentication
and signature schemes is captured by the security definition, and effects the possible
applications of these schemes.
From the point of view of their functionality, the difference between message-
authentication and signature schemes arises from the difference in the settings for which
they are intended, which amounts to a difference in the identity of the receiver and in
the level of trust that the sender has in the receiver. Typically, message-authentication
schemes are employed in cases where the receiver is predetermined (at the time of
message transmission) and is fully trusted by the sender, whereas signature schemes
allow verification of the authenticity of the data by anybody (which is certainly not
trusted by the sender). In other words, signature schemes allow for universal verifica-
tion, whereas message-authentication schemes may only allow predetermined parties
to verify the authenticity of the data. Thus, in signature schemes the verification-key
must be known to anybody, and in particular is known to the adversary. In contrast, in
message-authentication schemes, the verification-key is only given to a set of predeter-
mined receivers that are all trusted not to abuse this knowledge; that is, in such schemes
it is postulated that the verification-key is not (a priori) known to the adversary. (See
Figure 6.1.)

Summary and Terminology. Message-authentication and signature schemes differ

in the question of whether the verification-key is “private” (i.e., a secret unknown to
the adversary) or “public” (i.e., known to everybody and in particular known to the
adversary). Thus, in a sense, these are private-key and public-key versions of a task that

Type Verification-key known Verification possible

Message auth. to the designated for the designated
schemes (trusted) receiver(s) only (trusted) receiver(s) only
Signature to everybody for anybody
schemes (including the adversary) (including the adversary)
Figure 6.1: Message-authentication versus signature schemes.

500

www.Ebook777.com
 6.1 THE SETTING AND DEFINITIONAL ISSUES

lacks a good name (since both authentication and signatures are already taken by one
of the two versions). Still, seeking a uniform terminology, we shall sometimes refer to
message-authentication schemes (also known as Message Authentication Codes [mac])
as to private-key signature schemes. Analogously, we shall sometimes refer to signature
schemes as to public-key signature schemes.

6.1.3. Basic Mechanism

We start by defining the basic mechanism of message-authentication and signature
schemes. Recall that this basic mechanism will support both the private-key and public-
key versions, and the difference between the two versions will only be reflected in
the definition of security. Indeed, the definition of the basic mechanism says nothing
about the security of the scheme (which is the subject of the next section), and thus is
the same for both the private-key and public-key versions. In both cases, the scheme
consists of three efficient algorithms: key generation, signing (or authenticating), and
verification. The basic requirement is that signatures that are produced by the signing
algorithm be accepted as valid by the verification algorithm, when fed a verification-key
corresponding to the signing-key used by the signing algorithm.

Definition 6.1.1 (signature scheme): A signature scheme is a triple, (G, S, V ), of

probabilistic polynomial-time algorithms satisfying the following two conditions:
1. On input 1n , algorithm G (called the key-generator) outputs a pair of bit strings.
2. For every pair (s, v) in the range of G(1n ), and for every α ∈ {0, 1}∗ , algorithms S
(signing) and V (verification) satisfy
Pr[V (v, α, S(s, α)) = 1] = 1

where the probability is taken over the internal coin tosses of algorithms S and V .

The integer n serves as the security parameter of the scheme. Each (s, v) in the range
of G(1n ) constitutes a pair of corresponding signing/verification keys.

We sometimes call S(s, α) a signature to the document α produced using the

signing-key s . Likewise, when V (v, α, β) = 1, we say that β is a valid signature
to α with respect to the verification-key v. (Indeed, at this point, we may assume that
algorithm V is deterministic, but see subsequent comments.) This definition asserts that
any signature to α produced using the signing-key s is a valid signature to α with respect
to the corresponding verification-key v. Note that there may be valid signatures (with
respect to v) that are not produced by the signing process (using the corresponding s).
We stress that Definition 6.1.1 says nothing about security, and so trivial (i.e., inse-
def def
cure) triples of algorithms may satisfy it (e.g., S(s, α) = 0 and V (v, α, β) = 1, for all
s, v, α and β). Furthermore, Definition 6.1.1 does not distinguish private-key signature
schemes from public-key ones. The difference between the two types is introduced in
the security definitions: In a public-key scheme, the “adversary” gets the verification-
key (i.e., v) as an additional input (and thus v = s follows), whereas in private-key
501
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

schemes, v is not given to the “adversary” (and thus one may assume, without loss of
generality, that v = s).

Notation. In the rest of this work, we shall write Ss (α) instead of S(s, α) and Vv (α, β)
instead of V (v, α, β). Also, we let G 1 (1n ) (resp., G 2 (1n )) denote the first (resp., second)
element in the pair G(1n ). That is, G(1n ) = (G 1 (1n ), G 2 (1n )). Without loss of generality,
we may assume that |G 1 (1n )| and |G 2 (1n )| are polynomially related to n, and that each
of these integers can be efficiently computed from the other.

Comments: A Few Relaxations

Definition 6.1.1 may be relaxed in several ways without significantly harming its useful-
ness. For example, we may relax Condition (2) and allow a negligible verification error
(e.g., Pr[Vv (α, Ss (α)) = 1] < 2−n ). Alternatively, one may postulate that Condition (2)
holds for all but a negligible measure of the key-pairs generated by G(1n ). At least one
of these relaxations is essential for many suggestions of (public-key) signature schemes.
Especially in the case where we adopt the first relaxation of Condition (2), it makes
sense to consider also randomized verification algorithms. However, all natural signa-
ture schemes happen to employ a deterministic verification algorithm (see Exercise 1).
Still, in the case of probabilistic verification algorithms, we may define β as a valid
signature of α (with respect to v) if Pr[Vv (α, β) = 1] ≥ 1/2. The threshold 1/2 used
here is quite arbitrary, and the definition is essentially robust under the replacement of
1/2 by either 1/poly(n) or 1 − 2−poly(n) .1 Alternatively, we may view β as a “fractionally
valid” signature of α with respect to v (i.e., valid with probability Pr[Vv (α, β) = 1]).
Another relaxation of Definition 6.1.1 consists of restricting the domain of pos-
sible documents. However, unlike the situation with respect to encryption schemes,
such a restriction is non-trivial in the current context, and is discussed at length in
Section 6.2.

6.1.4. Attacks and Security

Loosely speaking, secure signature schemes should prevent an adversary from generat-
ing valid signatures to “unauthentic” documents (i.e., documents that were not approved
by the legitimate signer). Thus, the potential adversary is “active” at least in the mild
sense that it attempts to “generate” something new and different from all that it holds
(rather than to “extract” information that is implicit in something that is given to it).2

1 Indeed, robustness follows by “amplification” (i.e., error- reduction) of the verification algorithm. For example,
given V as here, one may consider V  that applies V to the tested pair for a linear number of times and accepting
if and only if V has accepted in all tries.
2 Indeed, in general, the distinction between “generating something new” and “extracting something implicit”
cannot be placed on firm grounds. However, our reference to this distinction is merely at the motivational
level. Furthermore, this distinction can be formalized in the context that we care about, which is the context
of comparing encryption and signature schemes (or, rather, the adversaries attacking these schemes). In the
case of encryption schemes, we consider adversaries that try to extract information about the plaintext from
the ciphertext. That is, the desired object is a function of the given input. In contrast, in the case of signature
schemes, we consider adversaries that try to generate a valid signature with respect to a certain verification-key.
That is, the desired object is not a function of the given input.

502

www.Ebook777.com
 6.1 THE SETTING AND DEFINITIONAL ISSUES

Furthermore, the typical applications of signature schemes are to setting in which the
adversary may obtain from the legitimate signer valid signatures to some documents
of the adversary’s choice. For this reason, the basic definition of security of signature
schemes refers to such “chosen message attacks” (to be discussed and defined next).
(Indeed, the situation here is different from the case of encryption schemes, where the
basic definition refers to a “passive” adversary that only wire-taps a communication
line, in encrypted form, over this line.)
We shall consider a very strong definition of security (against “chosen message
attacks”). That is, we consider very powerful attacks on the signature scheme, as well
as a very liberal notion of breaking it. Specifically, during the course of the attack,
the attacker is allowed to obtain signatures to any document of its choice. One may
argue that in many applications, such a general attack is not possible (because, in
these applications, documents to be signed must have a specific format). Yet our view
is that it is impossible to define a general (i.e., application-independent) notion of
admissible documents, and thus a general/robust definition of an attack seems to have
to be formulated as suggested here. (Note that at worst, our approach is overly cautious.)
Likewise, the attacker is said to be successful if it can produce a valid signature to any
document for which it has not asked for a signature during its attack. Again, this defines
the ability to form signatures to possibly “nonsensical” documents as a breaking of the
scheme. Yet, again, we see no way to have a general (i.e., application-independent)
notion of “meaningful” documents (so that only forging signatures to them will be
considered a breaking of the scheme). This discussion leads to the following (slightly
informal) formulation:

r A chosen message attack is a process that can obtain signatures to strings of its
choice, relative to some fixed signing-key that is generated by G. We distinguish two
cases:
The private-key case: Here the attacker is given 1n as input, and the signatures are
produced relative to s, where (s, v) ← G(1n ).
The public-key case: Here the attacker is given v as input, and the signatures are
produced relative to s, where (s, v) ← G(1n ).
r Such an attack is said to succeed (in existential forgery) if it outputs a valid signature
to a string for which it has not requested a signature during the attack. That is, the
attack is successful if it outputs a pair (α, β) such that Vv (α, β) = 1 (where v is as in
the previous item) and α is different from all strings for which a signature has been
required during the attack.
r A signature scheme is secure (or unforgeable) if every feasible chosen message
attack succeeds with at most negligible probability.

Formally, a chosen message attack is modeled by a probabilistic polynomial-time

oracle machine that is given oracle access to a “keyed signing process” (i.e., the signing
algorithm combined with a signing-key). Depending on the version (i.e., public-key
or not), the attacker may get the corresponding verification-key as input. We stress
that this is the only difference between the two cases (i.e., private-key and public-key),
503
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

which are spelled out in Definition 6.1.2. We refer the reader to the clarifying discussion
that follows Definition 6.1.2; in fact, some readers may prefer to read that discussion
first.

Definition 6.1.2 (unforgeable signatures): For a probabilistic oracle machine, M, we

denote by Q O
M (x) the set of queries made by M on input x and access to oracle O.
O
As usual, M (x) denotes the output of the corresponding computation. We stress that
QO O
M (x) and M (x) are dependent random variables that represents two aspects of the
same probabilistic computation.

The private-key case: A private-key signature scheme is secure if for every proba-
bilistic polynomial-time oracle machine M, every positive polynomial p, and all
sufficiently large n, it holds that


Vv (α, β) = 1 & α ∈ Q SMs (1n ) 1
Pr <
where (s, v) ← G(1 ) and (α, β) ← M (1 )
n Ss n
p(n)
where the probability is taken over the coin tosses of algorithms G, S, and V, as well
as over the coin tosses of machine M.
The public-key case: A public-key signature scheme is secure if for every probabilistic
polynomial-time oracle machine M, every positive polynomial p, and all sufficiently
large n, it holds that


Vv (α, β) = 1 & α ∈ Q SMs (v) 1
Pr <
where (s, v) ← G(1 ) and (α, β) ← M (v)
n Ss
p(n)
where the probability is taken over the coin tosses of algorithms G, S, and V , as
well as over the coin tosses of machine M.

The definition refers to the following experiment. First a pair of keys, (s, v), is generated
by invoking G(1n ), and is fixed for the rest of the discussion. Next, an attacker is invoked
on input 1n or v, depending on whether we are in the private-key or public-key case.
In both cases, the attacker is given oracle access to Ss , where the latter may be a
probabilistic oracle rather than a standard deterministic one (e.g., if queried twice for
the same value, then the probabilistic signing-oracle may answer in different ways).
Finally, the attacker outputs a pair of strings (α, β). The attacker is deemed successful
if and only if the following two conditions hold:

1. The string α is different from all queries (i.e., requests for signatures) made by the
attacker; that is, the first string in the output pair (α, β) = M Ss (x) is different from
any string in Q SMs (x), where x = 1n or x = v, depending on whether we are in the
private-key or public-key case.
We stress that both M Ss (x) and Q SMs (x) are random variables that are defined based
on the same random execution of M (on input x and oracle access to Ss ).
2. The pair (α, β) corresponds to a valid document-signature pair relative to the verifi-
cation key v. In case V is deterministic (which is typically the case) this means that
504

www.Ebook777.com
 6.1 THE SETTING AND DEFINITIONAL ISSUES

Vv (α, β) = 1. The same applies also in case V is probabilistic, and when viewing
Vv (α, β) = 1 as a random variable. (Alternatively, in the latter case, a condition such
as Pr[Vv (α, β) = 1] ≥ 1/2 may replace the condition Vv (α, β) = 1.)

6.1.5.* Variants
Clearly, any signature scheme that is secure in the public-key model is also secure in
the private-key model. The converse is not true: Consider, for example, the private-
key scheme presented in Construction 6.3.1 (as well as any other “natural” message-
authentication scheme). Following are a few other comments regarding the definitions.

6.1.5.1. Augmenting the Attack with a Verification Oracle

It is natural to augment Definition 6.1.2 by providing the adversary with unlimited access
to the corresponding verification-oracle Vv . We stress that (in this augmented definition)
the documents that (only) appear in the verification queries are not added to the set Q SMs ;
that is, the output (α, β) is considered a successful forgery even if the adversary made
a verification-query of the form (α, ·), but provided (as in Definition 6.1.2) that the
adversary did not make the signing-query α (and that Vv (α, β) = 1).
Indeed, in the public-key case, the verification-oracle adds no power to the adversary,
because the adversary (which is given the verification-key) can emulate the verification-
oracle by itself. Furthermore, typically, also in the private-key model, the verification-
oracle does not add much power. Specifically, we have:

Proposition 6.1.3 (cases in which security extends to the augmented model):

1. Any secure public-key signature scheme is secure also under attacks that utilize a
verification-oracle (in addition to the signing-oracle).
2. Any secure private-key signature scheme that has unique valid signatures (as defined
next) is secure also under attacks that utilize a verification-oracle (in addition to the
signing-oracle).
A signature scheme (G, S, V ) is said to have unique valid signatures if for every
verification-key v and document α, there exists a unique β such that Vv (α, β) = 1
(or, such that Pr[Vv (α, β) = 1] > 1/ poly(|v|)). As discussed in Section 6.5.1 (see also
Exercises 1 and 2), any secure private-key signature scheme can be transformed into
one having a deterministic verification algorithm and unique valid signatures. In fact,
all private-key signature schemes presented in Section 6.3 have unique valid signatures.
We comment that the unique signature property is essential for the validity of Part 2;
see Exercise 3.
Proof Sketch: As stated previously, Part 1 is obvious (because a standard adversary can
emulate the verification-oracle by using the verification-key given to it). We prove Part 2
by showing that also in that case, a standard adversary can emulate the verification-
oracle. However, in this case, the emulation is less obvious, because the standard ad-
versary cannot test the validity of signatures by itself. Still, considering an arbitrary
505
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

combined attack on such a private-key signature scheme, we emulate the verification-

queries (in the standard model) as follows:

r For a verification-query (α, β), if α equals a previous signing-query, then we can

emulate the answer by ourselves. Specifically, if the signing-query α was answered
with β, then we answer the verification-query positively; otherwise we answer it
negatively. The correctness of the emulation follows from the hypothesis that this
signature scheme has unique valid signatures.
r Otherwise (i.e., for a verification-query (α, β) such that α does not equal any previous
signing-query), we may choose to either halt and output (α, β) as a candidate forgery
(gambling on Vv (α, β) = 1) or continue and emulate a negative answer by ourselves
(gambling on Vv (α, β) = 0). Specifically, for every such verification-query, we may
choose the first possibility with probability 1/t(n) and the second possibility oth-
erwise, where t(n) is a bound on the number of verification-queries performed by
the original augmented attack (which we emulate). It can be shown that the success
probability of the resulting standard adversary is at least a 1/t(n) fraction of the
success probability of the given adversary. For details see Exercise 3.
Thus, insecurity in the augmented model implies insecurity in the original model, and
the proposition follows.

6.1.5.2. Inessential Generalities

The definitions presented here (specifically, Definition 6.1.1) were aimed at generality
and flexibility. We comment that several levels of freedom can be eliminated without loss
of generality (but with some loss of convenience). Firstly, as in the case of encryption
schemes, one may modify the key-generation algorithm so that on input 1n it outputs a
pair of n-bit long keys. Two more fundamental restrictions, which actually do not affect
the existence of secure schemes, follow.

Randomization in the Signing Process. In contrast to the situation with respect to

encryption schemes (see Sections 5.2 and 5.3), randomization is not essential to the
actual signing and verifying processes (but is, as usual, essential to key-generation). That
is, without loss of generality (but with possible loss in efficiency), the signing algorithm
may be deterministic, and in all of the schemes we present (in the current chapter), the
verification algorithm is deterministic. For further discussion, see Exercise 1.

Canonical Verification in the Private-Key Version. As hinted earlier, in the private-

def
key case, we may just identify the signing and verification keys (i.e., k = s = v).
Furthermore (following the comment about deterministic signing), without loss of
generality, verification may amount to comparing the alleged signature to one pro-
duced by the verification algorithm itself (which may just produce signatures exactly
as the signing algorithm). That is, for a deterministic signing process Sk , we may let
def
Vk (α, β) = 1 if and only if β = Sk (α). For details, see Exercise 2.
506

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

6.1.5.3. Weaker Notions of Security and Some Popular Schemes

Weaker notions of security have been considered in the literature. The various notions
refer to two parameters: (1) the type of attack, and (2) when the adversary is consid-
ered to be successful. Indeed, Definition 6.1.2 refers to the most severe type of attacks
(i.e., unrestricted chosen message attacks) and to the most liberal notion of success
(i.e., the ability to produce a valid signature to any new message). For further discus-
sion, the interested reader is referred to Section 6.6.3. In particular, we note that plain
RSA, as well as plain versions of Rabin’s scheme and the DSS, are not secure under
Definition 6.1.2. However, these schemes satisfy weaker notions of security, provided
that some (standard) intractability assumptions hold. Furthermore, variants of these
signature schemes (in which the function is not applied directly to the document itself)
may be secure (under Definition 6.1.2).

6.2. Length-Restricted Signature Scheme

Restricted types of (public-key and private-key) signature schemes play an important

role in our exposition. The first restriction we consider is the restriction of signature
schemes to (apply only to) documents of a certain predetermined length. We call the re-
sulting schemes length-restricted. The effect of the length-restriction is more dramatic
here (in the context of signature schemes) than it is in the context of encryption schemes;
this can be appreciated by comparing (the length of) Section 6.2.2 to (the length of)
Section 5.3.2.2. Nevertheless, as we shall show (see Theorem 6.2.2), if the length re-
striction is not too low, then the full power of signature schemes can be regained; that
is, length-restricted signature schemes yield full-fledged ones.

6.2.1. Definition
The essence of the length-restriction is that security is guaranteed only with respect to
documents of the predetermined length. Note that the question of what is the result of
invoking the signature algorithm on a document of improper length is immaterial. What
is important is that an attacker (of a length-restricted scheme) is deemed successful only
if it produces a signature to a (different) document of proper length. Still, for the sake of
concreteness (and simplicity of subsequent treatment), we define the basic mechanism
only for documents of proper length.

Definition 6.2.1 (signature scheme for fixed-length documents): Let  : N → N. An

-restricted signature scheme is a triple, (G, S, V ), of probabilistic polynomial-time
algorithms satisfying the following two conditions:

1. As in Definition 6.1.1, on input 1n , algorithm G outputs a pair of bit strings.

2. Analogously to Definition 6.1.1, for every n and every pair (s, v) in the range of G(1n ),
and for every α ∈ {0, 1}(n) , algorithms S and V satisfy Pr[Vv (α, Ss (α)) = 1] = 1.

507
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Such a scheme is called secure (in the private-key or public-key model) if the (corre-
sponding) requirements of Definition 6.1.2 hold when restricted to attackers that only
make queries of length (n) and output a pair (α, β) with |α| = (n).

We stress that the essential modification is presented in the security condition. The latter
considers an adversary to be successful only in case it forges a signature to a (different)
document α of the proper length (i.e., |α| = (n)).

6.2.2. The Power of Length-Restricted Signature Schemes

We comment that -restricted private-key signature schemes for (n) = O(log n) are
trivial (since the signing and verification keys may contain a table look-up associating
a secret with each of the 2(n) = poly(n) possible documents).3 In contrast, this trivi-
ality does not hold for public-key signature schemes. (For details on both claims, see
Exercise 5.) On the other hand, in both (private-key and public-key) cases, -restricted
signature schemes for any super-logarithmic  (e.g., (n) = n or even (n) = log22 n)
are as powerful as ordinary signature schemes:

Theorem 6.2.2: Suppose that  is a super-logarithmically growing function. Then,

given an -restricted signature scheme that is secure in the private-key (resp., public-
key) model, one can construct a full-fledged signature scheme that is secure in the same
model.

Results of this flavor can be established in two different ways, corresponding to two
methods of converting an -restricted signature scheme into a full-fledged one. Both
methods are applicable both to private-key and public-key signature schemes. The
first method (presented in Section 6.2.2.1) consists of parsing the original document
into blocks (with adequate “linkage” between blocks), and applying the -restricted
scheme to each block. The second method (presented in Section 6.2.2.2) consists of
hashing the document into an (n)-bit long value (via an adequate hashing scheme),
and applying the restricted scheme to the resulting value. Thus, the second method
requires an additional assumption (i.e., the existence of “collision-free” hashing), and
so Theorem 6.2.2 (as stated) is actually proved using the first method. The second
method is presented because it offers other benefits; in particular, it yields signatures
of fixed length (i.e., the signature-length only depends on the key-length) and uses a
single invocation of the restricted scheme. The latter feature will play an important role
in subsequent sections (e.g., in Sections 6.3.1.2 and 6.4.1.3).

6.2.2.1. Signing (Augmented) Blocks

In this subsection we present a simple method for constructing general signature
schemes out of length-restricted ones, and in doing so we establish Theorem 6.2.2.

3 Recall that such triviality does not hold in the context of encryption schemes, not even in the private-key case.
See Section 5.3.2.

508

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

Loosely speaking, the method consists of parsing the original document into blocks
(with adequate “linkage” between blocks), and applying the length-restricted scheme
to each (augmented) block.
Let  and (G, S, V ) be as in Theorem 6.2.2. We construct a general signature scheme,
(G  , S  , V  ), with G  = G, by viewing documents as sequences of strings, each of length
 (n) = (n)/O(1). That is, we associate α = α1 · · · αt with the sequence (α1 , ..., αt ),
where each αi has length  (n). (At this point, the reader may think of  (n) = (n),
but actually we will use  (n) = (n)/4 in order to make room for some auxiliary
information.)
To motivate the actual construction, we consider first the following simpler schemes
all aimed at producing secure signatures for arbitrary (documents viewed as) sequences
of  (n)-bit long strings. The simplest scheme consists of just signing each of the strings
in the sequence. That is, the signature to the sequence (α1 , ..., αt ), is a sequence of
βi ’s, each being a signature (with respect to the length-restricted scheme) to the cor-
responding αi . This will not do, because an adversary, given the (single) signature
(β1 , β2 ) to the sequence (α1 , α2 ) with α1 = α2 , can present (β2 , β1 ) as a valid signature
to (α2 , α1 ) = (α1 , α2 ). So how about foiling this forgery by preventing a reordering
of the “atomic” signatures (i.e., the βi ’s); that is, how about signing the sequence
(α1 , ..., αt ) by applying the restricted scheme to each pair (i, αi ), rather than to αi it-
self? This will not do either, because an adversary, given a signature to the sequence
(α1 , α2 , α3 ), can easily present a signature to the sequence (α1 , α2 ). So we also need
to include in each (n)-bit string the total number of αi ’s in the sequence. But even
this is not enough, because given signatures to the sequences (α1 , α2 ) and (α1 , α2 ),
with α1 = α1 and α2 = α2 , an adversary can easily generate a signature to (α1 , α2 ).
Thus, we have to prevent the forming of new sequences of “basic signatures” by com-
bining elements from different signature sequences. This can be done by associating
(say, at random) an identifier with each sequence and incorporating this identifier in
each (n)-bit string to which the basic (restricted) signature scheme is applied. This
discussion yields the signature scheme presented next, where a signature to a message
(α1 , ..., αt ) consists of a sequence of (basic) signatures to statements of the (effec-
tive) form “the string αi is the i-th block, out of t blocks, in a message associate with
identifier r .”

Construction 6.2.3 (signing augmented blocks): Let  and (G, S, V ) be as in The-

orem 6.2.2. We construct a general signature scheme, (G  , S  , V  ), with G  = G, by
considering documents as sequences of strings. We construct S  and V  as follows,
using G  = G and  (n) = (n)/4:

Signing with S  : On input a signing-key s (in the range of G 1 (1n )) and a document
α ∈ {0, 1}∗ , algorithm S  first parses α into a sequence of blocks (α1 , ..., αt ), such
that α is uniquely reconstructed from the αi ’s and each αi is an  (n)-bit long string.4

4 The parsing rule should apply to strings of arbitrary length, regardless of whether or not this length is a multiple
of  (n). For example, we may parse α as (α1 , ..., αt ) such that α1 · · · αt = α · 10 j and j ∈ {0, 1, ...,  (n) − 1}.
(Note that under this parsing rule, if |α| is a multiple of  (n), then |α1 · · · αt | = |α| +  (n).)

509
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION


Next, S  uniformly selects r ∈ {0, 1} (n) . For i = 1, ..., t, algorithm S  computes

βi ← Ss (r, t, i, αi )

where i and t are represented as  (n)-bit long strings. That is, βi is essentially
a signature to the statement “αi is the i-th block, out of t blocks, in a sequence
associate with identifier r .” Finally, S  outputs as signature the sequence

(r, t, β1 , ...., βt )

Verification with V  : On input a verifying-key v (in the range of G 2 (1n )), a docu-
ment α ∈ {0, 1}∗ , and a sequence (r, t, β1 , ...., βt ), algorithm V  first parses α into
α1 , ..., αt  , using the same parsing rule as used by S  . Algorithm V  accepts if and
only if the following two conditions hold:

1. t  = t, where t  is obtained in the parsing of α and t is part of the alleged signature.

2. For i = 1, ..., t, it holds that Vv ((r, t, i, αi ), βi ) = 1, where αi is obtained in the
parsing of α and the rest are as in the corresponding parts of the alleged signature.

Clearly, the triplet (G  , S  , V  ) satisfies Definition 6.1.1. We need to show that is also
inherits the security of (G, S, V ). That is:

Proposition 6.2.4: Suppose that (G, S, V ) is an -restricted signature scheme that is

secure in the private-key (resp., public-key) model. Then (G  , S  , V  ), as defined in
Construction 6.2.3, is a full-fledged signature scheme that is secure in the private-key
(resp., public-key) model.

Theorem 6.2.2 follows immediately from Proposition 6.2.4.

Proof: Intuitively, ignoring the unlikely case that two messages signed by Ss were
assigned the same random identifier, a forged signature with respect to (G  , S  , V  )
must contain some Ss -signature that was not contained in any of the Ss -signatures
(provided in the attack). Thus, forgery with respect to (G  , S  , V  ) yields forgery with
respect to (G, S, V ). Indeed, the proof is by a reducibility argument, and it holds for
both the private-key and the public-key models.
Given an adversary A attacking the complex scheme (G  , S  , V  ), we construct an
adversary A that attacks the -restricted scheme, (G, S, V ). In particular, A invokes A
with input identical to its own input (which is the security parameter or the verification-
key, depending on the model), and uses its own oracle in order to emulate the oracle
Ss for A . This can be done in a straightforward manner; that is, algorithm A will
act as Ss does by using the oracle Ss . Specifically, A parses each query α  of A into
a corresponding sequence (α1 , ..., αt ), uniformly selects an identifier r  , and obtains
Ss -signatures to (r  , t  , j, α j ), for j = 1, ..., t  . When A outputs a document-signature
pair relative to the complex scheme (G  , S  , V  ), algorithm A tries to use this pair in
order to form a document-signature pair relative to the -restricted scheme, (G, S, V ).
We stress that from the point of view of adversary A , the distribution of keys and
oracle answers that A provides it with is exactly as in a real attack on (G  , S  , V  ).
510

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

This is a crucial point, because we use the fact that events that occur in a real attack
of A on (G  , S  , V  ) occur with the same probability in the emulation of (G  , S  , V  )
by A.
Assume that with (non-negligible) probability ε  (n), the (probabilistic polynomial-
time) algorithm A succeeds in existentially forging relative to the complex scheme
(G  , S  , V  ). We consider the following cases regarding the forging event:

1. The identifier supplied in the forged signature is different from all the random identi-
fiers supplied (by A) as part of the signatures given to A . In this case, each -restricted
signature supplied as part of the forged (complex) signature yields existential forgery
relative to the -restricted scheme.
Formally, let α(1) , ..., α (m) be the sequence of queries made by A  , and let
(1) (m)
(r (1) , t (1) , β ), ..., (r (m) , t (m) , β ) be the corresponding (complex) signatures sup-
(i) (i)
plied to A by A (using Ss to form the β ’s). It follows that each β consists of a
sequence of Ss -signatures to (n)-bit strings starting with r (i) ∈ {0, 1}(n)/4 , and that
the oracle Ss was invoked (by A) only on strings of this form. Let (α, (r, t, β1 , ..., βt ))
be the output of A , where α is parsed as (α1 , ..., αt ), and suppose that applying Vv
to the output of A yields 1 (i.e., the output is a valid document-signature pair for the
complex scheme). The case hypothesis states that r = r (i) , for all i’s. It follows that
each of the β j ’s is an Ss -signature to a string starting with r ∈ {0, 1}(n)/4 , and thus
different from all queries made to the oracle Ss . Thus, each pair ((r, t, i, αi ), βi )
is a valid document-signature pair (because Vv (α, (r, t, β1 , ..., βt )) = 1 implies
Vv ((r, t, i, αi ), βi ) = 1), with a document different from all queries made to Ss .
This yields a successful forgery with respect to the -restricted scheme.
2. The identifier supplied in the forged signature equals the random identifier supplied
(by A) as part of exactly one of the signatures given to A . In this case, existen-
tial forgery relative to the -restricted scheme is obtained by considering the rela-
tion between the output of A and the single supplied signature having the same
identifier.
As in the previous case, let α (1) , ..., α (m) be the sequence of queries made by A , and
(1) (m)
let (r (1) , t (1) , β ), ..., (r (m) , t (m) , β ) be the corresponding (complex) signatures
supplied to A by A. Let (α, (r, t, β1 , ..., βt )) be the output of A , where α is parsed as
(α1 , ..., αt ), and suppose that α = α (i) for all i’s and that Vv (α, (r, t, β1 , ..., βt )) = 1.
The hypothesis of the current case is that there exists a unique i so that r = r (i) .
We consider two subcases regarding the relation between t and t (i) :
r t = t (i) . In this subcase, each -restricted signature supplied as part of the forged
(complex) signature yields existential forgery relative to the -restricted scheme.
The argument is analogous to the one employed in the previous case. Specifically,
here each of the β j ’s is an Ss -signature to a string starting with (r, t), and thus dif-
ferent from all queries made to the oracle Ss (because these queries either start with

r (i ) = r or start with (r (i) , t (i) ) = (r, t)). Thus, each pair ((r, t, j, α j ), β j ) is a valid
document-signature pair with a document different from all queries made to Ss .
r t = t (i) . In this subcase, we use the hypothesis α = α (i) , which (combined with
(i) (i)
t = t (i) ) implies that there exists a j such that α j = α j , where α j is the j th
511
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

block in the parsing of α (i) . For this j, the string β j (supplied as part of the forged
complex-signature) yields existential forgery relative to the -restricted scheme.
Specifically, we have Vv ((r, t, j, α j ), β j ) = 1, whereas (r, t, j, α j ) is different
  (i  )
from each query (r (i ) , t (i ) , j  , α j  ) made by A to Ss .
  (i  )
Justification for (r, t, j, α j ) = (r (i ) , t (i ) , j  , α j  ): In case i  = i, it must hold

that r (i ) = r (by the [Case 2] hypothesis regarding the uniqueness of i s.t.
(i  )
r (i) = r ). Otherwise (i.e., in case i  = i), either j  = j or α j  = α j = α j ,
(i)

where the inequality is due to the hypothesis regarding j.

Thus, ((r, t, j, α j ), β j ) is a valid document-signature pair with a document dif-

ferent from all queries made to Ss .
3. The identifier supplied in the forged signature equals the random identifiers supplied
(by A) as part of at least two signatures given to A . In particular, it follows that two
signatures given to A use the same random identifier. The probability that this event
occurs is at most
 
m 
· 2− (n) < m 2 · 2−(n)/4
2
However, m = poly(n) (since A runs in polynomial-time), and 2−(n)/4 is negligible
(since  is super-logarithmic). So this case occurs with negligible probability and
may be ignored.
Note that A can easily determine which of the cases occurs and act accordingly.5 Thus,
assuming that A forges relative to the complex scheme with non-negligible probabil-
ity ε  (n), it follows that A forges relative to the length-restricted scheme with non-
negligible probability ε(n) ≥ ε  (n) − poly(n) · 2−(n)/4 , in contradiction to the proposi-
tion’s hypothesis.
Comment. We call the reader’s attention to the essential role of the hypothesis that
 is super-logarithmic in the proof of Proposition 6.2.4. Indeed, Construction 6.2.3 is
insecure in case (n) = O(log n). The reason is that by asking for polynomially many
signatures, the adversary may obtain two Ss -signatures that use the same (random)
identifier. Furthermore, with some care, these signatures yield existential forgery (see
Exercise 6).

6.2.2.2. Signing a Hash Value

In this subsection, we present an alternative method for constructing general signature
schemes out of length-restricted ones. Loosely speaking, the method consists of hashing
the document into a short (fixed-length) string (via an adequate hashing scheme), and
applying the length-restricted signature scheme to the resulting hash-value. This two-
stage process is referred to as the hash and sign paradigm.

5 This observation only saves us a polynomial factor in the forging probability. That is, if A did not know which
part of the forged complex-signature to use for its own forgery, it could have just selected one at random (and
be correct with probability 1/poly(n) because there are only poly(n)-many possibilities).

512

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

Let  and (G, S, V ) be as in Theorem 6.2.2. The second method of constructing a

general signature scheme out of (G, S, V ) consists of first hashing the document into
an (n)-bit long value and then applying the -restricted scheme to the hashed value.
Thus, in addition to an -restricted scheme, this method employs an adequate hashing
scheme. In particular, one way of implementing this method is based on “collision-free
hashing” (defined next). An alternative implementation, based on “universal one-way
hashing,” is deferred to Section 6.4.3.

Collision-Free Hashing Functions. Loosely speaking, a collision-free hashing scheme

(aka a collision-resistent hashing scheme) consists of a collection of functions {h s :
{0, 1}∗ → {0, 1}|s| }s∈{0,1}∗ such that given s and x it is easy to compute h s (x), but given
a random s it is hard to find x = x  such that h s (x) = h s (x  ).

Definition 6.2.5 (collision-free hashing functions): Let  : N → N. A collection of

functions {h s : {0, 1}∗ → {0, 1}(|s|) }s∈{0,1}∗ is called collision-free hashing if there ex-
ists a probabilistic polynomial-time algorithm I such that the following holds:

1. (admissible indexing – technical):6 For some polynomial p, all sufficiently large n’s,
and every s in the range of I (1n ), it holds that n ≤ p(|s|). Furthermore, n can be
computed in polynomial-time from s.
2. (efficient evaluation): There exists a polynomial-time algorithm that, given s and x,
returns h s (x).
3. (hard-to-form collisions): We say that the pair (x, x  ) forms a collision under the
function h if h(x) = h(x  ) but x = x  . We require that every probabilistic polynomial-
time algorithm, given I (1n ) as input, outputs a collision under h I (1n ) with negligible
probability. That is, for every probabilistic polynomial-time algorithm A, every pos-
itive polynomial p, and all sufficiently large n’s,

  1
Pr A(I (1n )) is a collision under h I (1n ) <
p(n)

where the probability is taken over the internal coin tosses of algorithms I and A.

The function  is called the range specifier of the collection.

Note that the range specifier must be super-logarithmic (or else one may easily find a
collision by selecting 2(n) + 1 different pre-images and computing their image under
the function). In Section 6.2.3, we show how to construct collision-free hashing func-
tions using claw-free collections. But first, we show how to use the former in order to
convert a length-restricted signature scheme into a full-fledged one.

6 This condition is made merely in order to avoid annoying technicalities. In particular, this condition allows the
collision-forming adversary to run for poly(n)-time (because by this condition n = poly(|s|)), as well as allows
for determining n from s. Note that |s| = poly(n) holds by definition of I .

513
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Construction 6.2.6 (hash and sign): Let  and (G, S, V ) be as in Theorem 6.2.2, and
let {h r : {0, 1}∗ → {0, 1}(|r |) }r ∈{0,1}∗ be as in Definition 6.2.5. We construct a general
signature scheme, (G  , S  , V  ), as follows:

Key-generation with G  : On input 1n , algorithm G  first invokes G to obtain (s, v) ←

G(1n ). Next, it invokes I , the indexing algorithm of the collision-free hashing col-
lection, to obtain r ← I (1n ). Finally, G  outputs the pair ((r, s), (r, v)), where (r, s)
serves as a signing-key and (r, v) serves as a verification-key.
Signing with S  : On input a signing-key (r, s) (in the range of G 1 (1n )) and a document
α ∈ {0, 1}∗ , algorithm S  invokes S once to produce and output Ss (h r (α)).
Verification with V  : On input a verifying-key (r, v) (in the range of G 2 (1n )), a docu-
ment α ∈ {0, 1}∗ , and an alleged signature β, algorithm V  invokes V and outputs
Vv (h r (α), β).
Note that the resulting signature scheme applies the original one once (per each invo-
cation of the resulting scheme). We stress that the length of resulting signatures only
depend on the length of the signing-key and is independent of the document being

signed; that is, |Sr,s (α)| = |Ss (h r (α))|, which in turn is bounded by poly(|s|, (|r |)).

Proposition 6.2.7: Suppose that (G, S, V ) is an -restricted signature scheme that

is secure in the private-key (resp., public-key) model. Suppose that {h r : {0, 1}∗ →
{0, 1}(|r |) }r ∈{0,1}∗ is indeed a collision-free hashing collection. Then (G  , S  , V  ), as
defined in Construction 6.2.6, is a full-fledged signature scheme that is secure in the
private-key (resp., public-key) model.

Proof: Intuitively, the security of (G  , S  , V  ) follows from the security of (G, S, V )

and the collision-freeness property of the collection {h r }. Specifically, forgery relative
to (G  , S  , V  ) can be obtained either by a forged S-signature to a hash-value different
from all hash-values that appeared in the attack or by forming a collision under the
hash function. The actual proof is by a reducibility argument. Given an adversary A
attacking the complex scheme (G  , S  , V  ), we construct an adversary A that attacks the
-restricted scheme, (G, S, V ), as well as an algorithm B forming collisions under the
hashing collection {h r }. Both A and B will have running time related to that of A . We
show if A is successful with non-negligible probability, than the same holds for either
A or B. Thus, in either case, we reach a contradiction. We start with the description of
algorithm A, which is designed to attack the -restricted scheme (G, S, V ). We stress
that almost the same description applies in both the private-key and public-key case.
On input x, which equals the security parameter 1n in the private-key case and a
verification-key v otherwise (i.e., in the public-key case), the adversary A operates as
follows. First, A uses I (the indexing algorithm of the collision-free hashing collection)
to obtain r ← I (1n ), exactly as done in the second step of G  . Next, A invokes A (on
input 1n or (r, v), depending on the case) and uses r as well as its own oracle Ss in order

to emulate the oracle Sr,s for A . The emulation is done in a straightforward manner;

that is, algorithm A will act as Sr,s does by using the oracle Ss (i.e., to answer query
q, algorithm A makes the query h r (q)). When A outputs a document-signature pair
514

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

relative to the complex scheme (G  , S  , V  ), algorithm A tries to use this pair in order
to form a document-signature pair relative to the -restricted scheme, (G, S, V ). That
is, if A outputs the document-signature pair (α, β), then A will output the document-
signature pair (h r (α), β).
As in the proof of Proposition 6.2.4, we stress that the distribution of keys and oracle
answers that A provides A is exactly as in a real attack of A on (G  , S  , V  ). This is
a crucial point, because we use the fact that events that occur in a real attack of A on
(G  , S  , V  ) occur with the same probability in the emulation of (G  , S  , V  ) by A.
Assume that with (non-negligible) probability ε  (n), the (probabilistic polynomial-
time) algorithm A succeeds in existentially forging relative to the complex scheme
(G  , S  , V  ). We consider the following two cases regarding the forging event, letting
(α (i) , β (i) ) denote the i-th query and answer pair made by A , and (α, β) denote the
forged document-signature pair that A outputs (in case of success):

Case 1: h r (α) = h r (α (i) ) for all i’s. (That is, the hash-value used in the forged signature
is different from all hash-values used in the queries to Ss .) In this case, the pair
(h r (α), β) constitutes a success in existential forgery relative to the -restricted
scheme.
Case 2: h r (α) = h r (α (i) ) for some i. (That is, the hash-value used in the forged sig-
nature equals the hash-value used in the i-th query to Ss , although α = α (i) .) In this
case, the pair (α, α (i) ) forms a collision under h r (and we do not obtain success in
existential forgery relative to the -restricted scheme).

Thus, if Case 1 occurs with probability at least ε  (n)/2, then A succeeds in its attack
on (G, S, V ) with probability at least ε (n)/2, which contradicts the security of the
-restricted scheme (G, S, V ). On the other hand, if Case 2 occurs with probability
at least ε (n)/2, then we derive a contradiction to the collision-freeness of the hashing
collection {h r : {0, 1}∗ → {0, 1}(|r |) }r ∈{0,1}∗ . Details (regarding the second case) follow.
We construct an algorithm, denoted B, that given r ← I (1n ), attempts to form col-
lisions under h r as follows. On input r , algorithm B generates (s, v) ← G(1n ) and
emulates the attack of A on this instance of the -restricted scheme, with the exception
that B does not invoke algorithm I to obtain an index of a hash function but rather uses
the index r (given to it as input). Recall that A, in turn, emulates an attack of A on

the signing-oracle Sr,s , and that A answers the query q  made by A by forwarding the
query q = h r (q ) to Ss . Thus, B actually emulates the attack of A (on the signing-oracle


Sr,s ) and does so in a straightforward manner; that is, to answer query q  made by A ,
algorithm B first obtains q = h r (q  ) (using its knowledge of r ) and then answers with
Ss (q) (using its knowledge of s). Finally, when A outputs a forged document-signature
pair, algorithm B checks whether Case 2 occurs (i.e., whether h r (α) = h r (α (i) ) holds
for some i), in which case it obtains (and outputs) a collision under hr . (Note that in
the public-key case, B invokes A on input (r, v), whereas in the private-key case, B
invokes A on input 1n . Thus, in the private-key case, B actually does not use r but
rather only uses an oracle access to hr .)
We stress that from the point of view of the emulated adversary A, the execu-
tion is distributed exactly as in its attack on (G, S, V ). Thus, since we assumed that
515
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

the second case occurs with probability at least ε  (n)/2 in a real attack, it follows
that B succeeds in forming a collision under h I (1n ) with probability at least ε (n)/2.
This contradicts the collision-freeness of the hashing functions, and the proposition
follows.

Comment. For the private-key case, the proof of Proposition 6.2.7 actually established
a stronger claim than stated. Specifically, the proof holds even for a weaker definition of
collision-free hashing in which the adversary is not given a description of the hashing
function, but can rather obtain its value at any pre-image of its choice. This observation
is further pursued in Section 6.3.1.3.

On Using the Hash-and-Sign Paradigm in Practice. The hash-and-sign paradigm,

underlying Construction 6.2.6, is often used in practice. Specifically, a document is
signed using a two-stage process: First, the document is hashed into a (relatively) short
bit string, and next, a basic signature scheme is applied to the resulting string. One
appealing feature of this process is that the length of resulting signatures only depends
on the length of the signing-key (and is independent of the document being signed). We
stress that this process yields a secure signature scheme only if the hashing scheme is
collision-free (as defined previously). In Section 6.2.3, we present several constructions
of collision-free hashing functions (based on general assumptions). Alternatively, one
may indeed postulate that certain off-the-shelf products (such as MD5 or SHA) are
collision-free, but such assumptions need to be seriously examined (and indeed may
turn out false).7 We stress that using a hashing scheme, in the two-stage (hash-and-
sign) process, without seriously evaluating whether or not it is collision-free is a very
dangerous practice.
We comment that a variant on the hash-and-sign paradigm will be presented in
Construction 6.4.30. The two variants are compared in Section 6.4.3.4.

6.2.3.* Constructing Collision-Free Hashing Functions

In view of the relevance of collision-free hashing to signature schemes, we now take a
small detour from the main topic and consider the construction of collision-free hash-
ing. Most importantly, we show how to construct collision-free hashing functions using
a claw-free collection of permutations. In addition, we show two different construc-
tions that use a restricted type of collision-free hashing in order to obtain full-fledged
collision-free hashing.

6.2.3.1. A Construction Based on Claw-Free Permutations

In this subsection, we show how to construct collision-free hashing functions using a
claw-free collection of permutations as defined in Section 2.4.5 of Volume 1. Recall
that such a collection consists of pairs of permutations, ( f s0 , fs1 ), such that both f sσ ’s

7 See, for example, [76].

516

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

are permutations over a set Ds , augmented with a probabilistic polynomial-time index

selection algorithm I such that the following conditions hold:

1. The domain is easy to sample: There exists a probabilistic polynomial-time algorithm

that, given s, outputs a string uniformly distributed over Ds .
2. The permutations are easy to evaluate: There exists a polynomial-time algorithm
that, given s, σ and x ∈ Ds , outputs f sσ (x).
3. It is hard to form claws: Every probabilistic polynomial-time algorithm, given
s ← I (1n ), outputs a pair (x, y) such that f s0 (x) = f s1 (y) with at most negligible
probability. That is, a pair (x, y) satisfying f s0 (x) = f s1 (y) is called a claw for index
s . (We stress that x = y may hold.) Then, it is required that for every probabilistic
polynomial-time algorithm, A , every positive polynomial p(·), and all sufficiently
large n’s
  1
Pr A (I (1n )) ∈ C I (1n ) <
p(n)

where Cs denote the set of claws for index s.

Note that since f s0 and f s1 are permutations over the same set, many claws do exists
(i.e., |Cs | = |Ds |). However, the third condition postulates that for s generated by I (1n ),
such claws are hard to find. We may assume, without loss of generality, that for some
 : N → N and all s’s, it holds that Ds ⊆ {0, 1}(|s|) . Indeed,  must be polynomially
bounded. For simplicity, we assume that I (1n ) ∈ {0, 1}n . Recall that such collections
of permutation pairs can be constructed based on the standard DLP or factoring in-
tractability assumptions (see Section 2.4.5).

Construction 6.2.8 (collision-free hashing based on claw-free permutations pairs):

Given an index selecting algorithm I for a collection of permutation pairs {( f s0 , f s1 )}s
as in the foregoing discussion, we construct a collection of hashing functions {h (s,r ) :
{0, 1}∗ → {0, 1}|r | }(s,r )∈{0,1}∗ ×{0,1}∗ as follows:

Index selection algorithm: On input 1n , we first invoke I to obtain s ← I (1n ), and next
use the domain sampler to obtain a string r that is uniformly distributed in Ds .
We output the index (s, r ) ∈ {0, 1}n × {0, 1}(n) , which corresponds to the hashing
function

def
h (s,r ) (x) = f sy1 f sy2 · · · f syt (r )

where y1 · · · yt is a prefix-free encoding of x; that is, for any x = x  the coding

of x is not a prefix of the coding of x  . For example, we may code x 1 x 2 · · · xm by
x1 x1 x 2 x 2 · · · x m xm 01.
Evaluation algorithm: Given an index (s, r ) and a string x, we compute h (s,r ) (x) in
a straightforward manner. That is, first we compute the prefix-free encoding of x,
denoted y1 · · · yt . Next, we use the evaluation algorithm of the claw-free collection
y y y
to compute f s 1 f s 2 · · · f s t (r ), which is the desired output.
517
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Actually, as will become evident from the proof of Proposition 6.2.9, as far as
Construction 6.2.8 is concerned, the definition of claw-free permutations can be re-
laxed: We do not need an algorithm that, given an index s, generates a uniformly
distributed element in Ds ; any efficient algorithm that generates elements in Ds will do
(regardless of the distribution induced on Ds , and in particular, even if the algorithm
always outputs the same element in Ds ).

Proposition 6.2.9: Suppose that the collection of permutation pairs {( f s0 , f s1 )}s , to-
gether with the index-selecting algorithm I , constitutes a claw-free collection. Then,
the function ensemble {h (s,r ) : {0, 1}∗ → {0, 1}|r | }(s,r )∈{0,1}∗ ×{0,1}∗ as defined in Con-
struction 6.2.8 constitutes a collision-free hashing with a range specifying function 
satisfying  (n + (n)) = (n).

Proof: Intuitively, forming collisions under h (s,r ) means finding two different sequences
of functions from { f s0 , f s1 } that (when applied to r ) yield the same image (e.g., f s1 ◦
f s0 ◦ f s0 (r ) = f s1 ◦ f s1 (r ) ◦ f s1 (r )). Since these two sequences cannot be a prefix of one
another, it must be that somewhere along the process (of applying these f sσ ’s), the
application of two different functions yields the same image (i.e., a claw).
The proof is by a reducibility argument. Given an algorithm A that on input (s, r )
forms a collision under h (s,r ) , we construct an algorithm A that on input s forms a
claw for index s. On input s (supposedly generated by I (1n )), algorithm A selects r
(uniformly) in Ds , and invokes algorithm A on input (s, r ). Suppose that A outputs a
pair (x, x  ) so that h (s,r ) (x) = h (s,r ) (x  ) but x = x  . Without loss of generality,8 assume
that the coding of x equals y1 · · · yi−1 0zi+1 · · · z t , and that the coding of x  equals

y1 · · · yi−1 1z i+1 · · · z t  . By the definition of h (s,r ) , it follows that

z z
f sy1 · · · f syi−1 f s0 f szi+1 · · · f sz t (r ) = f sy1 · · · f syi−1 f s1 f s i+1 · · · f s t (r ) (6.1)

Since each of the f sσ ’s is 1-1, Eq. (6.1) implies that

z z
f s0 f szi+1 · · · f sz t (r ) = f s1 f s i+1 · · · f s t (r ) (6.2)

def def z z

Computing w = f s i+1 · · · f szt (r ) and w  = f s i+1 · · · f s t (r ), algorithm A obtains a pair
z

(w, w  ) such that f s0 (w) = f s1 (w  ). Thus, algorithm A forms claws for index I (1n )
with probability that is lower-bounded by the probability that A forms a collision
under h I  (1n ) , where I  is the index-selection algorithm as defined in Construction 6.2.8.
Using the hypothesis that the collection of pairs (together with I ) is claw-free, the
proposition follows.

8 Let C(x) (resp., C(x  )) denote the prefix-free coding of x (resp., x  ). Then C(x) is not a prefix of C(x  ), and
C(x  ) is not a prefix of C(x). It follows that C(x) = uv and C(x  ) = uv  , where v and v  differ in their leftmost
bit. Without loss of generality, we may assume that the leftmost bit of v is 0, and the leftmost bit of v  is 1.

518

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

6.2.3.2. Collision-Free Hashing via Block-Chaining

In this subsection, we show how a restricted type of Collision-Free Hashing (CFH) can
be used to obtain full-fledge collision-free hashing (CFH). Specifically, we refer to the
following restriction of Definition 6.2.5:

Definition 6.2.10 (length-restricted collision-free hashing functions): Let  ,  : N →


N. A collection of functions {h s : {0, 1} (|s|) → {0, 1}(|s|) }s∈{0,1}∗ is called  -restricted
collision-free hashing if there exists a probabilistic polynomial-time algorithm I such
that the following holds:
1. (admissible indexing – technical): As in Definition 6.2.5.
2. (efficient evaluation): There exists a polynomial-time algorithm that, given s and

x ∈ {0, 1} (|s|) , returns h s (x).
3. (hard-to-form collisions): As in Definition 6.2.5, we say that the pair (x, x  ) forms
a collision under the function h if h(x) = h(x  ) but x = x  . We require that ev-
ery probabilistic polynomial-time algorithm, given I (1n ) as input, outputs a pair
 
in {0, 1} (|s|) × {0, 1} (|s|) that forms a collision under h I (1n ) with negligible proba-
bility. That is, for every probabilistic polynomial-time algorithm A, every positive
polynomial p, and all sufficiently large n’s,
  n  1
Pr A(I (1n )) ∈ {0, 1}2· (|I (1 )|) is a collision under h I (1n ) <
p(n)
where the probability is taken over the internal coin tosses of algorithms I and A.
Indeed, we focus on the case  (n) = poly(n), or else the hardness condition holds
vacuously (since no polynomial-time algorithm can print a pair of strings of super-
polynomial length). On the other hand, we only care about the case  (n) > (n) (oth-
erwise the functions may be 1-1). Finally, recall that  must be super-logarithmic.
Following is a simple construction of full-fledge collision-free hashing based on any
2-restricted collision-free hashing (see also Figure 6.2).

Construction 6.2.11 (from 2-restricted CFH to full-fledged CFH): Let {h s :

{0, 1}2(|s|) → {0, 1}(|s|) }s∈{0,1}∗ be a collection of functions. Consider the collection
{h s : {0, 1}∗ → {0, 1}2(|s|) }s∈{0,1}∗ , where h s (x) is defined by the following process,
which we call block-chaining:
def
1. Break x into t = |x|/(|s|) consecutive blocks, while possibly padding the last
block with 0’s, such that each block has length (|s|). Denote these (|s|)-bit long
blocks by x1 , ..., xt . That is, x 1 · · · x t = x0t·(|s|)−|x| .

X1 X2 X3 X4 X5 X6 X7

y1 h’s y2 h’s y3 h’s y4 h’s y5 h’s y6 h’s y7

Figure 6.2: Collision-free hashing via block-chaining (for t = 7).

519
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

For the sake of uniformity, in case |x| ≤ (|s|), we let t = 2 and x 1 x2 = x02(|s|)−|x| .
On the other hand, we may assume that |x| < 2(|s|) , and so |x| can be represented
by an (|s|)-bit long string.9
def
2. Let y1 = x1 . For i = 2, ..., t, compute yi = h s (yi−1 xi ).
3. Set h s (x) to equal (yt , |x|).
An interesting property of Construction 6.2.11 is that it allows for computing the hash-
value of an input string while processing the input in an on-line fashion; that is, the
implementation of the hashing process may process the input x in a block-by-block
manner, while storing only the current block and a small amount of state information
(i.e., the current yi and the number of blocks encountered so far). This property is
important in applications in which one wishes to hash a long stream of input bits.

Proposition 6.2.12: Let {h s : {0, 1}2(|s|) → {0, 1}(|s|) }s∈{0,1}∗ and {h s : {0, 1}∗ →
{0, 1}2(|s|) }s∈{0,1}∗ be as in Construction 6.2.11, and suppose that the former is a col-
lection of 2-restricted collision-free hashing functions. Then the latter constitutes a
(full-fledged) collection of collision-free hashing functions.

Proof: Recall that forming a collision under h s means finding x = x  such that h s (x) =
h s (x  ). By the definition of h s , this means that (yt , |x|) = h s (x) = h s (x  ) = (yt , |x  |),
where t, t  and yt , yt are determined by h s (x) and h s (x  ). In particular, it follows that
|x| = |x  | and so t = t  (where, except when |x| ≤ (|s|), it holds that t = |x|/(|s|) =
|x  |/(|s|) = t  ). Recall that yt = yt and consider two cases:

Case 1: If (yt−1 , xt ) = (yt−1 , x t ), then we obtain a collision under h s (since
   
h s (yt−1 xt ) = yt = yt = h s (yt−1 x t )), and derive a contradiction to its collision-free
hypothesis.

Case 2: Otherwise (yt−1 , x t ) = (yt−1 , x t ), and we consider the two corresponding cases
 
with respect to the relation of (yt−2 , x t−1 ) to (yt−2 , xt−1 ); that is, we further consider
 
whether or not (yt−2 , xt−1 ) equals (yt−2 , x t−1 ).
Eventually, since x = x  , we get to a situation in which yi = yi and (yi−1 , xi ) =

(yi−1 , xi ), which is handled as in the first case.
We now provide a formal implementation of this intuitive argument. Suppose toward the
contradiction that there exists a probabilistic polynomial-time algorithm A that on input
s forms a collision under h s (with certain probability). Then, we construct an algorithm
that will, with similar probability, succeed to form a suitable (i.e., length-restricted)
collision under h s . Algorithm A (s) operates as follows:
1. A (s) invokes A(s) and obtains (x, x  ) ← A(s).
If either h s (x) = h s (x  ) or x = x  , then A failed, and A halts without output. In the
sequel, we assume that h s (x) = h s (x  ) and x = x  .

9 The adversary trying to form collisions with respect to h s runs in poly(|s|)-time. Using (|s|) = ω(log |s|), it
follows that such an adversary cannot output a string of length 2(|s|) . (The same also holds, of course, for
legitimate usage of the hashing function.)

520

www.Ebook777.com
 6.2 LENGTH-RESTRICTED SIGNATURE SCHEME

2. A (s) computes t, x1 , ..., x t and y1 , ..., yt (resp., t  , x1 , ..., xt and y1 , ..., yt ) as
in Construction 6.2.11. Next, A (s) finds an i ∈ {2, ..., t} such that yi = yi and

(yi−1 , xi ) = (yi−1 , xi ), and outputs the pair (yi−1 xi , yi−1

xi ). (We will show next that
such an i indeed exists.)
Note that (since h s (x) = h s (x  )) it holds that t = t  and yt = yt . On the other hand,
(x1 , ..., x t ) = (x1 , ..., xt ). As argued in the motivating discussion, it follows that there
exists an i ∈ {2, ..., t} such that yi = yi and (yi−1 , xi ) = (yi−1 
, xi ).

On the existence of a suitable i (more details): Suppose, toward the contra-

diction that, for every i ∈ {2, ..., t}, it holds that either yi = yi or (yi−1 , xi ) =

(yi−1 , xi ). Using the hypothesis yt = yt , it follows (by descending induction on
j) that (y j−1 , x j ) = (y j−1 , x j ), for j = t, ..., 2. Using y1 = x 1 and y1 = x1 , it
follows that x j = x j for every j = 1, ..., t, which contradicts the hypothesis
(x1 , ..., x t ) = (x 1 , ..., xt ).

Clearly, the output pair (yi−1 xi , yi−1 xi ) constitutes a collision under h s (because
     
h s (yi−1 xi ) = yi = yi = h s (yi−1 xi ), whereas yi−1 xi = yi−1 xi ).
Thus, whenever A(s) forms a collision under h s , it holds that A (s) outputs a pair of
2(s)-bit long strings that form a collision under h s . The proposition follows.

Variants on Construction 6.2.11. The said construction can be generalized to use

any (non-trivial) length-restricted collision-free hashing. That is, for any  > , let

{h s : {0, 1} (|s|) → {0, 1}(|s|) }s∈{0,1}∗ be a collection of  -restricted collision-free hash-
ing functions, and consider a parsing of the input string x into a sequence x1 , ..., x t
of ( (|s|) − (|s|))-bit long blocks. Then we get a full-fledged collision-free hashing
family {h s : {0, 1}∗ → {0, 1}2(|s|) } by letting h s (x) = (yt , |x|), where yi = h s (yi−1 xi )
for i = 2, ..., t. (Construction 6.2.11 is obtained as a special case, for  (n) = 2(n).) In
case  (n) − (n) = ω(log n), we obtain another variant by letting h s (x) = h s (yt , |x|)
(rather than h s (x) = (yt , |x|)), where yt is as in Construction 6.2.11. The latter variant is
quite popular. In establishing its security, when considering a collision h s (x) = h s (x  ),
we distinguish the case (yt , |x|) = (yt , |x  |) (which is handled as in the proof of
Proposition 6.2.12) from the case (yt , |x|) = (yt , |x  |) (which yields an immediate col-
lision under h s ).

6.2.3.3. Collision-Free Hashing via Tree-Hashing

Using 2-restricted collision-free hashing functions, we now present an alternative
construction of (full-fledged) collision-free hashing functions. The alternative con-
struction will have the extra property of supporting verification of a bit in the input
(with respect to the hash-value) within complexity that is independent of the length of
the input.

Construction 6.2.13 (from 2-restricted CFH to full-fledged CFH – an alternative

construction (see also Figure 6.3.)): Let {h s : {0, 1}2(|s|) → {0, 1}(|s|) }s∈{0,1}∗ be a
521
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

X1 X2 X3 X4 X5 X6 X7 X8

y3,1 y y y y y y y
3,2 3,3 3,4 3,5 3,6 3,7 3,8

h’s h’s h’s h’s

y y y y
2,1 2,2 2,3 2,4

h’s h’s
y y
1,1 1,2

h’s
y
0,1
Figure 6.3: Collision-free hashing via tree-chaining (for t = 8).

collection of functions. Consider the collection {h s : {0, 1}∗ → {0, 1}2(|s|) }s∈{0,1}∗ ,
where h s (x) is defined by the following process, called tree-hashing:
def
1. Break x into t = 2 log2 (|x|/(|s|)) consecutive blocks, while possibly adding dummy
0-blocks and padding the last block with 0’s, such that each block has length (|s|).
Denote these (|s|)-bit long blocks by x 1 , ..., xt . That is, x 1 · · · x t = x0t·(|s|)−|x| .
Let d = log2 t, and note that d is a positive integer.
Again, for the sake of uniformity, in case |x| ≤ (|s|), we let t = 2 and x 1 x 2 =
x02(|s|)−|x| . On the other hand, again, we assume that |x| < 2(|s|) , and so |x| can be
represented by an (|s|)-bit long string.
def
2. For i = 1, ..., t, let yd,i = xi .
3. For j = d −1, ..., 1, 0 and i = 1, ..., 2 j , compute y j,i = h s (y j+1,2i−1 y j+1,2i ).
4. Set h s (x) to equal (y0,1 , |x|).

That is, hashing is performed by placing the (|s|)-bit long blocks of x at the leaves of
a binary tree of depth d, and computing the values of internal nodes by applying h s to
the values associated with the two children (of the node). The final hash-value consists
of the value associated with the root (i.e., the only level-0 node) and the length of x.

Proposition 6.2.14: Let {h s : {0, 1}2(|s|) → {0, 1}(|s|) }s∈{0,1}∗ and {h s : {0, 1}∗ →
{0, 1}2(|s|) }s∈{0,1}∗ be as in Construction 6.2.13, and suppose that the former is a col-
lection of 2-restricted collision-free hashing functions. Then the latter constitutes a
(full-fledged) collection of collision-free hashing functions.

Proof Sketch: Recall that forming a collision under h s means finding x = x  such that
h s (x) = h s (x  ). By the definition of h s , this means that (y0,1 , |x|) = h s (x) = h s (x  ) =
 
(y0,1 , |x  |), where (t, d) and y0,1 (resp., (t  , d  ) and y0,1 ) are determined by h s (x) (resp.,
h s (x )). In particular, it follows that |x| = |x | and so d = d  (because 2d = t = t  =
 

522

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

 
2d ). Recall that y0,1 = y0,1 , and let us state this fact by saying that for j = 0 and for
every i ∈ {1, ..., 2 }, it holds that y j,i = y j,i . Starting with j = 0, we consider two cases
j

(for level j + 1 in the tree):

Case 1: If for some i ∈ {1, ..., 2 j+1 } it holds that y j+1,i = y j+1,i , then we obtain a
collision under h s , and derive a contradiction to its collision-free hypothesis. Specif-
def
ically, the collision is obtained because z = y j+1,2 i/2 −1 y j+1,2 i/2 is different from
 def
z = y j+1,2 i/2 −1 y j+1,2 i/2 , whereas h s (z) = y j, i/2 = y j, i/2 = h s (z  ).
Case 2: Otherwise for every i ∈ {1, ..., 2 j+1
}, it holds that y j+1,i = y j+1,i . In this case,
we consider the next level.
Eventually, since x = x  , we get to a situation in which for some j ∈ {1, ..., d −
def
1} and some i ∈ {1, ..., 2 j+1 }, it holds that z = y j+1,2 i/2 −1 y j+1,2 i/2 is different
def
from z  = y j+1,2 i/2 −1 y j+1,2 i/2 , whereas h s (z) = y j, i/2 = y j, i/2 = h s (z  ). This
situation is handled as in the first case.
The actual argument proceeds as in the proof of Proposition 6.2.12.

A Local Verification Property. Construction 6.2.13 has the extra property of support-
ing efficient verification of bits in x with respect to the hash-value. That is, suppose
that for a randomly selected h s , one party holds x and the other party holds h s (x).
Then, for every i, the first party may provide a short (efficiently verifiable) certifi-
cate that xi is indeed the i-th block of x. The certificate consists of the sequence
of pairs (yd,2 i/2 −1 , yd,2 i/2 ), ..., (y1,2 i/2d −1 , y1,2 i/2d ), where d and the y j,k ’s are
computed as in Construction 6.2.13 (and (y0,1 , |x|) = h s (x)). The certificate is ver-
ified by checking whether or not y j−1, i/2d− j+1 = h s (y j,2 i/2d− j+1 −1 y j,2 i/2d− j+1 ), for
every j ∈ {1, ..., d}. Note that if the first party can present two different values for
the i-th block of x along with corresponding certificates, then it can also form col-
lisions under h s . Construction 6.2.13 and its local-verification property were already
used in this work (i.e., in the construction of highly- efficient argument systems, pre-
sented in Section 4.8.4 of Volume 1). Jumping ahead, we note the similarity between
the local-verification property of Construction 6.2.13 and the authentication-tree of
Section 6.4.2.2.

6.3. Constructions of Message-Authentication Schemes

In this section, we present several constructions of secure message-authentication

schemes (referred to earlier as secure private-key signature schemes). Here, we some-
times refer to such a scheme by the popular abbreviation MAC (which actually abbre-
viates the more traditional term of a Message Authentication Code).

6.3.1. Applying a Pseudorandom Function to the Document

A scheme for message authentication can be obtained by applying a pseudorandom
function (specified by the key) to the message (which one wishes to authenticate). The
523
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

simplest implementation of this idea is presented in Section 6.3.1.1, whereas more

sophisticated implementations are presented in Sections 6.3.1.2 and 6.3.1.3.

6.3.1.1. A Simple Construction and a Plausibility Result

Message-authentication schemes can be easily constructed using pseudorandom
functions (as defined in Section 3.6 of Volume 1). Specifically, by Theorem 6.2.2,
it suffices to construct an -restricted message-authentication scheme for any super-
logarithmically growing . Indeed, this is our starting point.

Construction 6.3.1 (an -restricted MAC based on pseudorandom functions): Let 

be a super-logarithmically growing function, and { f s : {0, 1}(|s|) → {0, 1}(|s|) }s∈{0,1}∗
be as in Definition 3.6.4. We construct an -restricted message-authentication scheme,
(G, S, V ), as follows:

Key-generation with G: On input 1n , we uniformly select s ∈ {0, 1}n , and output the
key-pair (s, s). (Indeed, the verification-key equals the signing-key.)
Signing with S: On input a signing-key s ∈ {0, 1}n and an (n)-bit string α, we compute
and output f s (α) as a signature of α.
Verification with V : On input a verification-key s ∈ {0, 1}n , an (n)-bit string α, and
an alleged signature β, we accept if and only if β = f s (α).

Indeed, signing amounts to applying f s to the given document string, and verification
amounts to comparing a given value to the result of applying f s to the document. Analo-
gous constructions can be presented by using the generalized notions of pseudorandom
functions defined in Definitions 3.6.9 and 3.6.12 (see further comments in the follow-
ing subsections). In particular, using a pseudorandom function ensemble of the form
{ f s : {0, 1}∗ → {0, 1}|s| }s∈{0,1}∗ , we obtain a general message-authentication scheme
(rather than a length-restricted one). In the following proof, we only demonstrate the
security of the -restricted message-authentication scheme of Construction 6.3.1. (The
security of the general message-authentication scheme can be established analogously;
see Exercise 8.)

Proposition 6.3.2: Suppose that { f s : {0, 1}(|s|) → {0, 1}(|s|) }s∈{0,1}∗ is a pseudoran-
dom function, and that  is a super-logarithmically growing function. Then Construc-
tion 6.3.1 constitutes a secure -restricted message-authentication scheme.

Proof: The proof follows the general methodology suggested in Section 3.6.3. Specif-
ically, we consider the security of an ideal scheme in which the pseudorandom function
is replaced by a truly random function (mapping (n)-bit long strings to (n)-bit long
strings). Clearly, an adversary that obtains the values of this random function at ar-
guments of its choice cannot predict its value at a new point with probability greater
than 2−(n) . Thus, an adversary attacking the ideal scheme may succeed in existen-
tial forgery with at most negligible probability. The same must hold for any efficient
adversary that attacks the actual scheme, because otherwise such an adversary yields
524

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

a violation of the pseudorandomness of { f s : {0, 1}(|s|) → {0, 1}(|s|) }s∈{0,1}∗ . Details

follow.
The actual proof is by a reducibility argument. Given a probabilistic polynomial-time
A attacking the scheme (G, S, V ), we consider what happens when A attacks an ideal
scheme in which a random function is used instead of a pseudorandom one. That is, we
refer to two experiments:

1. Machine A attacks the actual scheme: On input 1n , machine A is given oracle access
to (the signing process) f s : {0, 1}(n) → {0, 1}(n) , where s is uniformly selected in
{0, 1}n . After making some queries of its choice, A outputs a pair (α, β), where
α is different from all its queries. Machine A is deemed successful if and only if
β = f s (α).
2. Machine A attacks the ideal scheme: On input 1n , machine A is given oracle access
to a function φ : {0, 1}(n) → {0, 1}(n) , uniformly selected among all such possible
functions. After making some queries of its choice, A outputs a pair (α, β), where
α is different from all its queries. Again, A is deemed successful if and only if
β = φ(α).
Clearly, A’s success probability in this experiment is at most 2−(n) , which is a
negligible function (since  is super-logarithmic).

Assuming that A’s success probability in the actual attack is non-negligible, we derive
a contradiction to the pseudorandomness of the function ensemble { fs }. Specifically,
we consider a distinguisher D that, on input 1n and oracle access to a function f :
{0, 1}(n) → {0, 1}(n) , behaves as follows: First D emulates the actions of A, while
answering A’s queries using its oracle f . When A outputs a pair (α, β), the distinguisher
makes one additional oracle query to f and outputs 1 if and only if f (α) = β.
Note that when f is selected uniformly among all possible {0, 1}(n) → {0, 1}(n)
functions, D emulates an attack of A on the ideal scheme, and thus outputs 1 with
negligible probability (as explained in the beginning of the proof). On the other hand,
if f is uniformly selected in { f s }s∈{0,1}n , then D emulates an attack of A on the actual
scheme, and thus (due to the contradiction hypothesis) outputs 1 with non-negligible
probability. We reach a contradiction to the pseudorandomness of { f s }s∈{0,1}n . The
proposition follows.

A Plausibility Result. Combining Theorem 6.2.2, Proposition 6.3.2, and Corol-

lary 3.6.7, it follows that the existence of one-way functions implies the existence
of message-authentication schemes. The converse also holds; see Exercise 7. Thus, we
have:

Theorem 6.3.3: Secure message-authentication schemes exist if and only if one-way

functions exist.

In contrast to the feasibility result stated in Theorem 6.3.3, we now present alterna-
tive ways of using pseudorandom functions to obtain secure message-authentication
schemes (MACs). These alternatives yield more efficient schemes, where efficiency is
525
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

measured in terms of the length of the signatures and the time it takes to produce and
verify them.

6.3.1.2.* Using the Hash-and-Sign Paradigm

Theorem 6.3.3 was proved by combining the length-restricted MAC of Construc-
tion 6.3.1 with the simple but wasteful idea of providing signatures (authentication
tags) for each block of the document (i.e., Construction 6.2.3). In particular, the signa-
ture produced this way is longer than the document. Instead, here we suggest using the
second method of converting length-restricted MACs into full-fledged ones; that is, the
hash-and-sign method of Construction 6.2.6. This will yield signatures of a fixed length
(i.e., independent of the length of the document). Combining the hash-and-sign method
with a length-restricted MAC of Construction 6.3.1 (which is based on pseudorandom
functions), we obtain the following construction:

Construction 6.3.4 (hash and sign using pseudorandom functions): Let { f s :

{0, 1}|s| → {0, 1}|s| }s∈{0,1}∗ be a pseudorandom function ensemble and {h r : {0, 1}∗ →
{0, 1}|r | }r ∈{0,1}∗ be a collection of collision-free hashing functions. Furthermore, for
simplicity we assume that, when invoked on input 1n , the indexing algorithm I of the
collision-free hashing collection outputs an n-bit long index. The general message-
authentication scheme, (G, S, V ), is as follows:

Key-generation with G: On input 1n , algorithm G selects uniformly s ∈ {0, 1}n , and

invokes the indexing algorithm I to obtain r ← I (1n ). The key-pair output by G is
((r, s), (r, s)).
Signing with S: On input a signing-key (r, s) in the range of G 1 (1n ) and a document
α ∈ {0, 1}∗ , algorithm S outputs the signature/tag f s (h r (α)).
Verification with V : On input a verification-key (r, s) in the range of G 2 (1n ), a doc-
ument α ∈ {0, 1}∗ , and an alleged signature β, algorithm outputs 1 if and only if
f s (h r (α)) = β.

Combining Propositions 6.2.7 and 6.3.2, it follows that Construction 6.3.4 constitutes
a secure message-authentication scheme (MAC), provided that the ingredients are
as postulated. In particular, this means that Construction 6.3.4 yields a secure MAC,
provided that collision-free hashing functions exist (and are used in Construction 6.3.4).
While this result uses a seemingly stronger assumption than the existence of one-way
functions (used to establish the Theorem 6.3.3), it yields more efficient MACs, both
in terms of signature length (as discussed previously) and authentication time (to be
discussed next).
Construction 6.3.4 yields faster signing and verification algorithms than the construc-
tion resulting from combining Constructions 6.2.3 and 6.3.1, provided that hashing a
long string is less time-consuming than applying a pseudorandom function to it (or to all
its blocks). The latter assumption is consistent with the current state of the art regarding
the implementation of both primitives. Further speed improvements are discussed in
Section 6.3.1.3.
526

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

An Alternative Presentation. Construction 6.3.4 was analyzed by invoking the hash-

and-sign paradigm (i.e., Proposition 6.2.7), while referring to the fixed-length MAC
arising from the pseudorandom function ensemble { f s : {0, 1}|s| → {0, 1}|s| }s∈{0,1}∗ . An
alternative analysis may proceed by first establishing that {gs,r = f s ◦ h r }s∈{0,1}∗ ,r ←I (1|s| )
is a generalized pseudorandom function (as per Definition 3.6.12), and next observing
that any such ensemble yields a full-fledged MAC (see Exercise 8).

6.3.1.3.* A Variation on the Hash-and-Sign Paradigm

(or Using Non-Cryptographic Hashing Plus Hiding)
Construction 6.3.4 combines the use of a collision-free hashing function with the ap-
plication of a pseudorandom function. Here we take another step toward speeding-up
message authentication by showing that the collision-free hashing can be replaced
with ordinary (i.e., non-cryptographic) hashing, provided that a pseudorandom func-
tion (rather than a generic MAC) is applied to the result. Consequently, we also
reduce the intractability assumptions used in the analysis of the construction. Be-
fore getting into details, let us explain why we can use non-cryptographic hash-
ing and why this may lead to reduced intractability assumptions and to efficiency
improvements.
r Since we are in the private-key setting, the adversary does not get the description
of the hash function used in the hash-and-sign process. Furthermore, applying the
pseudorandom function to the hash-value hides it from the adversary. Thus, when
trying to form collisions under the hash function, the adversary is in “total darkness”
and may only rely on the collision probability of the hashing function (as defined
next). (Recall that in case the adversary fails to form a collision, it must succeed in
forging with respect to the length-restricted scheme if it wishes to forge with respect
to the full-fledged scheme.)
r Using an ordinary hashing instead of a collision-free hash function means that the
only intractability assumption used is the existence of pseudorandom functions (or,
equivalently, of one-way functions).
The reason that applying an ordinary hashing, rather than a collision-free hash func-
tion, may yield an efficiency improvement is that the former is likely to be more
efficient than the latter. This is to be expected, given that ordinary hashing need only
satisfy a weak (probabilistic) condition, whereas collision-free hashing refers to a
more complicated (intractability) condition.10
By ordinary hashing we mean function ensembles as defined in Section 3.5.1.1 of
Volume 1. For starters, recall that these are collections of functions mapping (n)-
bit strings to m(n)-bit strings. These collections are associated with a set of strings,
m(n) m(n) m(n)
denoted S(n) , and we may assume that S(n) ≡ {0, 1}n . Specifically, we call {S(n) }n∈N

10 This intuition may not hold when comparing a construction of ordinary hashing that is rigorously ana-
lyzed with an ad hoc suggestion of a collision-free hashing. But it certainly holds when comparing the
former to the constructions of collision-free hashing that are based on a well-established intractability
assumption.

527
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

a hashing ensemble if it satisfies the following three conditions:

1. Succinctness: n = poly((n), m(n)).
2. Efficient evaluation: There exists a polynomial-time algorithm that, on input a rep-
resentation of a function, h (in S(n) ), and a string x ∈ {0, 1}(n) , returns h(x).
m(n)

3. Pairwise independence: For every x = y ∈ {0, 1}(n) , if h is uniformly selected in

m(n)
S(n) , then h(x) and h(y) are independent and uniformly distributed in {0, 1}m(n) .
That is, for every α, β ∈ {0, 1}m(n) ,
Prh [h(x) = α ∧ h(y) = β] = 2−2m(n)

In fact, for the current application, we can replace the third condition by the following
weaker condition, parameterized by a function cp : N →[0, 1] (s.t. cp(n) ≥ 2−m(n) ): For
every x = y ∈ {0, 1}(n) ,
Prh [h(x) = h(y)] ≤ cp(n) (6.3)
Indeed, the pairwise independence condition implies that Eq. (6.3) is satisfied with
cp(n) = 2−m(n) . Note that Eq. (6.3) asserts that the collision probability of S(n)
m(n)
is
at most cp(n), where the collision probability refers to the probability that h(x) =
h(y) when h is uniformly selected in S(n) and x = y ∈ {0, 1}(n) are arbitrary fixed
m(n)

strings.
Hashing ensembles with n ≤ (n) + m(n) and cp(n) = 2−m(n) can be constructed
(for a variety of functions , m : N → N, e.g., (n) = 2n/3 and m(n) = n/3); see Ex-
ercise 22. Using such ensembles, we first present a construction of length-restricted
message-authentication schemes (and later show how to generalize the construction to
obtain full-fledged message-authentication schemes).

Construction 6.3.5 (Construction 6.3.4, revisited – length-restricted version): Let

{h r : {0, 1}(|r |) → {0, 1}m(|r |) }r ∈{0,1}∗ and { f s : {0, 1}m(|s|) → {0, 1}m(|s|) }s∈{0,1}∗ be effi-
ciently computable function ensembles. We construct the following -restricted scheme,
(G, S, V ):
Key-generation with G: On input 1n , algorithm G selects independently and uniformly
r, s ∈ {0, 1}n . The key-pair output by G is ((r, s), (r, s)).
Signing with S: On input a signing-key (r, s) in the range of G 1 (1n ) and a document
α ∈ {0, 1}(n) , algorithm S outputs the signature/tag f s (h r (α)).
Verification with V : On input a verifying-key (r, s) in the range of G 2 (1n ), a docu-
ment α ∈ {0, 1}(n) , and an alleged signature β, algorithm outputs 1 if and only if
f s (h r (α)) = β.

Note that a generalization of Construction 6.3.5 in which the pseudorandom function is

replaced by an arbitrary (length-restricted) secure message-authentication scheme may
be insecure; see Exercise 9.

Proposition 6.3.6: Suppose that { f s : {0, 1}m(|s|) → {0, 1}m(|s|) }s∈{0,1}∗ is a pseudoran-
dom function, and that the collision probability of the collection {h r : {0, 1}(|r |) →
528

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

{0, 1}m(|r |) }r ∈{0,1}∗ is a negligible function of |r |. Then Construction 6.3.5 constitutes a

secure -restricted message-authentication scheme.

In particular, the second hypothesis requires that 2−m(n) be a negligible function in n. By

the previous discussion, adequate collections of hashing functions (i.e., with collision
probability 2−m(n) ) exists for (n) = 2n/3 (and m(n) = n/3). We comment that, under
the hypothesis of Proposition 6.3.6, the collection {gs,r : f s ◦ h r }|s|=|r | constitutes a
pseudorandom function ensemble. This is implicitly shown in the following proof, and
is related to Exercise 31 in Chapter 3.

Proof Sketch: As in the proof of Proposition 6.3.2, we first consider the security of
an ideal scheme in which the pseudorandom function is replaced by a truly random
function (mapping m(n)-bit long strings to m(n)-bit long strings). Consider any (proba-
bilistic polynomial-time) adversary attacking the ideal scheme. Such an adversary may
obtain the signatures to polynomially -many (n)-bit long strings of its choice. How-
ever, except with negligible probability, these strings are hashed to different m(n)-bit
long strings, which in turn are mapped by the random function to totally independent
and uniformly distributed m(n)-bit long strings. Furthermore, except with negligible
probability, the (n)-bit long string α contained in the adversary’s (alleged message-
signature) output pair is hashed to an m(n)-bit long string that is different from all the
previous hash-values, and so the single valid signature corresponding to α is a uniformly
distributed m(n)-bit long string that is independent of all previously seen signatures.

On the distribution of signatures in the ideal scheme: Suppose that the hashing
collection {h r : {0, 1}(|r |) → {0, 1}m(|r |) }r ∈{0,1}n has collision probability cp(n), and
φ : {0, 1}m(n) → {0, 1}m(n) is a random function. Then, we claim that an adversary
that obtains signatures to t(n) − 1 strings of its choice succeeds in forging a signature
to a new string with probability at most t(n)2 · cp(n) + 2−m(n) , regardless of its
computational powers. The claim is proved by showing that, except with probability
at most t(n)2 · cp(n), the t(n) strings selected by the adversary are mapped by
h r to distinct values. The latter claim is proved by induction on the number of
selected strings, denoted i, where the base case (i.e., i = 1) holds vacuously. Let
s1 , ..., si denote the strings selected so far, and suppose that with probability at
least 1 − i 2 · cp(n), the i hash-values h r (s j )’s are distinct. The adversary only sees
the corresponding φ(h r (s j ))’s, which are uniformly and independently distributed
(in a way independent of the values of the h r (s j )’s). Thus, loosely speaking, the
adversary’s selection of the next string, denoted si+1 , is independent of the values
of the h r (s j )’s, and so a collision of h r (si+1 ) with one of the previous h r (s j )’s occurs
with probability at most i · cp(n). The induction step follows (since 1 − i 2 · cp(n) −
i · cp(n) > 1 − (i + 1)2 · cp(n)).

It follows that any adversary attacking the ideal scheme may succeed in existential
forgery with at most negligible probability (provided it makes at most polynomially
many queries). The same must hold for any efficient adversary that attacks the actual
scheme, since otherwise such an adversary yields a violation of the pseudorandomness
of { f s : {0, 1}m(|s|) → {0, 1}m(|s|) }s∈{0,1}∗ . The exact implementation of this argument
follows the details given in the proof of Proposition 6.3.2.
529
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Obtaining Full-Fledged MACs. Construction 6.3.5 can be generalized to obtain full-

fledged MACs by using generalized hashing families that map arbitrary strings (rather
than fixed-length ones) to fixed-length strings. Specifically, for  : N → N and cp :
N →[0,1], we call {h r : {0, 1}∗ → {0, 1}m(|r |) }n∈N a generalized hashing ensemble
with a (, cp)-collision property if it satisfies the following two conditions:

1. Efficient evaluation: There exists a polynomial-time algorithm that, on input r (rep-

resenting the function h r ) and a string x ∈ {0, 1}∗ , returns h r (x).
2. Collision probability:11 For every n ∈ N and x = y such that |x|, |y| ≤ (n), the
probability that h r (x) = h r (y) when r is uniformly selected in {0, 1}n is at most
cp(n).

For our construction of a full-fledged MAC, we need a generalized hashing ensemble

with an (, cp)-collision property for some super-polynomial (n) and negligible cp(n)
ε
(e.g., (n) = 1/cp(n) = 2n for some constant ε > 0). The existence of such ensembles
will be discussed after the proof of Proposition 6.3.7.

Proposition 6.3.7 (Construction 6.3.4, revisited – full-fledged version): Suppose that

{ f s : {0, 1}m(|s|) → {0, 1}m(|s|) }s∈{0,1}∗ is a pseudorandom function ensemble. For some
super-polynomial  : N → N and negligible cp : N →[0, 1], suppose that {h r :
{0, 1}∗ → {0, 1}m(|r |) }r ∈{0,1}∗ is a generalized hashing ensemble with an (, cp)-collision
property. Then (G, S, V ) as in Construction 6.3.4 constitutes a secure MAC. That is,
we refer to the following scheme:
Key-generation with G: On input 1n , algorithm G selects independently and uniformly
r, s ∈ {0, 1}n , and outputs ((r, s), (r, s)).
Signing with S: On input a signing-key (r, s) and a document α ∈ {0, 1}∗ , algorithm
S outputs the signature/tag f s (h r (α)).
Verification with V : On input a verifying-key (r, s), a document α ∈ {0, 1}∗ , and an
alleged signature β, algorithm outputs 1 if and only if f s (h r (α)) = β.

Proof Sketch: The proof is identical to the proof of Proposition 6.3.6, except that here
the (polynomial-time) adversary attacking the scheme may query for the signatures
of strings of various lengths. Still, all these queries (as well as the final output) are
of polynomial length and thus shorter than (n). Thus, the (, cp)-collision property
implies that, except with negligible probability, all these queries (as well as the relevant
part of the output) are hashed to different values.
ε
On Constructing Adequate Hashing Ensembles. For some ε > 0 and f (n) = 2n ,
generalized hashing ensembles with a ( f, 1/ f )-collision property can be constructed
is several ways. One such way is by applying a tree-hashing scheme as in Construc-
tion 6.2.13; see Exercise 23. For further details about constructions of generalized

11 Note that it is essential to restrict the collision condition to strings of bounded length. In contrast, for every finite
family of functions H , there exist two different strings that are mapped to the same image by each function in
H . For details, see Exercise 21.

530

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

hashing ensembles, see Section 6.6.5. Combining any of these constructions with
Proposition 6.3.7, we get:

Theorem 6.3.8: Assuming the existence of one-way functions, there exist message-
authentication schemes with fixed-length signatures; that is, signatures of length that
depend on the length of the signing-key but not on the length of the document.

An Alternative Presentation. The proofs of Propositions 6.3.6 and 6.3.7 actually

establish that {gs,r = f s ◦ h r }s∈{0,1}∗ ,r ∈{0,1}|s| is a generalized pseudorandom function
(as per Definition 3.6.12). For further discussion of this aspect, see Section C.2 in
Appendix C. Hence, the actual claim of these propositions (i.e., the security of the
constructed MAC) can be derived from the fact that any generalized pseudorandom
function yields a full-fledged MAC (see Exercise 8).

6.3.2.* More on Hash-and-Hide and State-Based MACs

The basic idea underlying Construction 6.3.5 (as well as Proposition 6.3.7) is to combine
a “weak tagging scheme” with an adequate “hiding scheme.” Specifically, the weak
tagging scheme should be secure against forgery provided that the adversary does not
have access to the scheme’s outcome, and the hiding scheme implements the latter
provision in a setting in which the actual adversary does obtain the value of the MAC.
In Construction 6.3.5 (and in Proposition 6.3.7), the tagging scheme was implemented
by ordinary hashing and hiding was obtained by applying a pseudorandom function to
the string that one wishes to hide.12
One more natural “hiding scheme” (which can also be implemented using pseu-
dorandom functions) is obtained by using certain private-key encryption schemes. For
example, we may use Construction 5.3.9 (in which the plaintext x is encrypted/hidden by
the pair (u, x ⊕ f s (u)), where u is uniformly selected), instead of hiding x by the value
f s (x) (as in Construction 6.3.5 and Proposition 6.3.7). The resulting MAC is as follows:
Key-generation: On input 1n , we select independently and uniformly r, s ∈ {0, 1}n ,
where r specifies a hashing13 function h r : {0, 1}∗ → {0, 1}m(|r |) and s specifies
a pseudorandom function f s : {0, 1}m(|s|) → {0, 1}m(|s|) . We output the key-pair
((r, s), (r, s)).
Signing: On input a signing-key (r, s) and a document α ∈ {0, 1}∗ , we uniformly select
u ∈ {0, 1}m(|s|) , and output the signature/tag (u, h r (α) ⊕ f s (u)).
Verification: On input a verifying-key (r, s), a document α ∈ {0, 1}∗ , and an alleged
signature (u, v), we output 1 if and only if v = hr (α) ⊕ f s (u).
Alternative implementations of the same underlying idea are more popular, especially
in the context of state-based MACs. We start by defining state-based MACs, and

12 We comment that this specific hiding method is not 1-1, and furthermore, it is not clear whether it can also be
efficiently inverted when given the “secret key” (i.e., the seed of the pseudorandom function). In contrast, the
alternative hiding method described next is 1-1 and can be efficiently inverted when given the secret key.
13 The hashing function should belong to an AXU family, as defined in Section 6.3.2.2.

531
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

then show how to construct them based on the hash-and-hide (or rather tag-and-hide)
paradigm.

6.3.2.1. The Definition of State-Based MACs

As in the case of steam-ciphers discussed in Section 5.3.1, we extend the mechanism
of message-authentication schemes (MACs) by allowing the signing and verification
processes to maintain and update a state. Formally, both the signing and the verification
algorithms take an additional input and emit an additional output, corresponding to their
state before and after the operation. The length of the state is not allowed to grow by too
much during each application of the algorithm (see Condition 3 in Definition 6.3.9), or
else efficiency of the entire “repeated signing” process cannot be guaranteed. For the
sake of simplicity, we incorporate the key in the state of the corresponding algorithm.
Thus, the initial state of each of the algorithms is set to equal its corresponding key.
Furthermore, one may think of the intermediate states as of updated values of the
corresponding key.
In the following definition, we follow conventions similar to those used in defin-
ing state-based ciphers (i.e., Definition 5.3.1). Specifically, for simplicity, we assume
that the verification algorithm (i.e., V ) is deterministic (otherwise the formulation
would be more complex). Intuitively, the main part of the verification condition (i.e.,
Condition 2) is that the (proper) iterative signing-verifying process always accepts.
The additional requirement in Condition 2 is that the state of the verification algorithm
be updated correctly as long as it is fed with strings of length equal to the length of
the valid document-signature pairs. The importance of this condition was discussed in
Section 5.3.1 and is further discussed following Definition 6.3.9.

Definition 6.3.9 (state-based MAC – the mechanism): A state-based message-

authentication scheme is a triple, (G, S, V ), of probabilistic polynomial-time algo-
rithms satisfying the following three conditions:

1. On input 1n , algorithm G outputs a pair of bit strings.

2. For every pair (s (0) , v (0) ) in the range of G(1n ), and every sequence of α (i) ’s, the
following holds: If (s (i) , β (i) ) ← S(s (i−1) , α (i) ) and (v (i) , γ (i) ) ← V (v (i−1) , α (i) , β (i) )
for i = 1, 2, ..., then γ (i) = 1 for every i.
Furthermore, for every i and every (α, β) ∈ {0, 1}|α | × {0, 1}|β | , it holds that
(i) (i)

V (v (i−1) , α, β) = (v (i) , ·). That is, v (i) is actually determined by v (i−1) and
(|α (i) |, |β (i) |).14
3. There exists a polynomial p such that for every pair (s (0) , v (0) ) in the range of
G(1n ), and every sequence of α (i) ’s and s (i) ’s as in Condition 2, it holds that |s (i) | ≤
|s (i−1) | + |α (i) | · p(n). Similarly for the v (i) ’s.

14 Alternatively, we may decompose the verification (resp., signing) algorithm into two algorithms, where the first
takes care of the actual verification (resp., signing) and the second takes care of updating the state. For details,
see Exercise 18.

532

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

That is, as in Definition 6.1.1, the signing-verification process operates properly pro-
vided that the corresponding algorithms get the corresponding keys (states). Note that
in Definition 6.3.9, the keys are modified by the signing-verification process, and so
correct verification requires holding the correctly updated verification-key. We stress
that the furthermore-clause in Condition 2 guarantees that the verification-key is cor-
rectly updated as long as the verification process is fed with strings of the correct lengths
(but not necessarily with the correct document-signature pairs). This extra requirement
implies that, given the initial verification-key and the current document-signature pair,
as well as the lengths of all previous pairs (which may be actually incorporated in
the current signature), one may correctly decide whether or not the current document-
signature pair is valid. As in the case of state-based ciphers (cf. Section 5.3.1), this fact
is interesting for two reasons:
A theoretical reason: It implies that without loss of generality (alas, with possible loss
in efficiency), the verification algorithm may be stateless. Furthermore, without loss
of generality (alas, with possible loss in efficiency), the state of the signing algorithm
may consist of the initial signing-key and the lengths of the messages signed so far.
(We assume here that the length of the signature is determined by the length of the
message and the length of the signing-key.)
A practical reason: It allows for recovery from the loss of some of the message-
signature pairs. That is, assuming that all messages have the same length (which
is typically the case in MAC applications), if the receiver knows (or is given) the
total number of messages sent so far, then it can verify the authenticity of the current
message-signature pair, even if some of the previous message-signature pairs were
lost.

We stress that Definition 6.3.9 refers to the signing of multiple messages (and
is meaningless when considering the signing of a single message). However, Defi-
nition 6.3.9 (by itself) does not explain why one should sign the i-th message us-
ing the updated signing-key s (i−1) , rather than by reusing the initial signing-key s (0)
(where all corresponding verifications are done by reusing the initial verification-key
v (0) ). Indeed, the reason for updating these keys is provided by the following secu-
rity definition that refers to the signing of multiple messages, and holds only in case
the signing-keys in use are properly updated (in the multiple-message authentication
process).

Definition 6.3.10 (security of state-based MACs):

r A chosen message attack on a state-based MAC, (G, S, V ), is an interactive
process that is initiated with (s (0) , v (0) ) ← G(1n ), and proceeds as follows: In the
i-th iteration, based on the information gathered so far, the attacker selects a string
α (i) , and obtains β (i) , where (s (i) , β (i) ) ← S(s (i−1) , α (i) ).
r Such an attack is said to succeed if it outputs a valid signature to a string for which
it has not requested a signature during the attack. That is, the attack is successful
if it outputs a pair (α, β) such that α is different from all signature-queries made
533
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

during the attack, and V (v (i−1) , α, β) = (·, 1) holds for some intermediate state
(verification-key) v (i−1) (as in Definition 6.3.9).15
r A state-based MAC is secure if every probabilistic polynomial-time chosen message
attack as in the first item succeeds with at most negligible probability.

Note that Definition 6.3.10 (only) differs from Definition 6.1.2 in the way that the
signatures β (i) ’s are produced (i.e., using the updated signing-key s (i−1) , rather than the
initial signing-key s (0) ). Furthermore, Definition 6.3.10 guarantees nothing regarding
a signing process in which the signature to the i-th message is obtained by invoking
S(s (0) , ·) (as in Definition 6.1.2).

6.3.2.2. State-Based Hash-and-Hide MACs

We are now ready to present alternative implementations of the hash-and-hide paradigm.
Recall that in Section 6.3.1.3, the document was hashed (by using an adequate hashing
function), and the resulting hash-value was (authenticated and) hidden by applying a
pseudorandom function to it. In the current subsection, hiding will be obtained in a
more natural (and typically more efficient) way, that is, by XORing the hash-value with
a new portion of a (pseudorandom) one-time pad. Indeed, the state is used in order to
keep track of what part of the (one-time) pad was already used (and should not be used
again). Furthermore, to obtain improved efficiency, we let the state encode information
that allows fast generation of the next portion of the (pseudorandom) one-time pad. This
is obtained using an (on-line) pseudorandom generator (see Sections 3.3.3 and 5.3.1).
Recall that on-line pseudorandom generators are a special case of variable-output
pseudorandom generators (see Section 3.3.3), in which a hidden state is maintained
and updated so as to allow generation of the next output bit in time polynomial in the
length of the initial seed, regardless of the number of bits generated so far. Specifically,
the next (hidden) state and output bit are produced by applying a (polynomial-time
computable) function g : {0, 1}n → {0, 1}n+1 to the current state (i.e., (s  , σ ) ← g(s),
where s is the current state, s  is the next state and σ is the next output bit). Analogously
to Construction 5.3.3, the suggested state-based MAC will use an on-line pseudorandom
generator in order to generate the required pseudorandom one-time pad, and the latter
will be used to hide (and authenticate) the hash-value (obtained by hashing the original
document).

Construction 6.3.11 (a state-based MAC): Let g : {0, 1}∗ → {0, 1}∗ such that |g(s)| =
|s| + 1, for every s ∈ {0, 1}∗ . Let {h r : {0, 1}∗ → {0, 1}m(|r |) }r ∈{0,1}∗ be a family of func-
tions having an efficient evaluation algorithm.

15 In fact, one may strengthen the definition by using a weaker notion of success in which it is only required that
α = α (i) (rather than requiring that α ∈ {α ( j) } j ). That is, the attack is successful if, for some i, it outputs a
pair (α, β) such that α = α (i) and V (v (i−1) , α, β) = (·, 1), where the α ( j) ’s and v ( j) ’s are as in Definition 6.3.9.
The stronger definition provides “replay protection” (i.e., even if the adversary obtains a valid signature that
authenticates α as the j-th message, it cannot produce a valid signature that authenticates α as the i-th message,
unless α was actually authenticated as the i-th message).

534

www.Ebook777.com
 6.3 CONSTRUCTIONS OF MESSAGE-AUTHENTICATION SCHEMES

Key-generation and initial state: Uniformly select s, r ∈ {0, 1}n , and output the key-
pair ((s, r ), (s, r )). The initial state of each algorithm is set to (s, r, 0, s).
(We maintain the initial key (s, r ) and a step-counter in order to allow recovery from
loss of message-signature pairs.)
def
Signing message α with state (s, r, t, s  ): Let s0 = s  . For i = 1, ..., m(n), compute
si σi = g(si−1 ), where |si | = n and σi ∈ {0, 1}. Output the signature h r (α) ⊕
σ1 · · · σm(n) , and set the new state to (s, r, t + m(n), sm(n) ).
Verification of the pair (α, β) with respect to the state (s, r, t, s  ): Compute σ1· · · σm(n)
and sm(n) as in the signing process; that is, for i = 1, ..., m(n), compute si σi =
def
g(si−1 ), where s0 = s  . Set the new state to (s, r, t + m(n), sm(n) ), and accept if and
only if β = hr (α) ⊕ σ1 · · · σm(n) .

Special recovery procedure: When notified that some message-signature pairs may
have been lost and that the current message-signature pair has index t  , one first
recovers the correct current state, which as in the ordinary verification (of the pre-
def
vious paragraph) will be denoted s0 . This is done by setting s−t  = s and computing

si−t  σi−t  = g(si−t  −1 ), for i = 1, ..., t . Indeed, recovery of s0 is required only if
t  = t.16
Note that both the signing and verification algorithms are deterministic, and that the
state after authentication of t messages has length 3n + log2 (t · m(n)) < 4n, provided
that t < 2n /m(n).
We now turn to the analysis of the security of Construction 6.3.11. The hashing
property of the collection of h r ’s should be slightly stronger than the one used in
Section 6.3.1.3. Specifically, rather than a bound on the collision probability (i.e., the
probability that h r (x) = h r (y) for any relevant fixed x, y and a random r ), we need
a bound on the probability that h r (x) ⊕ h r (y) equals any fixed string (again, for any
relevant fixed x, y and a random r ). This property is commonly referred to by the name
Almost-Xor-Universal (AXU). That is, {h r : {0, 1}∗ → {0, 1}m(|r |) }r ∈{0,1}∗ is called an
(, ε)-AXU family if for every n ∈ N, every x = y such that |x|, |y| ≤ (n), and every
z, it holds that

Pr[h Un (x) ⊕ h Un (y) = z] ≤ ε(n) (6.4)

References to constructions of such families are provided in Section 6.6.5.

Proposition 6.3.12: Suppose that g is a pseudorandom generator,17 and that {h r } is

a (, ε)-AXU family, for some super-polynomial  and negligible ε. Then Construc-
tion 6.3.11 constitutes a secure state-based MAC. Furthermore, security holds even
with respect to the stronger notion discussed in footnote 15.

16 More generally, if the verification procedure holds the state at time t < t  , then it need only compute
st+1−t  , ..., s0 .
17 In fact, as shown in the proof, it suffices to assume that g is a next-step function of an on-line pseudorandom

generator.

535
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Proof Sketch: By Exercise 21 of Chapter 3, if g is a pseudorandom generator, then for

p p
every polynomial p the ensemble {G n }n∈N is pseudorandom, where G n is defined by
the following random process:
Uniformly select s0 ∈ {0, 1}n ;
For i = 1 to p(n), let si σi ← g(si−1 ), where σi ∈ {0, 1} (and si ∈ {0, 1}n );
Output σ1 σ2 · · · σ p(n) .
Recall that, in such a case, we said that g is a next-step function of an on-line pseu-
dorandom generator.
As in previous cases, it suffices to establish the security of an ideal scheme in which
the sequence (of m(n)-bit long blocks) produced by iterating the next-step function g
is replaced by a truly random sequence (of m(n)-bit long blocks). In the ideal scheme,
all that the adversary may obtain via a chosen message attack is a sequence of m(n)-bit
long blocks, which is uniformly distributed among all such possible sequences. Note
that each of the signatures obtained during the attack, as well as the forged signature,
refers to a single block in this sequence (e.g., the i-th obtained signature refers to the
i-th block). We consider two types of forgery attempts:
1. In case the adversary tries to forge a signature referring to an unused (during the
attack) block, it may succeed with probability at most 2−m(n) , because we may think
of this block as being chosen after the adversary makes its forgery attempt. Note
that 2−m(n) is negligible, because ε(n) ≥ 2−m(n) must hold (i.e., 2−m(n) lower-bounds
the collision probability).
2. The more interesting case is when the adversary tries to forge a signature referring to
a block, say the i-th one, that was used (to answer the i-th query) during the attack.
Denote the j-th query by α ( j) , the (random) j-th block by b( j) , and the forged
document by α. Then, at the time of outputting the forgery attempt (α, β), the
adversary only knows the sequence of b( j) ⊕ h r (α ( j) )’s (as well as the α ( j) ’s that
were chosen by it), but this yields no information on r (because the b( j) ’s are random
and unknown to the adversary). Note that the adversary succeeds if and only if
def
b(i) ⊕ h r (α) = β, where β (i) = b(i) ⊕ h r (α (i) ) is known to it. Thus, the adversary
succeeds if and only if h r (α (i) ) ⊕ h r (α) = β (i) ⊕ β, where α (i) , β (i) , α, β are known
to the adversary and r is uniformly distributed.

Further clarification: Considering the distribution of r conditioned on partial

transcripts of the attack (i.e., the sequence of queries and answers), we claim
that at any time, r is uniformly distributed in {0, 1}n . The claim holds because,
for each possible value of r , the answers to the different queries are all uniformly
distributed (because they are XORed with random b( j) ’s). Thus, r is uniformly
distributed also conditioned on the transcript at the time that the adversary outputs
its forgery attack, which in turn is successful if and only if b(i) ⊕ h r (α) = β holds,
where b (i) = h r (α (i) ) ⊕ β (i) and α (i) , β (i) , α, β are fixed by this transcript. Thus,
a successful forgery implies h r (α (i) ) ⊕ h r (α) = β (i) ⊕ β, for fixed α (i) , β (i) , α, β
and uniformly distributed r .

Hence, by the AXU property, the probability that the adversary succeeds is at most
ε(n).
536

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

The security of the real scheme follows (or else one could have distinguished
the sequence produced by iterating the next-step function g from a truly random
sequence).

Construction 6.3.11 Versus the Constructions of Section 6.3.1.3: Recall that all
these schemes are based on the hash-and-hide paradigm. The difference between the
schemes is that in Section 6.3.1.3, a pseudorandom function is applied to the hash-value
(i.e., the signature to α is f s (h r (α))), whereas in Construction 6.3.11, the hash-value
is XORed with a pseudorandom value (i.e., we may view the signature as consisting
of (c, h r (α) ⊕ f s (c)), where c is a counter value and f s (c) is the c-th block produced
by iterating the next-step function g starting with the initial seed s). We note two ad-
vantages of the state-based MAC over the MACs presented in Section 6.3.1.3: First,
applying an on-line pseudorandom generator is likely to be more efficient than ap-
plying a pseudorandom function. Second, a counter allows for securely authenticating
more messages than can be securely authenticated by applying a pseudorandom func-
tion to the hashed value. Specifically, the use of an an m-bit long counter allows for
securely authenticating 2m messages, whereas using an m-bit long √ hash-value suffers
from the “birthday effect” (i.e., collisions are likely to occur when 2m messages are
authenticated). Indeed, these advantages are relevant only in applications in which us-
ing state-based MACs is possible, and are most advantageous in applications where
verification is performed in the same order as signing (e.g., in fifo communication).
In the latter case, Construction 6.3.11 offers another advantage: “replay protection” (as
discussed in footnote 15).

6.4. Constructions of Signature Schemes

In this section, we present several constructions of secure public-key signature schemes.

In the sequel, we refer to such schemes as signature schemes, which is indeed the
traditional term.
Two central paradigms in the construction of signature schemes are the “refreshing”
of the “effective” signing-key (see Section 6.4.2.1), and the usage of an “authentication-
tree” (see Section 6.4.2.2). In addition, the “hash-and-sign paradigm” (employed also
in the construction of message-authentication schemes) plays an even more crucial role
in the following presentation. In addition, we use the notion of a one-time signature
scheme (see Section 6.4.1).
The current section is organized as follows. In Section 6.4.1 we define and construct
various types of one-time signature schemes. The hash-and-sign paradigm plays a
crucial role in one of these constructions, which in turn is essential for Section 6.4.2.
In Section 6.4.2 we show how to use one-time signature schemes to construct general
signature schemes. This construction utilizes the “refreshing paradigm” (as applied to
one-time signature schemes) and an authentication-tree. Thus, assuming the existence
of collision-free hashing, we obtain (general) signature schemes.
In Section 6.4.3, wishing to relax the conditions under which signature schemes
can be constructed, we define universal one-way hashing functions, and show how to
537
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

use them instead of collision-free hashing (in the aforementioned constructions and,
in particular, within a modified hash-and-sign paradigm). Indeed, the gain in using
universal one-way hashing (rather than collision-free hashing) is that the former can be
constructed based on any one-way function (whereas this is not known for collision-free
hashing). Thus, we obtain:

Theorem 6.4.1: Secure signature schemes exist if and only if one-way functions exist.

The difficult direction is to show that the existence of one-way functions implies the
existence of signature schemes. For the opposite direction, see Exercise 7.

6.4.1. One-Time Signature Schemes

In this section we define and construct various types of one-time signature schemes.
Specifically, we first define one-time signature schemes, next define a length-restricted
version of this notion (analogous to Definition 6.2.1), then present a simple construction
of the latter, and finally show how such a construction, combined with collision-free
hashing, yields a general one-time signature scheme.

6.4.1.1. Definitions
Loosely speaking, one-time signature schemes are signature schemes for which the
security requirement is restricted to attacks in which the adversary asks for at most
one string to be signed. That is, the mechanics of one-time signature schemes is as
of ordinary signature schemes (see Definition 6.1.1), but the security requirement is
relaxed as follows:
r A chosen one-message attack is a process that can obtain a signature to at most
one string of its choice. That is, the attacker is given v as input, and obtains a signature
relative to s, where (s, v) ← G(1n ) for an adequate n.
(Note that in this section, we focus on public-key signature schemes and thus present
only the definition for this case.)
r Such an attack is said to succeed (in existential forgery) if it outputs a valid signature
to a string for which it has not requested a signature during the attack.
(Indeed, the notion of success is exactly as in Definition 6.1.2.)
r A one-time signature scheme is secure (or unforgeable) if every feasible chosen
one-message attack succeeds with at most negligible probability.

Moving to the formal definition, we again model a chosen message attack as a proba-
bilistic oracle machine; however, since here we care only about one-message attacks, we
consider only oracle machines that make at most one query. Let M be such a machine.
As before, we denote by Q O M (x) the set of queries made by M on input x and access
to oracle O, and let M O (x) denote the output of the corresponding computation. Note
that here |Q OM (x)| ≤ 1 (i.e., M may make either no queries or a single query).

538

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

Definition 6.4.2 (security for one-time signature schemes): A one-time signature

scheme is secure if for every probabilistic polynomial-time oracle machine M that
makes at most one query, every positive polynomial p, and all sufficiently large n, it
holds that


Vv (α, β) = 1 & α ∈ Q SMs (1n ) 1
Pr <
where (s, v) ← G(1 ) and (α, β) ← M (v)
n Ss
p(n)
where the probability is taken over the coin tosses of algorithms G, S, and V, as well
as over the coin tosses of machine M.

We now define a length-restricted version of one-time signature schemes. The definition

is indeed analogous to Definition 6.2.1:

Definition 6.4.3 (length-restricted one-time signature schemes): Let  : N → N. An

-restricted one-time signature scheme is a triple, (G, S, V ), of probabilistic
polynomial-time algorithms satisfying the the mechanics of Definition 6.2.1. That is, it
satisfies the following two conditions:

1. As in Definition 6.1.1, on input 1n , algorithm G outputs a pair of bit strings.

2. Analogously to Definition 6.1.1, for every n and every pair (s, v) in the range of G(1n ),
and for every α ∈ {0, 1}(n) , algorithms S and D satisfy Pr[Vv (α, Ss (α)) = 1] = 1.

Such a scheme is called secure (in the one-time model) if the requirement of Defini-
tion 6.4.2 holds when restricted to attackers that only make queries of length (n) and
output a pair (α, β) with |α| = (n). That is, we consider only attackers that make at
most one query, with the requirements that this query be of length (n) and that the
output (α, β) satisfies |α| = (n).

Note that even the existence of secure 1-restricted one-time signature schemes implies
the existence of one-way functions, see Exercise 13.

6.4.1.2. Constructing Length-Restricted One-Time Signature Schemes

We now present a simple construction of length-restricted one-time signature schemes.
The construction works for any length-restriction function , but the keys will have
length greater than . The latter fact limits the applicability of such schemes and will be
removed in the next subsection. But first, we construct -restricted one-time signature
schemes that are based on any one-way function f . Loosely speaking, the verification-
key will consist of  pairs of images (of f ), and a signature will consist of  pre-images
(under f ) corresponding to  out of these 2 images, where the selection of images is
determined by the corresponding bits of the message. We may assume for simplicity
that f is length-preserving.

Construction 6.4.4 (an -restricted one-time signature scheme): Let  : N → N be

polynomially bounded and polynomial-time computable, and f : {0, 1}∗ → {0, 1}∗ be
539
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

polynomial-time computable and length-preserving. We construct an -restricted one-

time signature scheme, (G, S, V ), as follows:

Key-generation with G: On input 1n , we uniformly select s10 , s11 , ..., s(n)

0 1
, s(n) ∈
j j
{0, 1}n , and compute vi = f (si ), for i = 1, ..., (n) and j = 0, 1. We let s =
((s10 , s11 ), ..., (s(n)
0 1
, s(n) )), and v = ((v10 , v11 ), ..., (v(n)
0
, v(n)
1
)), and output the key-pair
(s, v).
(Note that |s| = |v| = 2 · (n) · n.)
Signing with S: On input a signing-key s = ((s10 , s11 ), ..., (s(n)0 1
, s(n) )) and an (n)-bit
σ1 σ(n)
string α = σ1 · · · σ(n) , we output (s1 , ..., s(n) ) as a signature of α.
Verification with V : On input a verification-key v = ((v10 , v11 ), ..., (v(n)
0
, v(n)
1
)), an
(n)-bit string α = σ1 · · · σ(n) , and an alleged signature β = (β1 , ..., β(n) ), we ac-
cept if and only if viσi = f (βi ), for i = 1, ..., (n).

Proposition 6.4.5: If f is a one-way function, then Construction 6.4.4 constitutes a

secure -restricted one-time signature scheme.

Note that Construction 6.4.4 does not constitute a (general) -restricted signature
scheme: An attacker that obtains signatures to two strings (e.g., to the strings 0(n)
and 1(n) ), can present a valid signature to any (n)-bit long string (and thus totally
break the system). However, here we consider only attackers that may ask for at most
one string (of their choice) to be signed. As a corollary to Proposition 6.4.5, we obtain:

Corollary 6.4.6: If there exist one-way functions, then for every polynomially bounded
and polynomial-time computable  : N → N, there exist secure -restricted one-time
signature schemes.

Proof of Proposition 6.4.5: Intuitively, forging a signature (after seeing at most one sig-
nature to a different message) requires inverting f on some random image (correspond-
ing to a bit location on which the two (n)-bit long messages differ). The actual proof
is by a reducibility argument. Given an adversary A attacking the scheme (G, S, V ),
while making at most one query, we construct an algorithm A  for inverting f .
As a warm-up, let us first deal with the case in which A makes no queries at all. In
this case, on input y (supposedly in the range of f ), algorithm A proceeds as follows.
First A selects uniformly and independently a position p in {1, ..., (n)}, a bit b, and
a sequence of (2(n) many) n-bit long strings s10 , s11 , ..., s(n)0 1
, s(n) . (Actually, s bp is not
used and needs not be selected.) For every i ∈ {1, ..., (n)} \ { p}, and every j ∈ {0, 1},
algorithm A computes vi = f (si ). Algorithm A also computes v 1−b
j j
p = f (s 1−b
p ), and
sets v p = y and v = ((v1 , v1 ), ..., (v(n) , v(n) )). Note that if y = f (x), for a uniformly
b 0 1 0 1

distributed x ∈ {0, 1}n , then for each possible choice of p and b, the sequence v is
distributed identically to the public-key generated by G(1n ). Next, A invokes A on
input v, hoping that A will forge a signature, denoted β = τ1 · · · τ(n) , to a message
α = σ1 · · · σ(n) so that σ p = b. If this event occurs, A obtains a pre-image of y under
540

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

σ
f , because the validity of the signature implies that f (τ p ) = v p p = v bp = y. Observe
that conditioned on the value of v and the internal coin tosses of A, the value b is
uniformly distributed in {0, 1}. Thus, A inverts f with probability ε(n)/2, where ε(n)
denotes the probability that A succeeds in forgery.
We turn back to the actual case in which A may make a single query. Without loss
of generality, we assume that A always makes a single query; see Exercise 11. In this
case, on input y (supposedly in the range of f ), algorithm A selects p, b and the si ’s,
j
j
and forms the vi ’s and v exactly as in the previous warm-up discussion. Recall that if
y = f (x), for a uniformly distributed x ∈ {0, 1}n , then for each possible choice of p
and b, the sequence v is distributed identically to the public-key generated by G(1n ).
Also note that for each vi other than v bp = y, algorithm A holds a random pre-image
j

(of vi ) under f . Next, A invokes A on input v, and tries to answer its query, denoted
j

α = σ1 · · · σ(n) . We consider two cases regarding this query:

1. If σ p = b, then A cannot supply the desired signature because it lacks a pre-image

of s bp = y under f . Thus, in this case A aborts. However, this case occurs with
probability 12 , independently of the actions of A (because v yields no information
on either p or b).
(That is, conditioned on the value of v and the internal coin tosses of A, this case
occurs with probability 12 .)18
2. If σ p = 1 − b, then A can supply the desired signature because it holds all the
relevant si ’s (i.e., random pre-images of the relevant vi ’s under f ). In particular, A
j j
j  σ1 σ(n)
holds both si ’s, for i = p, as well as s 1−b
p . Thus, A answers with (s1 , ..., s(n) ).

Note that conditioned on the value of v, on the internal coin tosses of A, and on the
second case occuring, p is uniformly distributed in {1, ..., (n)}. When the second case
occurs, A obtains a signature to α, and this signature is distributed exactly as in a
real attack. We stress that since A asks at most one query, no additional query will be
asked by A. Also note that, in this case (i.e., σ p = 1 − b), algorithm A outputs a forged
message-signature pair, denoted (α  , β  ), with probability exactly as in a real attack.
We now turn to the analysis of A , and consider first the emulated attack of A. Recall
that α = σ1 · · · σ(n) denotes the (single) query19 made by A, and let α  = σ1 · · · σ(n) 
    
and β = s1 · · · s(n) , where (α , β ) is the forged message-signature pair output by A.
By our hypothesis (that this is a forgery-success event), it follows that α  = α and that
σ
f (si ) = vi i for all i’s. Now, considering the emulation of A by A , recall that (under
all these conditions) p is uniformly distributed in {1, ..., (n)}. Hence, with probability
|{i:σi =σi }|
(n)
≥ (n)
1
, it holds that σ p = σ p , and in that case, A obtains a pre-image of y
σ 1−σ p
under f (since s p satisfies f (s p ) = v p p , which in turn equals v p = v bp = y).

18 This follows from an even stronger statement by which conditioned on the value of v, on the internal coin tosses
of A, and on the value of p, the current case happens with probability 12 . The stronger statement holds because
under all these conditions, b is uniformly distributed in {0, 1} (and so σ p = b happens with probability exactly
1
2 ).
19 Recall that, without loss of generality, we may assume that A always makes a single query; see Exercise 11.

541
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

To summarize, assuming that A succeeds in a single-message attack on (G, S, V )

with probability ε(n), algorithm A inverts f on a random image (i.e., on f (Un )) with
probability
1 |{i : σi = σi }| ε(n)
ε(n) · · ≥
2 (n) 2(n)
Thus, if A is a probabilistic polynomial-time chosen one-message attack that forges
signatures with non-negligible probability, then A is a probabilistic polynomial-time
algorithm that inverts f with non-negligible probability (in violation of the hypothesis
that f is a one-way function). The proposition follows.

6.4.1.3. From Length-Restricted Schemes to General Ones

Using the hash-and-sign paradigm (i.e., Construction 6.2.6), we transform length-
restricted one-time signature schemes into one-time signature schemes. That is, we
use collision-free hashing and apply Construction 6.2.6, except that here (G, S, V ) is
an -restricted one-time signature scheme, rather than an -restricted (general) signature
scheme. Analogously to Proposition 6.2.7, we obtain:

Proposition 6.4.7: Suppose that (G, S, V ) is a secure -restricted one-time signature

scheme, and that {hr : {0, 1}∗ → {0, 1}(|r |) }r ∈{0,1}∗ is a collision-free hashing collec-
tion. Then (G  , S  , V  ), as defined in Construction 6.2.6, is a secure one-time signature
scheme.

Proof: The proof is identical to the proof of Proposition 6.2.7; we merely notice that if
the adversary A , attacking (G  , S  , V  ), makes at most one query, then the same holds
for the adversary A that we construct (in that proof) to attack (G, S, V ). In general,
the adversary A constructed in the proof of Proposition 6.2.7 makes a single query per
each query of the adversary A .

Combining Proposition 6.4.7, Corollary 6.4.6, and the fact that collision-free hashing
collections, imply one-way functions (see Exercise 14), we obtain:

Corollary 6.4.8: If there exist collision-free hashing collections, then there exist secure
one-time signature schemes. Furthermore, the length of the resulting signatures depends
only on the length of the signing-key.

Comments. We stress that when using Construction 6.2.6, signing each document
under the (general) scheme (G  , S  , V  ) only requires signing a single string under the
-restricted scheme (G, S, V ). This is in contrast to Construction 6.2.3, in which signing
a document under the (general) scheme (G  , S  , V  ) requires signing many strings under
the -restricted scheme (G, S, V ), where the number of such strings depends (linearly)
on the length of the original document.
Construction 6.2.6 calls for the use of collision-free hashing. The latter can be con-
structed using any claw-free permutation collection (see Proposition 6.2.9); however,
542

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

it is not know whether collision-free hashing can be constructed based on any one-way
function. Wishing to construct signature schemes based on any one-way function, we
later avoid (in Section 6.4.3) the use of collision-free hashing. Instead, we use “universal
one-way hashing functions” (to be defined), and present a variant of Construction 6.2.6
that uses these functions, rather than collision-free ones.

6.4.2. From One-Time Signature Schemes to General Ones

In this section we show how to construct general signature schemes using one-time
signature schemes. That is, we shall prove:

Theorem 6.4.9: If there exist secure one-time signature schemes, then secure (general)
signature schemes exist as well.

Actually, we can use length-restricted one-time signature schemes, provided that the
length of the strings being signed is at least twice the length of the verification-key.
Unfortunately, Construction 6.4.4 does not satisfy this condition. Nevertheless, Corol-
lary 6.4.8 does provide one-time signature schemes. Thus, combining Theorem 6.4.9
and Corollary 6.4.8, we obtain:

Corollary 6.4.10: If there exist collision-free hashing collections, then there exist se-
cure signature schemes.

Note that Corollary 6.4.10 asserts the existence of secure (public-key) signature
schemes, based on an assumption that does not mention trapdoors. We stress this point
because of the contrast to the situation with respect to public-key encryption schemes,
where a trapdoor property seems necessary for the construction of secure schemes.

6.4.2.1. The Refreshing Paradigm

The so-called “refreshing paradigm” plays a central role in the proof of Theorem 6.4.9.
Loosely speaking, the refreshing paradigm suggests reducing the dangers of a chosen
message attack on the signature scheme by using “fresh” instances of the scheme for
signing each new document. Of course, these fresh instances should be authenticated
by the original instance (corresponding to the verification-key that is publicly known),
but such an authentication refers to a string selected by the legitimate signer, rather
than by the adversary.

Example. To demonstrate the refreshing paradigm, consider a basic signature scheme

(G, S, V ) used as follows. Suppose that the user U has generated a key-pair, (s, v) ←
G(1n ), and has placed the verification-key v on a public-file. When a party asks U to
sign some document α, the user U generates a new (fresh) key-pair, (s  , v  ) ← G(1n ),
signs v  using the original signing-key s, signs α using the new (fresh) signing-key s  ,
and presents (Ss (v  ), v  , Ss  (α)) as a signature to α. An alleged signature, (β1 , v  , β2 ), is
verified by checking whether both Vv (v  , β1 ) = 1 and Vv (α, β2 ) = 1 hold. Intuitively,
543
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

the gain in terms of security is that a full-fledged chosen message attack cannot be
launched on (G, S, V ). All that an attacker may obtain (via a chosen message attack
on the new scheme) is signatures, relative to the original signing-key s, to randomly
chosen strings (taken from the distribution G 2 (1n )), as well as additional signatures
each relative to a random and independently chosen signing-key.
We refrain from analyzing the features of the signature scheme presented in this
example. Instead, as a warm-up to the actual construction used in the next section (in
order to establish Theorem 6.4.9), we present and analyze a similar construction (which
is, in some sense, a hybrid of the two constructions). The reader may skip this warm-up,
and proceed directly to Section 6.4.2.2.

Construction 6.4.11 (a warm-up): Let (G, S, V ) be a signature scheme and

(G  , S  , V  ) be a one-time signature scheme. Consider a signature scheme, (G  , S  , V  ),
with G  = G, as follows:

Signing with S  : On input a signing-key s (in the range of G 1 (1n )) and a document
α ∈ {0, 1}∗ , first invoke G  to obtain (s  , v  ) ← G  (1n ). Next, invoke S to obtain
β1 ← Ss (v  ), and S  to obtain β2 ← Ss  (α). The final output is (β1 , v  , β2 ).
Verification with V  : On input a verifying-key v, a document α ∈ {0, 1}∗ , and an al-
leged signature β = (β1 , v  , β2 ), we output 1 if and only if both Vv (v  , β1 ) = 1 and
Vv (α, β2 ) = 1.

Construction 6.4.11 differs from the previous example only in that a one-time signature
scheme is used to generate the “second signature” (rather than using the same ordinary
signature scheme). The use of a one-time signature scheme is natural here, because it
is unlikely that the same signing-key s  will be selected in two invocations of S  .

Proposition 6.4.12: Suppose that (G, S, V ) is a secure signature scheme, and that
(G  , S  , V  ) is a secure one-time signature scheme. Then (G  , S  , V  ), as defined in
Construction 6.4.11, is a secure signature scheme.

We comment that the proposition holds even if (G, S, V ) is secure only against attackers
that select queries according to the distribution G 2 (1n ). Furthermore, (G, S, V ) need
only be -restricted, for some suitable function  : N → N.

Proof Sketch: Consider an adversary A attacking the scheme (G  , S  , V  ). We may

ignore the case in which two queries of A are answered by triplets containing the
same one-time verification-key v  (because if this event occurs with non-negligible
probability, then the one-time scheme (G  , S  , V  ) cannot be secure). We consider two
cases regarding the relation of the one-time verification-keys included in the signatures
provided by Ss and the one-time verification-key included in the signature forged by A .

1. In case, for some i, the one-time verification-key v  contained in the forged message
equals the one-time verification-key v (i) contained in the answer to the i-th query,
we derive violation to the security of the one-time scheme (G  , S  , V  ).
544

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

Specifically, consider an adversary A that on input a verification-key v  for the one-

time scheme (G  , S  , V  ), generates (s, v) ← G(1n ) at random, selects i at random
(among polynomially many possibilities), invokes A on input v, and answers its
queries as follows. The i-th query of A , denoted α (i) , is answered by making the only
query to Ss  , obtaining β  = Ss  (α (i) ), and returning (Ss (v  ), v  , β  ) to A . (Note that A
holds s.) Each other query of A , denoted α ( j) , is answered by invoking G  to obtain
(s ( j) , v ( j) ) ← G  (1n ), and returning (Ss (v ( j) ), v ( j) , Ss ( j) (α ( j) ) to A . If A answers
with a forged signature and v is the verification-key contained in it, then A obtains
a forged signature relative to the one-time scheme (G  , S  , V  ) (i.e., a signature to a
message different from α (i) , which is valid with respect to the verification-key v  ).
Furthermore, conditioned on the case hypothesis and a forgery event, the second
event (i.e., v  is the verification-key contained in the forged signature) occurs with
probability 1/poly(n). Note that, indeed, A makes at most one query to Ss  , and that
the distribution seen by A is exactly as in an actual attack on (G  , S  , V  ).
2. In case, for all i, the one-time verification-key v  contained in the forged message is
different from the one-time verification-key v (i) contained in the answer to the i-th
query, we derive violation to the security of the scheme (G, S, V ).
Specifically, consider an adversary A that on input a verification-key v for the scheme
(G, S, V ), invokes A on input v, and answers its queries as follows. To answer the
j-th query of A , denoted α ( j) , algorithm A invokes G  to obtain (s ( j) , v ( j) ) ←
G  (1n ), queries Ss for a signature to v ( j) , and returns (Ss (v ( j) ), v ( j) , Ss ( j) (α ( j) )) to A .
When A answers with a forged signature and v  ∈ {v ( j) : j = 1, ..., poly(n)} is the
one-time verification-key contained in it, A obtains a forged signature relative to the
scheme (G, S, V ) (i.e., a signature to a string v  different from all v ( j) ’s, which is
valid with respect to the verification-key v). (Note again that the distribution seen
by A is exactly as in an actual attack on (G  , S  , V  ).)20
Thus, in both cases we derive a contradiction to some hypothesis, and the proposition
follows.

6.4.2.2. Authentication-Trees
The refreshing paradigm by itself (i.e., as employed in Construction 6.4.11) does not
seem to suffice for establishing Theorem 6.4.9. Recall that our aim is to construct
a general signature scheme based on a one-time signature scheme. The refreshing
paradigm suggests using a fresh instance of a one-time signature scheme in order to
sign the actual document; however, whenever we do so (as in Construction 6.4.11), we
must authenticate this fresh instance relative to the single verification-key that is public.
A straightforward implementation of this scheme (as presented in Construction 6.4.11)
calls for many signatures to be signed relative to the single verification-key that is
public, and so a one-time signature scheme cannot be used (for this purpose). Instead,
a more sophisticated method of authentication is called for.

20 Furthermore, all queries to Ss are distributed according to G 2 (1n ), justifying the comment made just before the
proof sketch.

545
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

x
sx vx
+ authx

x0 x1
sx0 vx0 sx1 vx1
+ authx0 + authx1

The (verification-key of a) node labeled x authenticates (the verification-keys of)

its children, labeled x0 and x1, respectively. The authentication is via a one-time
signature of the text vx0 vx1 using the signing-key sx , and it is verifiable with
respect to the verification-key vx .

Figure 6.4: Authentication-trees: the basic authentication step.

Let us try to sketch the basic idea underlying the new authentication method. The
idea is to use the public verification-key (of a one-time signature scheme) in order
to authenticate several (e.g., two) fresh instances (of the one-time signature scheme),
use each of these instances to authenticate several fresh instances, and so on. We
obtain a tree of fresh instances of the one-time signature, where each internal node
authenticates its children. We can now use the leaves of this tree in order to sign
actual documents, where each leaf is used at most once. Thus, a signature to an actual
document consists of (1) a one-time signature to this document authenticated with
respect to the verification-key associated with some leaf, and (2) a sequence of one-
time verification-keys associated with the nodes along the path from the root to this leaf,
where each such verification-key is authenticated with respect to the verification-key
of its parent (see Figures 6.4 and 6.5). We stress that each instance of the one-time
signature scheme is used to sign at most one string (i.e., several verification-keys if the
instance resides in an internal node, and an actual document if the instance resides in a
leaf).
This description may leave the reader wondering how one actually signs (and verifies
signatures) using the process outlined here. We start with a description that does not fit
our definition of a signature scheme, because it requires the signer to keep a record of
its actions during all previous invocations of the signing process.21 We refer to such a
scheme as memory dependent, and define this notion first.

21 This (memory) requirement will be removed in the next section.

546

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

λ
sλ vλ
+ authλ

0 1
s0 v0 s1 v1
+ auth 0 + auth1

00 01
s00 v00 s01 v01
+ auth 00 + auth 01

010 011
s010 v010 s011 v011
+ auth 010 + auth 011

Figure 6.5: An authentication path for nodes 010 and 011.

Definition 6.4.13 (memory-dependent signature schemes):

Mechanics: Item 1 of Definition 6.1.1 stays as it is, and the initial state (of the signing
algorithm) is defined to equal the output of the key-generator. Item 2 is modified
such that the signing algorithm is given a state, denoted γ , as auxiliary input and
returns a modified state, denoted δ, as auxiliary output. It is required that for every
pair (s, v) in the range of G(1n ), and for every α, γ ∈ {0, 1}∗ , if (β, δ) ← Ss (α, γ ),
then Vv (α, β) = 1 and |δ| ≤ |γ | + |α| · poly(n).
(That is, the verification algorithm accepts the signature β and the state does not
grow by too much.)
Security: The notion of a chosen message attack is modified so that the oracle Ss now
maintains a state that it updates in the natural manner; that is, when in state γ and
faced with query α, the oracle sets (β, δ) ← Ss (α, γ ), returns β, and updates its
state to δ. The notions of success and security are defined as in Definition 6.1.2,
except that they now refer to the modified notion of an attack.

The definition of memory-dependent signature schemes (i.e., Definition 6.4.13) is re-

lated to the definition of state-based MACs (i.e., Definition 6.3.10). However, there
are two differences between these two definitions: First, Definition 6.4.13 refers
547
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

to (public-key) signature schemes, whereas Definition 6.3.10 refers to MACs. Sec-

ond, in Definition 6.4.13, only the signing algorithm is state-based (or memory-
dependent), whereas in Definition 6.3.10 also the verification algorithm is state-
based. The latter difference reflects the difference in the applications envisioned for
both types of schemes. (Typically, MACs are intended for communication between
a predetermined set of “mutually synchronized” parties, whereas signature schemes
are intended for production of signatures that may be universally verifiable at any
time.)
We note that memory-dependent signature schemes may suffice in many applications
of signature schemes. Still, it is preferable to have memoryless (i.e., ordinary) signa-
ture schemes. In the following, we use any one-time signature schemes to construct a
memory-dependent signature scheme. The memory requirement will be removed in the
next section, so as to obtain a (memoryless) signature scheme (as in Definition 6.1.1).
The memory-dependent signature scheme presented (in Construction 6.4.14) main-
tains a binary tree of depth n, associating to each node an instance of a one-time
signature scheme. Each node in the tree is labeled by a binary string, denoted σ1 · · · σi
for some i ∈ {0, 1, ..., n}, and is associated with a (signing and verification) key-pair,
denoted (sσ1 ···σi , vσ1 ···σi ). The root of the tree is labeled by the empty string, λ, and
the verification-key vλ associated with it is used as the verification-key of the en-
tire (memory-dependent) signature scheme. The children of an internal node labeled
σ1 · · · σi are labeled σ1 · · · σi 0 and σ1 · · · σi 1, and their verification-keys (i.e., vσ1 ···σi 0
and vσ1 ···σi 1 ) are authenticated with respect to the verification-key vσ1 ···σi . With the ex-
ception of the (one-time) instance associated with the root of the tree, all the other
instances are generated (when needed) on the fly, and are stored in memory (along with
their authentication with respect to their parents). A new document is signed by allocat-
ing a new leaf, authenticating the actual document with respect to the verification-key
associated with this leaf, and authenticating each relevant verification-key with respect
to the verification-key associated with its parent. The relevant key-pairs (as well as their
authentication with respect to their parents) are generated on the fly, unless they are
already stored in memory (which means that they were generated in the course of sign-
ing a previous document). Thus, the verification-key associated with the relevant leaf
is authenticated with respect to the verification-key associated with its parent, which in
turn is authenticated with respect to the verification-key associated with its own par-
ent, and so on up to the authentication (of the verification-keys of the root’s children)
with respect to the verification-key associated with the root. The latter sequence of
authentications (of each node’s verification-key with respect to the verification-key of
its parent) is called an authentication path (see Figure 6.5). We stress that the (one-time)
instance associated with each node is used to authenticate at most one string. A formal
description of this memory-dependent signature scheme follows:

Construction 6.4.14 (a memory-dependent signature scheme): Let (G, S, V ) be a one-

time signature scheme. Consider the following memory-dependent signature scheme,
(G  , S  , V  ), with G  = G. On security parameter n, the scheme uses a full binary tree
of depth n. Each of the nodes in this tree is labeled by a binary string so that the root is
labeled by the empty string, denoted λ, and the left (resp., right) child of a node labeled
548

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

by x is labeled by x0 (resp., x1). Here we refer to the current state of the signing process
as to a record.

Initiating the scheme: To initiate the scheme, on security parameter n, we invoke G(1n )
and let (s, v) ← G(1n ). We record (s, v) as the key-pair associated with the root,
and output v as the (public) verification-key.
In the rest of the description, we denote by (sx , vx ) the key-pair associated with the
node labeled x; thus, (sλ , vλ ) = (s, v).
Signing with S  using the current record: Recall that the current record contains the
signing-key s = sλ , which is used to produce authλ (defined in the sequel).
To sign a new document, denoted α, we first allocate an unused leaf. Let σ1 · · · σn be
the label of this leaf. For example, we may keep a counter of the number of documents
signed, and determine σ1 · · · σn according to the counter value (e.g., if the counter
value is c, then we use the c-th string in lexicographic order).22
Next, for every i = 1, ..., n and every τ ∈ {0, 1}, we try to retrieve from our record
the key-pair associated with the node labeled σ1 · · · σi−1 τ . In case such a pair is not
found, we generate it by invoking G(1n ) and store it (i.e., add it to our record) for
future use; that is, we let (sσ1 ···σi−1 τ , vσ1 ···σi−1 τ ) ← G(1n ).
Next, for every i = 1, ..., n, we try to retrieve from our record a signature to the
string vσ1 ···σi−1 0 vσ1 ···σi−1 1 relative to the signing-key sσ1 ···σi−1 . In case such a signature
is not found, we generate it by invoking Ssσ1 ···σi−1 , and store it for future use; that
is, we obtain Ssσ1 ···σi−1 (vσ1 ···σi−1 0 vσ1 ···σi−1 1 ). (The ability to retrieve this signature from
memory, for repeated use, is the most important place in which we rely on the memory
dependence of our signature scheme.)23 We let
 
def
authσ1 ···σi−1 = vσ1 ···σi−1 0 , vσ1 ···σi−1 1 , Ssσ1 ···σi−1 (vσ1 ···σi−1 0 vσ1 ···σi−1 1 )

(Intuitively, via authσ1 ···σi−1 , the node labeled σ1 · · · σi−1 authenticates the
verification-keys associated with its children.)
Finally, we sign α by invoking Ssσ1 ···σn , and output

(σ1 · · · σn , authλ , authσ1 , ..., authσ1 ···σn−1 , Ssσ1 ···σn (α))

Verification with V  : On input a verification-key v, a document α, and an alleged

signature β, we accept if and only if the following conditions hold:

1. β has the form

(σ1 · · · σn , (v0,0 , v0,1 , β0 ), (v1,0 , v1,1 , β1 ), ..., (vn−1,0 , vn−1,1 , βn−1 ), βn )

22 Alternatively, as done in Construction 6.4.16, we may select the leaf at random (while ignoring the negligible
probability that the selected leaf is not unused).
23 This allows the signing process S  to use each (one-time) signing-key s for producing a single S -signature.
s x sx
In contrast, the use of a counter for determining a new leaf can be easily avoided, by selecting a leaf at random.

549
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

where the σi ’s are bits and all other symbols represent strings.
(Jumping ahead, we mention that vi,τ is supposed to equal vσ1 ···σi τ ; that is,
the verification-key associated by the signing process with the node labeled
σ1 · · · σi τ . In particular, vi−1,σi is supposed to equal vσ1 ···σi .)
2. Vv (v0,0 v0,1 , β0 ) = 1.
(That is, the public-key (i.e., v) authenticates the two strings v0,0 and v0,1 claimed
to correspond to the instances of the one-time signature scheme associated with
the nodes labeled 0 and 1, respectively.)
3. For i = 1, ..., n − 1, it holds that Vvi−1,σi (vi,0 vi,1 , βi ) = 1.
(That is, the verification-key vi−1,σi , which is already believed to be authentic
and supposedly corresponds to the instance of the one-time signature scheme
associated with the node labeled σ1 · · · σi , authenticates the two strings vi,0 and
vi,1 that are supposed to correspond to the instances of the one-time signature
scheme associated with the nodes labeled σ1 · · · σi 0 and σ1 · · · σi 1, respectively.)
4. Vvn−1,σn (α, βn ) = 1.
(That is, the verification-key vn−1,σn , which is already believed to be authentic,
authenticates the actual document α.)

Regarding the verification algorithm, note that Conditions 2 and 3 establish that vi,σi+1 is
authentic (i.e., equals vσ1 ···σi σi+1 ). That is, v = vλ authenticates vσ1 , which authenticates
vσ1 σ2 , and so on up-to vσ1 ···σn . The fact that the vi,σ i+1 ’s are also proven to be authentic
(i.e., equal to the vσ1 ···σi σ i+1 ’s, where σ = 1 − σ ) is not really useful (when signing a
message using the leaf associated with σ1 · · · σn ). This excess is merely an artifact of
the need to use sσ1 ···σi only once during the entire operation of the memory-dependent
signature scheme: In the currently (constructed) Ss -signature, we may not care about the
authenticity of some vσ1 ···σi σ i+1 , but we may care about it in some other Ss -signatures.
For example, if we use the leaf labeled 0n to sign the first document and the leaf
labeled 0n−1 1 to sign the second, then in the first Ss -signature we care only about the
authenticity of v0n , whereas in the second Ss -signature we care about the authenticity
of v0n−1 1 .

Proposition 6.4.15: If (G, S, V ) is a secure one-time signature scheme, then Construc-

tion 6.4.14 constitutes a secure memory-dependent signature scheme.

Proof: Recall that a Ssλ -signature to a document α has the form

(σ1 · · · σn , authλ , authσ1 , ..., authσ1 ···σn−1 , Ssσ1 ···σn (α)) (6.5)

where the authx ’s, vx ’s, and sx ’s satisfy

authx = vx0 , vx1 , Ssx (vx0 vx1 ) (6.6)

(See Figure 6.4.) In this case, we say that this Ss -signature uses the leaf labeled
σ1 · · · σn . For every i = 1, ..., n, we call the sequence (authλ , authσ1 , ..., authσ1 ···σi−1 )

550

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

an authentication path for vσ1 ···σi ; see Figure 6.5. (Note that this sequence is also an
authentication path for vσ1 ···σi−1 σ i , where σ = 1 − σ .) Thus, a valid Ss -signature to a
document α consists of an n-bit string σ1 · · · σn , authentication paths for each vσ1 ···σi
(i = 1, ..., n), and a signature to α with respect to the one-time scheme (G, S, V ) using
the signing-key sσ1 ···σn .
Intuitively, forging an Ss -signature requires either using only verification-keys sup-
plied by the signer (i.e., supplied by Ss as part of an answer to a query) or producing
an authentication path for a verification-key that is different from all verification-keys
supplied by the signer. In both cases, we reach a contradiction to the security of the one-
time signature scheme (G, S, V ). Specifically, in the first case, the forged Ss -signature
contains a one-time signature that is valid with respect to the one-time verification-key
associated by the signing process with a leaf labeled σ1 · · · σn , because by the case’s
hypothesis, the forged signature utilizes only verification-keys supplied by the signer.
This yields forgery with respect to the instance of the one-time signature scheme as-
sociated with the leaf labeled σ1 · · · σn (because the document that is Ss -signed by the
forger must be different from all Ss -signed documents, and thus the forged document
is different from all strings to which a one-time signature associated with a leaf was
applied).24 We now turn to the second case (i.e., forgery with respect to (G  , S  , V  ) is
obtained by producing an authentication path for a verification-key that is different from
all verification-keys supplied by the signer). As in the first case, we denote by σ1 · · · σn
the label of the leaf used for the (forged) signature. Let i ∈ {0, ..., n − 1} be the largest
integer such that the signature produced by the forger refers to the verification-key
vσ1 ···σi (as supplied by the signer), rather than to a different value (claimed by the forger
to be the verification-key associated with the node labeled σ1 · · · σi ). (Note that i = 0
corresponds to the forger not even using vσ1 , whereas i < n by the case hypothesis.)
 
For this i, the triple authσ1 ···σi = (vi,0 , vi,1 , βi ) that is contained in the Ss -signature pro-
duced by the forger contains a one-time signature (i.e., βi ) that is valid with respect to
the one-time verification-key associated by the signing process with the node labeled
σ1 · · · σi (where vλ is always used by the signing process). Furthermore, by maximality
 
of i, the latter signature is to a string (i.e., vi,0 vi,1 ) that is different from the string
 
to which the Ss -signer has applied Ssσ1 ···σi (i.e., vi,σi+1 = vσ1 ···σi+1 ). This yields forgery
with respect to the instance of the one-time signature scheme associated with the node
labeled σ1 · · · σi .
The actual proof is by a reducibility argument. Given an adversary A attacking the
complex scheme (G  , S  , V  ), we construct an adversary A that attacks the one-time
signature scheme, (G, S, V ). In particular, the adversary A will use its (one-time) oracle
access to Ss in order to emulate the memory-dependent signing oracle for A . We stress
that the adversary A may make at most one query to its Ss -oracle. Following is a detailed
description of the adversary A. Since we care only about probabilistic polynomial-time
adversaries, we may assume that A makes at most t = poly(n) many queries, where n
is the security parameter.

24 Note that what matters is merely that the document Ss -signed by the forger is different from the (single) document
to which Ssσ1···σn was applied by the Ss -signer, in case Ssσ1···σn was ever applied by the Ss -signer.

551
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

The Construction of Adversary A: Suppose that (s, v) is in the range of G(1n ). On

input v and one-query oracle access to Ss , adversary A proceeds as follows:

1. Initial choice: A uniformly selects j ∈ {1, ..., (2n + 1) · t}.

(The integer j specifies an instance of (G, S, V ) generated during the emulated attack
of A on (G  , S  , V  ). This instance will be attacked by A. Note that since 2n + 1
instances of (G, S, V ) are referred to in each signature relative to (G  , S  , V  ), the
quantity (2n + 1) · t upper-bounds the total number of instances of (G, S, V ) that
appear during the entire attack of A . This upper bound is not tight.)

2. Invoking A : If j = 1, then A sets vλ = v and invokes A on input v. In this case A

does not know sλ , which is defined to equal s, yet A can obtain a single signature
relative to the signing-key s by making a (single) query to its own oracle (i.e., the
oracle Ss ).
Otherwise (i.e., j > 1), machine A invokes G, obtains (s  , v  ) ← G(1n ), sets
(sλ , vλ ) = (s  , v  ), and invokes A on input v  . We stress that in this case A
knows sλ .
Indeed, in both cases, A is invoked on input vλ . Also, in both cases, the one-
time instance associated with the root (i.e., the node labeled λ) is called the first
instance.

3. Emulating the memory-dependent signing oracle for A : The emulation is analogous

to the operation of the signing procedure as specified in Construction 6.4.14. The
only exception refers to the j-th instance of (G, S, V ) that occurs in the memory-
dependent signing process. Here, A uses the verification key v, and if an Ss -signature
needs to be produced, then A queries Ss for it. We stress that at most one signature
need ever be produced with respect to each instance of (G, S, V ) that occurs in the
memory-dependent signing process, and therefore Ss is queried at most once. Details
follow.
Machine A maintains a record of all key-pairs and one-time signatures it has gen-
erated and/or obtained from Ss . When A is asked to supply a signature to a new
document, denoted α, it proceeds as follows:

(a) A allocates a new leaf-label, denoted σ1 · · · σn , exactly as done by the signing

process.
(b) For every i = 1, ..., n and every τ ∈ {0, 1}, machine A tries to retrieve from its
record the one-time instance associated with the node labeled σ1 · · · σi−1 τ . If
such an instance does not exist in the record (i.e., the one-time instance associated
with the node labeled σ1 · · · σi−1 τ did not appear so far), then A distinguishes
two cases:
i. If the record so far contains exactly j − 1 one-time instances (i.e., the current
instance is the j-th one to be encountered), then A sets vσ1 ···σi−1 τ ← v, and

552

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

adds it to its record. In this case, A does not know sσ1 ···σi−1 τ , which is defined to
equal s, yet A can obtain a single signature relative to s by making a (single)
query to its own oracle (i.e., the oracle Ss ).
From this point on, the one-time instance associated with the node labeled
σ1 · · · σi−1 τ will be called the j-th instance.
ii. Otherwise (i.e., the current instance is not the j-th one to be encountered), A
acts as the signing process: It invokes G(1n ), obtains (sσ1 ···σi−1 τ , vσ1 ···σi−1 τ ) ←
G(1n ), and adds it to the record. (Note that in this case, A knows sσ1 ···σi−1 τ
and can generate by itself signatures relative to it.)
The one-time instance just generated is given the next serial number. That
is, the one-time instance associated with the node labeled σ1 · · · σi−1 τ will
be called the k-th instance if the current record (i.e., after the generation of
the one-time key-pair associated with the node labeled σ1 · · · σi−1 τ ) contains
exactly k instances.

(c) For every i = 1, ..., n, machine A tries to retrieve from its record a (one-time)
signature to the string vσ1···σi−1 0 vσ1···σi−1 1 , relative to the signing-key sσ1···σi−1 .
If such a signature does not exist in the record then A distinguishes two
cases:
i. If the one-time signature instance associated with the node labeled
σ1 · · · σi−1 is the j-th such instance, then A obtains the one-time signa-
ture Ssσ1···σi−1 (vσ1···σi−1 0 vσ1···σi−1 1 ) by querying Ss , and adds this signature to
the record.
Note that by the previous steps (i.e., Step 3(b)i as well as Step 2), s is identified
with sσ1···σi−1 , and that the instance associated with a node labeled σ1 · · · σi−1 is
only used to produce a single signature; that is, to the string vσ1···σi−1 0 vσ1···σi−1 1 .
Thus, in this case, A queries Ss at most once.
We stress that this makes crucial use of the fact that, for every τ , the
verification-key associated with the node labeled σ1 · · · σi−1 τ is identical in all
executions of the current step. This fact guarantees that A only needs a single
signature relative to the instance associated with a node labeled σ1 · · · σi−1 ,
and thus queries Ss at most once (and retrieves this signature from memory
if it ever needs this signature again).
ii. Otherwise (i.e., the one-time signature instance associated with the node
labeled σ1 · · · σi−1 is not the j-th such instance), A acts as the signing process:
It invokes Ssσ1···σi−1 , obtains the one-time signature Ssσ1···σi−1 (vσ1···σi−1 0 vσ1···σi−1 1 ),
and adds it to the record. (Note that in this case, A knows sσ1···σi−1 and can
generate by itself signatures relative to it.)
Thus, in both cases, A obtains authσ1···σi−1 = (vσ1···σi−1 0 , vσ1···σi−1 1 , βi−1 ), where
βi−1 = Ssσ1···σi−1 (vσ1···σi−1 0 vσ1···σi−1 1 ).

(d) Machine A now obtains a one-time signature of α relative to Ssσ1···σn . (Since

a new leaf is allocated for each query made by A , we need to generate at
most one signature relative to the one-time instance Ssσ1···σn associated with the

553
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

leaf σ1 · · · σn .) This is done analogously to the previous step (i.e., Step 3c).
Specifically:

i. If the one-time signature instance associated with the (leaf) node labeled
σ1 · · · σn is the j-th instance, then A obtains the one-time signature Ssσ1···σn (α)
by querying Ss .
Note that in this case, s is identified with sσ1···σn , and that an instance associated
with a leaf is only used to produce a single signature. Thus, also in this case
(which is disjoint of Case 3(c)i), A queries Ss at most once.
ii. Otherwise (i.e., the one-time signature instance associated with the node
labeled σ1 · · · σn is not the j-th instance), A acts as the signing process: It
invokes Ssσ1···σn and obtains the one-time signature Ssσ1···σn (α). (Again, in this
case A knows sσ1···σn and can generate by itself signatures relative to it.)
Thus, in both cases, A obtains βn = Ssσ1···σn (α).

(e) Finally, A answers the query α with

(σ1 · · · σn , authλ , authσ1 , ..., authσ1···σn−1 , βn )

4. Using the output of A : When A halts with output (α  , β  ), machine A checks

whether this is a valid document-signature pair with respect to Vvλ and whether the
document α  did not appear as a query of A . If both conditions hold, then A tries
to obtain forgery with respect to Ss . To explain how this is done, we need to take a
closer look at the valid document-signature pair, (α  , β  ), output by A . Specifically,
suppose that β  has the form

(σ1 · · · σn , (v0,0

 
, v0,1 , β0 ), (v1,0
 
, v1,1 , β1 ), ..., (vn−1,0
 
, vn−1,1 
, βn−1 ), βn )

and that the various components satisfy all conditions stated in the verification
 
procedure. (In particular, the sequence (v0,0 , v0,1 , β0 ), ..., (vn−1,0
 
, vn−1,1 
, βn−1 ) is
 
the authentication path (for vn−1,σn ) output by A .) Recall that strings of the form

vk,τ denote the verification-keys included in the output of A , whereas strings of the
form vx denote the verification-keys (as used in the answers given to A by A and)
as recorded by A.
   
Let i be maximal such that the sequence of key-pairs (v0,0 , v0,1 ), ..., (vi−1,0 , vi−1,1 )
 25
appears in some authentication path supplied to A (by A). Note that
 
i ∈ {0, ..., n}, where i = 0 means that (v0,0 , v0,1 ) differs from (v0 , v1 ), and
   
i = n means that the sequence ((v0,0 , v0,1 ), ..., (vn−1,0 , vn−1,1 )) equals the

sequence ((v0 , v1 ), ..., (vσ1···σn−1
 , v  
0 σ1···σn−1 1 )). In general, the sequence ((v0,0 ,
  
v0,1 ), ..., (vi−1,0 , vi−1,1 )) equals the sequence ((v0 , v1 ), ..., (vσ1···σi−1 0 , vσ1···σi−1 1 )). In
   

particular, for i ≥ 1, it holds that vi−1,σ  = vσ ···σ  , whereas for i = 0 we shall only
1 i
i
25 That is, i is such that for some β0 , ..., βi−1 (which may but need not equal β0 , ..., βi−1  ), the sequence
   
(v0,0 , v0,1 , β0 ), ..., (vi−1,0 , vi−1,1 , βi−1 ) is a prefix of some authentication path (for some vσ ···σ  σi+1···σn ) sup-
1 i
plied to A by A. We stress that here we only care about whether or not some vk,τ  ’s equal the corresponding

verification-keys supplied by A, and ignore the question of whether (in case of equality) the verification-keys
were authenticated using the very same (one-time) signature. We mention that things will be different in the
analogous part of the proof of Theorem 6.5.2 (which refers to super-security).

554

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

refer to vλ (which is the verification-key attacked by A ). In both cases, the output

of A contains a one-time signature relative to vσ1···σi , and this signature is to a string
different from the (possibly) only one to which a signature was supplied to A by
A. Specifically, as in the motivating discussion (in the beginning of the proof), we
distinguish the cases i = n and i < n:

(a) In case i = n, the output of A contains the (one-time) signature βn that satisfies
Vvσ ···σn (α  , βn ) = 1. Furthermore, α  is different from the (possibly) only docu-
1
ment to which Ssσ ···σn was applied during the emulation of the S  -signer by A,
1
since by our hypothesis the document α  did not appear as a query of A . (Re-
call that by the construction of A, instances of the one-time signature scheme
associated with leaves are only applied to the queries of A .)
(b) In case i < n, the output of A contains the (one-time) signature βi that satisfies
 
Vvσ ···σ  (vi,0 vi,1 , βi ) = 1. Furthermore, vi,0
 
vi,1 is different from vσ1···σi 0 vσ1···σi 0 ,
1 i
which is the (possibly) only string to which Ssσ ···σ  was applied during the emu-
1 i
lation of the S  -signer by A, where the last assertion is due to the maximality of
i (and the construction of A).

Thus, in both cases, A obtains from A a valid (one-time) signature relative to the
(one-time) instance associated with the node labeled σ1 · · · σi . Furthermore, in both
cases, this (one-time) signature is to a string that did not appear in the record of A.
The question is whether the instance associated with the node labeled σ1 · · · σi is
the j-th instance, for which A set v = vσ1···σi . In case the answer is yes, A obtains
forgery with respect to the (one-time) verification-key v (which it attacks).
In view of this discussion, A acts as follows. It determines i as in the beginning of the
current step (i.e., Step 4), and checks whether v = vσ1···σi (or, almost equivalently,
whether the j-th instance is the one associated with the node labeled σ1 · · · σi ). In
case i = n, machine A outputs the string-signature pair (α  , βn ); otherwise (i.e.,
 
i < n) it outputs the string-signature pair (vi,0 vi,1 , βi ).

This completes the (admittingly long) description of adversary A. We repeat again

some obvious observations regarding this construction. Firstly, A makes at most one
query to its (one-time) signing oracle Ss . Secondly, assuming that A is probabilistic
polynomial-time, so is A. Thus, all that remains is to relate the success probability of A
(when attacking a random instance of (G, S, V )) to the success probability of A (when
attacking a random instance of (G  , S  , V  )). As usual, the main observation is that the
view of A , during the emulation of the memory-dependent signing process (by A), is
identically distributed to its view in an actual attack on (G  , S  , V  ). Furthermore, this
holds conditioned on any possible fixed value of j (selected in the first step of A). It
follows that if A succeeds in forging signatures in an actual attack on (G  , S  , V  ) with
probability ε (n), then A succeeds in forging signatures with respect to (G, S, V ) with
ε (n)
probability at least (2n+1)·t , where the (2n + 1) · t factor is due to the probability that
the choice of j is a good one (i.e., so that the j-th instance is the one associated with the
node labeled σ1 · · · σi , where σ1 · · · σn and i are as defined in Step 4 of A’s construction).
555
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

We conclude that if (G  , S  , V  ) can be broken by a probabilistic polynomial-time

chosen message attack with non-negligible probability, then (G, S, V ) can be broken by
a probabilistic polynomial-time single-message attack with non-negligible probability,
in contradiction to the proposition’s hypothesis. The proposition follows.

6.4.2.3. The Actual Construction

In this section, we remove the memory dependency of Construction 6.4.14 and obtain
an ordinary (rather than memory-dependent) signature scheme. Toward this end, we
use pseudorandom functions (as defined in Definition 3.6.4). The basic idea is that
the record maintained in Construction 6.4.14 can be determined (on the fly) by an
application of a pseudorandom function to certain strings. For example, instead of
generating and storing an instance of a (one-time) signature scheme for each node that
we encounter, we can determine the randomness for the (corresponding invocation of
the) key-generation algorithm as a function of the label of that node. Thus, there is no
need to store the key-pair generated, because if we ever need it again, then regenerating
it (in the very same way) will yield exactly the same result. The same idea applies
also to the generation of (one-time) signatures. In fact, the construction is simplified,
because we need not check whether or not we are generating an object for the first time.
For simplicity, let us assume that on security parameter n, both the key-generation
and signing algorithms (of the one-time signature scheme (G, S, V )) use exactly n
internal coin tosses. (This assumption can be justified by using pseudorandom gen-
erators, which exist anyhow under the assumptions used here.) For r ∈ {0, 1}n , we
denote by G(1n , r ) the output of G on input 1n and internal coin-tosses r . Likewise,
for r ∈ {0, 1}n , we denote by Ss (α, r ) the output of S, on input a signing-key s and
a document α, when using internal coin-tosses r . For simplicity, we shall actually be
using generalized pseudorandom functions as in Definition 3.6.12 (rather than pseu-
dorandom functions as defined in Definition 3.6.4).26 Furthermore, for simplicity, we
shall consider applications of such pseudorandom functions to sequences of characters
containing {0, 1}, as well as a few additional special characters.

Construction 6.4.16 (Removing the memory requirement from Construction 6.4.14):

Let (G, S, V ) be a one-time signature scheme, and { f r : {0, 1}∗ → {0, 1}|r | }r ∈{0,1}∗ be
a generalized pseudorandom function ensemble as in Definition 3.6.12. Consider the
following signature scheme, (G  , S  , V  ), which refers to a full binary tree of depth n as
in Construction 6.4.14:
Key-generation algorithm G  : On input 1n , algorithm G  obtains (s, v) ← G(1n ) and
selects uniformly r ∈ {0, 1}n . Algorithm G  outputs the pair ((r, s), v), where (r, s)
is the signing-key and v is the verification-key.27

26 We shall make comments regarding the minor changes required in order to use ordinary pseudorandom functions.
The first comment is that we shall consider an encoding of strings of length up to n + 2 by strings of length
n + 3 (e.g., for i ≤ n + 2, the string x ∈ {0, 1}i is encoded by x10n+2−i ).
27 In case we use ordinary pseudorandom functions, rather than generalized ones, we select r uniformly in {0, 1}n+3

such that fr : {0, 1}n+3 → {0, 1}n+3 . Actually, we shall be using the function f r : {0, 1}n+3 → {0, 1}n derived
from the original fr by dropping the last 3 bits of the function value.

556

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

Signing algorithm S  : On input a signing-key (r, s) (in the range of G 1 (1n )) and a
document α, the algorithm proceeds as follows:
1. It selects uniformly σ1 · · · σn ∈ {0, 1}n .
(Algorithm S  will use the leaf labeled σ1 · · · σn ∈ {0, 1}n to sign the current doc-
ument. Indeed, with exponentially vanishing probability, the same leaf may be
used to sign two different documents, and this will lead to forgery [but only with
negligible probability].)
(Alternatively, to obtain a deterministic signing algorithm, one may set σ1 · · ·
σn ← fr (select-leaf, α), where select-leaf is a special character.)28
2. Next, for every i = 1, ..., n and every τ ∈ {0, 1}, the algorithm invokes G and sets
(sσ1···σi−1 τ , vσ1···σi−1 τ ) ← G(1n , fr (key-gen, σ1 · · · σi−1 τ ))
where key-gen is a special character.29
3. For every i = 1, ..., n, the algorithm invokes Ssσ1···σi−1 and sets
def
authσ1···σi−1 = vσ1···σi−1 0 , vσ1···σi−1 1 ,
Ssσ1···σi−1 (vσ1···σi−1 0 vσ1···σi−1 1 , fr (sign, σ1 · · · σi−1 )))

where sign is a special character.30

4. Finally, the algorithm invokes Ssσ1···σn and outputs31
(σ1 · · · σn , authλ , authσ1 , ..., authσ1···σn−1 , Ssσ1···σn (α, fr (sign, σ1 · · · σn )))
Verification algorithm V  : On input a verification-key v, a document α, and an alleged
signature β, algorithm V  behaves exactly as in Construction 6.4.14. Specifically,
assuming that β has the form
(σ1 · · · σn , (v0,0 , v0,1 , β0 ), (v1,0 , v1,1 , β1 ), ..., (vn−1,0 , vn−1,1 , βn−1 ), βn )
algorithm V  accepts if and only if the following three conditions hold:
r Vv (v0,0 v0,1 , β0 ) = 1.
r For i = 1, ..., n − 1, it holds that Vv (vi,0 vi,1 , βi ) = 1.
r Vv
i−1,σi
(α, βn ) = 1.
n−1,σn

Proposition 6.4.17: If (G, S, V ) is a secure one-time signature scheme and { f r :

{0, 1}∗ → {0, 1}|r | }r ∈{0,1}∗ is a generalized pseudorandom function ensemble, then Con-
struction 6.4.16 constitutes a secure (general) signature scheme.

28 In case we use ordinary pseudorandom functions, rather than generalized ones, this alternative can be (directly)
implemented only if it is guaranteed that |α| ≤ n. In such a case, we apply the f r to the (n + 3)-bit encoding of
00α.
29 In case we use ordinary pseudorandom functions, rather than generalized ones, the argument to f is the
r
(n + 3)-bit encoding of 10σ1 · · · σi−1 τ .
30 In case we use ordinary pseudorandom functions, rather than generalized ones, the argument to f is the
r
(n + 3)-bit encoding of 11σ1 · · · σi−1 .
31 In case we use ordinary pseudorandom functions, rather than generalized ones, the argument to f is the
r
(n + 3)-bit encoding of 11σ1 · · · σn .

557
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Proof: Following the general methodology suggested in Section 3.6.3, we consider an

ideal version of Construction 6.4.16 in which a truly random function is used (rather than
a pseudorandom one). The ideal version is almost identical to Construction 6.4.14, with
the only difference being the way in which σ1 · · · σn is selected. Specifically, applying
a truly random function to determine (one-time) key-pairs and (one-time) signatures is
equivalent to generating these keys and signatures at random (on the fly) and reusing
the stored values whenever necessary. Regarding the way in which σ1 · · · σn is selected,
observe that the proof of Proposition 6.4.15 is oblivious of this way, except for the
assumption that the same leaf is never used to sign two different documents. However,
the probability that the same leaf is used twice by the (memoryless) signing algorithm,
when serving polynomially many signing requests, is exponentially vanishing and thus
can be ignored in our analysis. We conclude that the ideal scheme (in which a truly
random function is used instead of f r ) is secure. It follows that also the actual signature
scheme (as in Construction 6.4.16) is secure, or else one can efficiently distinguish a
pseudorandom function from a truly random one (which is impossible). Details follow.
Assume toward the contradiction that there exists a probabilistic polynomial-time
adversary A that succeeds in forging signatures with respect to (G  , S  , V  ) with non-
negligible probability, but succeeds only with negligible probability when attacking
the ideal scheme. We construct a distinguisher D that on input 1n and oracle access to
f : {0, 1}∗ → {0, 1}n behaves as follows. Machine D generates ((r  , s), v) ← G  (1n )
and invokes A on input v. Machine D answers the queries of A by running the signing
process, using the signing-key (r  , s), with the exception that it replaces the values
fr  (x) by f (x). That is, whenever the signing process calls for the computation of the
value of the function fr  on some string x, machine D queries its oracle (i.e., f ) on
the string x, and uses the response f (x) instead of fr  (x). When A outputs an alleged
signature to a new document, machine M evaluates whether or not the signature is
valid (with respect to Vv ) and outputs 1 if and only if A has indeed succeeded (i.e., the
signature is valid). Observe that if D is given oracle access to a truly random function,
then the emulated A attacks the ideal scheme, whereas if D is given oracle access to
a pseudorandom function fr , then the emulated A attacks the real scheme. It follows
that D distinguishes the two cases, in contradiction to the pseudorandomness of the
ensemble { fr }.

6.4.2.4. Conclusions and Comments

Theorem 6.4.9 follows by combining Proposition 6.4.17 with the fact that the exis-
tence of secure one-time signature schemes implies the existence of one-way functions
(see Exercise 13), which in turn implies the existence of (generalized) pseudorandom
functions. Recall that by combining Theorem 6.4.9 and Corollary 6.4.8, we obtain
Corollary 6.4.10, which states that the existence of collision-free hashing collections
implies the existence of secure signature schemes. Furthermore, the length of the re-
sulting signatures depends only on the length of the signing-key.
We comment that Constructions 6.4.14 and 6.4.16 can be generalized as follows.
Rather than using a (depth n) full binary tree, one can use any tree that has a super-
polynomial (in n) number of leaves, provided that one can enumerate the leaves (resp.,
558

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

uniformly select a leaf) and generate the path from the root to a given leaf. We consider
a few possibilities:

r For any d : N → N bounded by a polynomial in n (e.g., d ≡ 2 or d(n) = n are indeed

“extreme” cases), we may consider a full d(n)-ary tree of depth e(n) so that d(n)e(n) is
greater than any polynomial in n. The choice of parameters in Constructions 6.4.14
and 6.4.16 (i.e., d ≡ 2 and e(n) = n) is probably the simplest one.
Natural complexity measures for a signature scheme include the length of signatures
and the signing and verification times. In a generalized construction, the length of the
signatures is linear in d(n) · e(n), and the number of applications of the underlying
one-time signature scheme (per each general signature) is linear in e(n), where in
internal nodes the one-time signature scheme is applied to a string of length linear in
d(n). Assuming that the complexity of one-time signatures is linear in the document
length, all complexity measures are linear in d(n) · e(n), and so d ≡ 2 is the best
generic choice. However, this assumption may not hold when some specific one-time
signatures are used. For example, the complexity of producing a signature to an -bit
long string in a one-time signature scheme may be of the form p(n) + p  (n) · , where
p  (n)  p(n). In such (special) cases, one may prefer to use a larger d : N → N (see
Section 6.6.5).
r For the memory-dependent construction, it may be preferable to use unbalanced
trees (i.e., having leaves at various levels). The advantage is that if one utilizes first
the leaves closer to the root, then one can obtain a saving on the cost of signing the
first documents.
For example, consider using a ternary tree of super-logarithmic depth (i.e., d ≡ 3
and e(n) = ω(log n)), in which each internal node of level i ∈ {0, 1, ..., e(n) − 2}
has two children that are internal nodes and a single child that is a leaf (and the
internal nodes of level e(n) − 1 have only leaves as children). Thus, for i ≥ 1, there
are 3i−1 leaves at level i. If we use all leaves of level i before using any leaf of level
i + 1, then the length of the j-th signature in this scheme is linear in log3 j (and so
is the number of applications of the underlying one-time signature scheme).

When actually applying these constructions, one should observe that in variants
of Construction 6.4.14, the size of the tree determines the number of documents that
can be signed, whereas in variants of Construction 6.4.16, the tree size has an even
more drastic effect on the number of documents that can be signed.32 In some cases, a
hybrid of Constructions 6.4.14 and 6.4.16 may be preferable: We refer to a memory-
dependent scheme in which leaves are assigned as in Construction 6.4.14 (i.e., according
to a counter), but the rest of the operation is done as in Construction 6.4.16 (i.e., the
one-time instances are regenerated on the fly, rather than being recorded and retrieved

32 In particular, the number of documents that can be signed should definitely be smaller than the square root of
the size of the tree (or else two documents are likely to be assigned the same leaf ). Furthermore, we cannot use
a small tree (e.g., of size 1,000) even if we know that the total number of documents that will ever be signed is
small (e.g., 10), because in this case, the probability that two documents are assigned the same leaf is too big
(e.g., 1/20).

559
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

from memory). In some applications, the introduction of a document-counter may be

tolerated, and the gain is the ability to use a smaller tree (i.e., of size merely greater
than the number of documents that should be ever signed).
More generally, we wish to stress that each of the following ingredients of the
previous constructions is useful in a variety of related and unrelated settings. We refer
specifically to the refreshing paradigm, the authentication-tree construction, and the
notion (and constructions) of one-time signatures. For example:
r It is common practice to authenticate messages sent during a “communication ses-
sion” via a (fresh) session-key that is typically authenticated by a master-key. One
of the reasons for this practice is the prevention of a chosen message attack on the
(more valuable) master-key. (Other reasons include allowing the use of a faster (alas,
less secure) authentication scheme for the actual communication and introducing in-
dependence between sessions.)
r Observe the analogy between the tree-hashing (of Construction 6.2.13) and the
authentication-tree (of Construction 6.4.14). Despite the many differences, in both
cases the value of each internal node authenticates the values of its children. Thus,
the value of the root may be used to authenticate a very large number of values
(associated with the leaves). Furthermore, the value associated with each leaf can
be verified within complexity that is linear in the depth of the tree.
r Recall the application of one-time signatures to the construction of CCA-secure
public-key encryption schemes (see the proof of Theorem 5.4.30).

6.4.3.* Universal One-Way Hash Functions and Using Them

So far, we have established that the existence of collision-free hashing collections
implies the existence of secure signature schemes (cf. Corollary 6.4.10). We seek to
weaken the assumption under which secure signature schemes can be constructed,
and bear in mind that the existence of one-way functions is certainly a necessary
condition (cf., for example, Exercise 13). In view of Theorem 6.4.9, we may focus
on constructing secure one-time signature schemes. Furthermore, recall that secure
length-restricted one-time signature schemes can be constructed based on any one-way
function (cf. Corollary 6.4.6). Thus, the only bottleneck we face (with respect to the
assumption used) is the transformation of length-restricted one-time signature schemes
into (general) one-time signature schemes. For the latter transformation, we have used
a specific incarnation of the “hash-and-sign paradigm” (i.e., Proposition 6.4.7, which
refers to Construction 6.2.6). This incarnation utilizes collision-free hashing, and our
goal is to replace it by a variant (of Construction 6.2.6) that uses a seemingly weaker
notion called Universal One-Way Hash Functions.

6.4.3.1. Definition
A collection of universal one-way hash functions is defined analogously to a collection of
collision-free hash functions. The only difference is that the hardness (to form collisions)
requirement is relaxed. Recall that in the case of (a collection of) collision-free hash
560

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

functions, it was required that, given the function’s description, it is hard to form an
arbitrary collision under the function. In the case of (a collection of) universal one-way
hash functions, we only require that, given the function’s description h and a pre-image
x0 , it is hard to find an x = x 0 so that h(x) = h(x0 ). We refer to this requirement as to
hardness to form designated collisions.
Our formulation of the hardness to form designated collisions is actually seem-
ingly stronger. Rather than being supplied with a (random) pre-image x0 , the collision-
forming algorithm is allowed to select x 0 by itself, but must do so before being presented
with the function’s description. That is, the attack of the collision-forming algorithm
proceeds in three stages: First the algorithm selects a pre-image x 0 , next it is given a
description of a randomly selected function h, and finally it is required to output x = x0
such that h(x) = h(x0 ). We stress that the third stage in the attack is also given the
random coins used for producing the initial pre-image (at the first stage). This yields
the following definition, where the first stage is captured by a deterministic polynomial-
time algorithm A0 (which maps a sequence of coin tosses, denoted Uq(n) , to a pre-image
of the function), and the third stage is captured by algorithm A (which is given the very
same coins Uq(n) as well as the function’s description).

Definition 6.4.18 (universal one-way hash functions – UOWHF): Let  : N → N. A

collection of functions {h s : {0, 1}∗ → {0, 1}(|s|) }s∈{0,1}∗ is called universal one-way
hashing (UOWHF) if there exists a probabilistic polynomial-time algorithm I so that
the following holds:

1. (admissible indexing – technical):33 For some polynomial p, all sufficiently large

n’s, and every s in the range of I (1n ), it holds that n ≤ p(|s|). Furthermore, n can
be computed in polynomial-time from s.
2. (efficient evaluation): There exists a polynomial-time algorithm that, given s and x,
returns h s (x).
3. (hard-to-form designated collisions): For every polynomial q, every deterministic
polynomial-time algorithm A0 , every probabilistic polynomial-time algorithm A,
every positive polynomial p, and all sufficiently large n’s


h I (1n ) (A(I (1n ), Uq(n) )) = h I (1n ) (A0 (Uq(n) )) 1
Pr < (6.7)
and A(I (1n ), Uq(n) ) = A0 (Uq(n) ) p(n)
where the probability is taken over Uq(n) and the internal coin tosses of algorithms
I and A.

The function  is called the range specifier of the collection.

We stress that the hardness to form designated collisions condition refers to the
following three-stage process: First, using a uniformly distributed r ∈ {0, 1}q(n) , the
(initial) adversary generates a pre-image x 0 = A0 (r ); next, a function h is selected (by
invoking I (1n )); and, finally, the (residual) adversary A is given h (as well as r used

33 This condition is made merely to avoid annoying technicalities. Note that |s| = poly(n) holds by definition of I .

561
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

at the first stage) and tries to find a pre-image x = x0 such that h(x) = h(x0 ). Indeed,
def
Eq. (6.7) refers to the probability that x = A(h, r ) = x0 and yet h(x) = h(x 0 ).
Note that the range specifier (i.e., ) must be super-logarithmic (or else, given s and
x0 ← Un , one is too likely to find an x = x0 such that h s (x) = h s (x0 ), by uniformly
selecting x in {0, 1}n ). Also note that any UOWHF collection yields a collection of
one-way functions (see Exercise 19). Finally, note that any collision-free hashing is
universally one-way hashing, but the converse is false (see Exercise 20). Furthermore,
it is not known whether collision-free hashing can be constructed based on any one-way
functions (in contrast to Theorem 6.4.29, to follow).

6.4.3.2. Constructions
We construct UOWHF collections in several steps, starting with a related but restricted
notion, and relaxing the restriction gradually (until we reach the unrestricted notion
of UOWHF collections). The aforementioned restriction refers to the length of the
arguments to the function. Most importantly, the hardness (to form designated colli-
sions) requirement will refer only to an argument of this length. That is, we refer to the
following technical definition:

Definition 6.4.19 ((d, r )-UOWHFs): Let d, r : N → N. A collection of functions {h s :

{0, 1}d(|s|) → {0, 1}r (|s|) }s∈{0,1}∗ is called (d, r )-UOWHF if there exists a probabilistic
polynomial-time algorithm I so that the following holds:
1. For all sufficiently large n’s and every s in the range of I (1n ), it holds that |s| = n.34
2. There exists a polynomial-time algorithm that, given s and x ∈ {0, 1}d(|s|) , returns
h s (x).
3. For every polynomial q, every deterministic polynomial-time algorithm A0 mapping
q(n)-bit long strings to d(|s|)-bit long strings, every probabilistic polynomial-time
algorithm A, every positive polynomial p, and all sufficiently large n’s, Eq. (6.7)
holds.

Of course, we care only about (d, r )-UOWHF for functions d, r : N → N satisfying

d(n) > r (n). (The case d(n) ≤ r (n) is trivial since collisions can be avoided altogether,
say, by the identity map.) The “minimal” non-trivial case is when d(n) = r (n) + 1.
Indeed, this is our starting point. Furthermore, the construction of such a minimal
(d, d − 1)-UOWHF (undertaken in the following first step) is the most interesting step
to be taken on our entire way toward the construction of full-fledged UOWHF. We start
with an overview of the steps taken along the way.
Step I: Constructing (d, d − 1)-UOWHFs: This construction utilizes a one-way per-
mutation f and a family of hashing functions mapping n-bit long strings to (n − 1)-
bit long strings. A generic function in the constructed collection is obtained by

34 Here we chose to make a more stringent condition, requiring that |s| = n, rather than n ≤ poly(|s|). In fact, one
can easily enforce this more stringent condition by modifying I into I  so that I  (1l(n) ) = I (1n ) for a suitable
function l : N → N satisfying l(n) ≤ poly(n) and n ≤ poly(l(n)).

562

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

composing a hashing function with f ; that is, the resulting function is h ◦ f :

{0, 1}n → {0, 1}n−1 , where h : {0, 1}n → {0, 1}n−1 is a hashing function. Hence, the
constructed functions shrink their input by a single bit.
Intuitively, a random hashing function h maps the f -images in a random manner,
whereas the pre-images under h ◦ f are the f -inverses of the pre-images under h.
Thus, seeking to invert f on y, we may select x0 ∈ {0, 1}n and h at random such that
h( f (x0 )) = h(y), and seek a collision with the designated pre-image x 0 under h ◦ f .
It follows that the ability to form designated collisions can be translated to inverting
f on a random image. Transforming this intuition into an actual proof is the most
technically challenging part of the current section.
Step II: Constructing (d  , d  /2)-UOWHFs: Here we merely compose random func-
tions taken from collections as constructed in Step I. Successively applying d  /2
such functions, we map the d  -bit long pre-image to a d  /2-bit long image.
Intuitively, the ability to form designated collisions with respect to the constructed
collection yields such an ability with respect to (one of) the original collections. (In
the actual argument, we rely on the fact that the definition of (d, d − 1)-UOWHF
refers also to adversaries that get the random coins used for producing the designated
pre-image, and not merely the designated preimage itself.)
Step III: In this step, we construct (length-unrestricted) quasi-UOWHFs that shrink
their input by a factor of two. These functions are constructed by applying a (single)
random function taken from a collection as constructed in Step II to each block of
d  consequtive bits of the pre-image. (Clearly, a collision of the entire sequence of
blocks yields collisions at some block.)
Step IV: Obtaining full-fledged UOWHFs: This construction is analogous to the one
used in Step II. We merely compose random functions taken from a collection as
constructed in Step III. Successively applying t such functions, we essentially map
2t n-bit long pre-images to n-bit long images.
Detailed descriptions of these four steps follow:

Step I: Constructing (d, d − 1)-UOWHFs. We show how to construct length-

restricted UOWHFs that shrink their input by a single bit. Our construction can be
carried out using any one-way permutation. In addition, we use a family of hashing
functions, Snn−1 , as defined in Section 3.5.1.1. Recall that a function selected uni-
formly in Snn−1 maps {0, 1}n to {0, 1}n−1 in a pairwise independent manner, that the
functions in Snn−1 are easy to evaluate, and that for some polynomial p it holds that
log2 |Snn−1 | = p(n).

Construction 6.4.20 (a (d, d − 1)-UOWHF): Let f : {0, 1}∗ → {0, 1}∗ be a 1-1 and
length-preserving function, and let Snn−1 be a family of hashing functions such that
log2 |Snn−1 | = p(n), for some polynomial p. (Specifically, suppose that log2 |Snn−1 | ∈
{3n − 2, 2n}, as in Exercises 22.2 and 23 of Chapter 3.) Then, for every s ∈ Snn−1 ≡
def
{0, 1} p(n) and every x ∈ {0, 1}n , we define h s (x) = h s ( f (x)).
563
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

def
Tedious details: In case |s| ∈ { p(n) : n ∈ N}, we define h s = h s  where s  is the
longest prefix of s satisfying |s  | ∈ { p(n) : n ∈ N}. We refer to an index selection
algorithm that, on input 1m , uniformly selects s ∈ {0, 1}m .

That is, h s : {0, 1}d(|s|) → {0, 1}d(|s|)−1 , where d(m) is the largest integer n satisfying
p(n) ≤ m. Note that d is monotonically non-decreasing, and that for 1-1 p’s, the cor-
responding d is onto (i.e., d( p(n)) = n for every n).
The following analysis uses, in an essential way, an additional property of the afore-
mentioned families of hashing functions; specifically, we assume that given two pre-
image–image pairs, it is easy to uniformly generate a hashing function (in the family)
that is consistent with these two mapping conditions. Furthermore, to facilitate the
analysis, we use a specific family of hashing functions, presented in Exercise 23 of
Chapter 3: Functions in Snn−1 are described by a pair of elements of the finite field
GF(2n ) so that the pair (a, b) describes the function h a,b that maps x ∈ GF(2n ) to the
(n − 1)-bit prefix of the n-bit representation of ax + b, where the arithmetic is of
the field GF(2n ). This specific family satisfies all the additional properties required in
the next proposition (see Exercise 24).

Proposition 6.4.21: Suppose that f is a one-way permutation, and that Snn−1 is a

family of hashing functions (as defined in Section 3.5.1.1) such that log2 |Snn−1 | = 2n.
Furthermore, suppose that Snn−1 satisfies the following two conditions:

C1 All but a negligible fraction of the functions in Snn−1 are 2-to-1.

C2 There exists a probabilistic polynomial-time algorithm that, given y1 , y2 ∈ {0, 1}n
and z 1 , z 2 ∈ {0, 1}n−1 , outputs a uniformly distributed element of {s ∈ Snn−1 :
h s (yi ) = z i ∀i ∈ {1, 2}}.
Then {h s }s∈{0,1}∗ as in Construction 6.4.20 is a (d, d − 1)-UOWHF, for d(m) = m/2.

Proof Sketch: Intuitively, forming designated collisions under h s ≡ h s ◦ f yields the

ability to invert f on a random y, because the collisions are due to h s , which may be
selected such that h s (y) = h s ( f (x 0 )) for any given y and x 0 . We stress that typically
there are only two pre-images of h s (x0 ) under h s , one being x 0 itself (which is given to
the collision-finder) and the other being f −1 (y). Thus, the ability to form a designated
collision with x0 yields an ability to invert f on a random y, by selecting a random s
such that h s (y) = h s (x 0 ), and forming a designated collision under h s . More precisely,
suppose we wish to invert f on a random image y. Then we may invoke a collision-
finder, which first outputs some x0 , supply it with a random s satisfying h s (y) = h s (x 0 ),
and hope that it forms a collision (i.e., finds a different pre-image x satisfying h s (x) =
h s (x0 )). Indeed, typically, the different pre-image must be f −1 (y), which means that
whenever the collision-finder succeeds, we also succeed (i.e., invert f on y). Details
follow.
Evidently, the proof is by a reducibility argument. Suppose that we are given a
probabilistic polynomial-time algorithm A that forms designated collisions under {h s },
with respect to pre-images produced by a deterministic polynomial-time algorithm A0 ,
564

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

which maps p(n)-bit strings to n-bit strings. Then, we construct an algorithm A that
inverts f . On input y = f (x), where n = |y| = |x|, algorithm A proceeds as follows:

(1) Select r0 uniformly in {0, 1} p(n) , and compute x0 = A0 (r0 ) and y0 = f (x0 ).
(2) Select s uniformly in {s ∈ Snn−1 : h s (y0 ) = h s (y)}.
(Recall that y is the input to A, and y0 is generated by A at Step (1).)
(3) Invoke A on input (s, r0 ), and output whatever A does.
By Condition C2, Step (2) can be implemented in probabilistic polynomial-time.
Turning to the analysis of algorithm A, we consider the behavior of A on input
y = f (x) for a uniformly distributed x ∈ {0, 1}n , which implies that y is uniformly
distributed over {0, 1}n . We first observe that for every fixed r0 selected in Step (1), if y
is uniformly distributed in {0, 1}n , then s as determined in Step (2) is almost uniformly
distributed in Snn−1 .

On the distribution of s as selected in Step (2): Fixing r0 ∈ {0, 1}q(n) means that
y0 = f (A0 (r0 )) ∈ {0, 1}n is fixed. Using the pairwise independence property of
def
Snn−1 , it follows that for each y ∈ {0, 1}n \ {y0 }, the cardinality of S y = {s ∈ Snn−1 :
h s (y0 ) = h s (y)} equals |Snn−1 |/2n−1 . Furthermore, in case h s is 2-to-1, the string
s resides in exactly two S y ’s (one being Sy0 ). Recalling that all but a negligible
fraction of the h s ’s are 2-to-1 (i.e., Condition C1), it follows that each such function
is selected with probability 2 · 2−n · (|Snn−1 |/2n−1 )−1 = |Snn−1 |−1 . Other functions
(i.e., non-2-to-1 functions) are selected with negligible probability.

By the construction of A (which ignores y in Step (1)), the probability that f (x0 ) = y is
negligible (but we could have taken advantage of this case, too, by augmenting Step (1)
such that if y0 = y, then A halts with output x0 ). Note that in case f (x 0 ) = y and
h s is 2-to-1, if A returns x  such that x  = x0 and h s (x  ) = h s (x0 ), then it holds that
f (x  ) = y.
def
Justifying the last claim: Let v = h s (y) and suppose that h s is 2-to-1. Then, by
Step (2) and f (x0 ) = y, it holds that x = f −1 (y) and x0 are the two pre-images of
v = h s (x) = h s (x 0 ) under h s , where h s = h s ◦ f is 2-to-1 because f is 1-to-1 and
h s is 2-to-1. Since x  = x0 is also a pre-image of v under h s , it follows that x  = x.

We conclude that if A forms designated collisions with probability ε  (n), then A inverts
f with probability ε  (n) − µ(n), where µ is a negligible function (accounting for the
negligible probability that h s is not 2-to-1). (Indeed, we rely on the fact that s as selected
in Step (2) is distributed almost uniformly, and furthermore that each 2-to-1 function
appears with exectly the right probability.) The proposition follows.

Step II: Constructing (d
, d
/2)-UOWHFs. We now take the second step on our
way, and use any (d, d − 1)-UOWHF in order to construct a (d  , d  /2)-UOWHF. That
is, we construct length-restricted UOWHFs that shrink their input by a factor of 2.
The construction is obtained by composing a sequence of different functions taken
from different (d, d − 1)-UOWHFs. That is, each function in the sequence shrinks the
input by one bit, and the composition of d  /2 functions shrinks the initial d  -bit long

565
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

input by a factor of 2. For simplicity, we assume that the function d : N → N is onto

and monotonically non-decreasing. In such a case we denote by d −1 (m) the smallest
natural number n satisfying d(n) = m (and so d −1 (d(n)) ≤ n).

Construction 6.4.22 (a (d  , d  /2)-UOWHF): Let {h s : {0, 1}d(|s|) → {0, 1}d(|s|)−1 }

s∈{0,1}∗ , where d : N → N is onto and non-decreasing. Then, for every s =
−1
(s1 , ..., s d(n)/2 ), where each si ∈ {0, 1}d (d(n)+1−i) , and every x ∈ {0, 1}d(n) , we define

h s1 ,...,s d(n)/2 (x) = h s d(n)/2 (· · · h s2 (h s1 (x))· · ·)

def

def
That is, letting x0 = x, and xi ← h si (xi−1 ) for i = 1, ..., d(n)/2 , we set h s (x0 ) =
x d(n)/2 . (Note that d(|si |) = d(n) + 1 − i and |xi | = d(n) + 1 − i indeed hold.)
Tedious details: We refer to an index selection algorithm that, on input 1m , deter-
def
mines the largest integer n such that m ≥ m  = i=1 d −1 (d(n) + 1 − i), uni-
d(n)/2

d −1 (d(n)+1−i) 
formly selects s1 , ..., s d(n)/2 such that si ∈ {0, 1} , and s0 ∈ {0, 1}m−m ,
def
and lets h s0 ,s1 ,...,s d(n)/2 = h s1 ,...,s d(n)/2 .

That is, for m = |s|, we have h s : {0, 1}d(n) → {0, 1}d(n)/2 , where n is the largest integer
d −1 (d(n) + 1 − i). Thus, d  (m) = d(n), where n is the length
d(n)/2
such that m ≥ i=1
 
of the index in the (d, d – 1) - UOWHF; that is, we have h s : {0, 1}d (|s|) → {0, 1} d (|s|)/2 ,
with d  (|s|) = d(n). Note that for d(n) = (n) (as in Construction 6.4.20), it holds that
√
d  (O(n 2 )) ≥ d(n) and d  (m) = ( m) follows. More generally, if for some polynomial
p it holds that p(d(n)) ≥ n ≥ d(n) (for all n’s), then for some polynomial p it holds
that p  (d  (m)) ≥ m ≥ d  (m) (for all m’s), because d  (d(n) · n) ≥ d(n). We call such a
function sufficiently growing; that is, d : N → N is sufficiently growing if there exists
a polynomial p so that for every n it holds that p(d(n)) ≥ n. (E.g., for every fixed
ε, ε  > 0, the function d(n) = ε n ε is sufficiently growing.)

Proposition 6.4.23: Suppose that {h s }s∈{0,1}∗ is a (d, d − 1)-UOWHF, where d : N → N

is onto, non-decreasing, and sufficiently growing. Then, for some sufficiently growing
function d  : N → N, Construction 6.4.22 is a (d  , d  /2 )-UOWHF.

Proof Sketch: Intuitively, a designated collision under h s1 ,...,sd/2 yields a desig-
def
nated collision under one of the h si ’s. That is, let x0 = x and xi ← h si (xi−1 ) for
i = 1, ..., d(n)/2 . Then if given x and s = (s1 , ..., sd/2 ), one can find an x  = x such
that h s (x) = h s (x  ); then there exists an i so that xi−1 = xi−1

and xi = h si (xi−1 ) =
  
h si (xi−1 ) = xi , where the x j ’s are defined analogously to the x j ’s. Thus, we obtain
a designated collision under h si . We stress that because h s does not shrink its in-
put too much, the length of si is polynomially related to the length of s (and thus,
forming collisions with respect to h si by using the collision-finder for h s yields a
contradiction).
The actual proof uses the hypothesis that it is hard to form designated collisions
when one is also given the coins used in the generation of the pre-image (and not
merely the pre-image itself). In particular, we construct an algorithm that forms des-
ignated collisions under one of the h si ’s, when given not only xi−1 but also x0 (which
566

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

actually yields xi−1 ). The following details are quite tedious and merely provide an
implementation of this idea.
As stated, the proof is by a reducibility argument. We are given a probabilistic
polynomial-time algorithm A that forms designated collisions under {h s }, with respect
to pre-images produced by a deterministic polynomial-time algorithm A0 that maps
p  (n)-bit strings to n-bit strings. We construct algorithms A0 and A such that A forms
designated collisions under {h s } with respect to pre-images produced by algorithm A0 ,
which maps p(n)-bit strings to n-bit strings, for a suitable polynomial p. (Specifically,
p : N → N is 1-1 and p(n) ≥ p (d −1 (2d(n))) + n + n · d −1 (2d(n)), where the factor
of 2 appearing in the expression is due to the shrinking factor of h s .)
We start with the description of A0 , that is, the algorithm that generates pre-images
of {h s }. Intuitively, A0 selects a random j, uses A0 to obtain a pre-image x 0 of {h s },
generates random s0 , ..., s j−1 , and outputs a pre-image x j−1 of {h s j }, computed by
xi = h si (xi−1 ) for i = 1, ..., j − 1. (Algorithm A will be given x j−1 (or rather the coins
used to generate x j−1 ) and a random h s j and will try to form a collision with x j−1 under
h s j .)

Detailed description of A0 : Recall that p is a polynomial, d(n) ≤ n and d −1 (n) =

def def
poly(n). Let p(n) = n + n · q(n) + p (q(n)), where q(n) = d −1 (2d(n)). On input
r ∈ {0, 1} p(n) , algorithm A0 proceeds as follows:

(1) Write r = r1 r2 r3 such that |r1 | = n, |r2 | = n · q(n), and |r3 | = p  (q(n)).
Using r1 , determine m in {n + 1, ..., n · q(n)} and j ∈ {1, ..., q(n)} such that
both m and j are almost uniformly distributed in the corresponding sets.
d(n  )/2
(2) Compute the largest integer n  such that m ≤ i=1 d −1 (d(n  ) + 1 − i).
−1 
(3) If d (d(n ) + 1 − j) = n, then output the d(n)-bit long suffix of r3 .
(Comment: the output in this case is immaterial to our proof.)
(4) Otherwise (i.e., n = d −1 (d(n  ) + 1 − j), which is the case we care about), do:
(4.1) Let s0 s1 · · · s j−1 be a prefix of r2 such that
d(n  )/2
|s0 | = m − i=1 d −1 (d(n  ) + 1 − i),
−1 
and |si | = d (d(n ) + 1 − i), for i = 1, ..., j − 1.
(4.2) Let x0 ← A0 (r  ), where r  is the p (d −1 (d(n  )))-bit long suffix of r3 .

(Comment: x0 ∈ {0, 1}d(n ) .)
(4.3) For i = 1, ..., j − 1, compute xi ← h si (xi−1 ).
Output x j−1 ∈ {0, 1}d(n) .
(Note that d(n) = d(n  ) − ( j − 1).)

As stated previously, we only care about the case in which Step (4) is applied.
This case occurs with noticeable probability, and the description of the following
algorithm A refers to it.

Algorithm A will be given x j−1 as produced by A0 (along with, or actually only, the
coins used in its generation), as well as a random h s j , and will try to form a collision with
x j−1 under h s j . On input s ∈ {0, 1}n (viewed as s j ) and the coins given to A0 , algorithm
A operates as follows. First, A selects j and s0 , s1 , ..., s j−1 exactly as A0 does (which is
the reason that A needs the coins used by A0 ). Next, A tries to obtain a collision under
h s by invoking A (r  , s  ), where r  is the sequence of coins that A0 handed to A0 and
567
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

s  = (s0 , s1 , ..., s j−1 , s, s j+1 , ..., sd(n)/2 ), where s j+1 , ..., sd(n)/2 are uniformly selected
by A. Finally, A outputs h s j−1 (· · · (h s1 (A (r  , s  ))· · ·).
Detailed description of A: On input s ∈ {0, 1}n and r ∈ {0, 1} p(n) , algorithm A
proceeds as follows.

(1–2) Using r , determine m, j, and n  exactly as done by A0 .

(3) If d −1 (d(n  ) + 1 − j) = n, then abort.
(4) Otherwise (i.e., n = d −1 (d(n  ) + 1 − j)), do:
(4.1) Determine s0 , s1 , ..., s j−1 and r  exactly as A0 does (at its Step (4)).
−1 
(4.2) Uniformly select s j+1 , ..., s d(n )/2 such that si ∈ {0, 1}d (d(n )+1−i) ,
and set s  = (s0 , s1 , ..., s j−1 , s, s j+1 , ..., s d(n )/2 ).
(4.3) Invoke A on input (s  , r  ), and obtain x0 ← A (s  , r  ).

(Comment: x0 ∈ {0, 1}d(n ) .)
(4.4) For i = 1, ..., j − 1, compute xi ← h si (xi−1 
).

Output x j−1 ∈ {0, 1} .d(n)

Clearly, if algorithms A and A0 run in polynomial-time, then so do A and A 0 (and if p

is a polynomial then so is p). We now lower-bound the probability that A succeeds in
forming designated collisions under {h s }, with respect to pre-images produced by A0 .
We start from the contradiction hypothesis by which the corresponding probability for
A (with respect to A0 ) is non-negligible.
Let use denote by ε (m) the success probability of A on uniformly distributed

input (s  , r  ) ∈ {0, 1}m × {0, 1} p (m) . Let n  be the largest integer so that m ≤

d −1 (d(n  ) + 1 − i). Then, there exists a j ∈ {1, ..., d(n  )} such that, with
d(n )/2
i=1
probability at least ε (m)/d  (n  ), on input (s  , r  ), where s  = s0 , s1 , ..., s d(n  )/2 is as in
def
Construction 6.4.22, A outputs an x  = x = A0 (r  ) such that h s j−1 (· · · (h s1 (x  ))· · ·) =
h s j−1 (· · · (h s1 (x))· · ·) and h s j (· · · (h s1 (x  ))· · ·) = h s j (· · · (h s1 (x))· · ·). Fixing these m,
j, and n  , let n = d −1 (d(n  ) + 1 − j), and consider what happens when A is in-
voked on uniformly distributed (s, r ) ∈ {0, 1}n × {0, 1} p(n) . With probability at least
def
δ(n) = 1/(nq(n))2 over the possible r ’s, the values of m and j are determined to equal
the aforementioned desired values. Conditioned on this case, A  is invoked on uni-

formly distributed input (s  , r  ) ∈ {0, 1}m × {0, 1} p (m) , and so a collision at the j-th
hashing function occurs with probability at least ε (m)/d  (n  ). Note that m = poly(n),
δ(n) ≥ 1/poly(n) and d  (n  ) = poly(n). This implies that A succeeds with probability at
 
least ε(n) = δ(n) · dε (m) = ε poly(n)
def (poly(n))
(n  )
, with respect to pre-images produced by A0 . Thus,
if ε is non-negligible, then so is ε, and the proposition follows.

Step III: Constructing (Length-Unrestricted) Quasi-UOWHFs That Shrink Their

Input by a Factor of Two. The third step on our way consists of using any (d, d/2)-
UOWHF in order to construct “quasi UOWHFs” that are applicable to any input length
but shrink each input to half its length (rather than to a fixed length that only depends on
the function description). The resulting construct does not fit Definition 6.4.19, because
the function’s output length depends on the function’s input length, yet the function can
be applied to any input length (rather than only to a single length determined by the
568

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

function’s description). Thus, the resulting construct yields a (d  , d  /2)-UOWHF for any
polynomially bounded function d  (e.g., d  (n) = n 2 ), whereas in Construction 6.4.22,
the function d  is fixed and satisfies d  (n)  n. The construction itself amounts to
parsing the input into blocks and applying the same function (taken from a (d, d/2)-
UOWHF) to each block.

Construction 6.4.24 (a (d  , d  /2)-UOWHF for any d  ): Let {h s : {0, 1}d(|s|) →

{0, 1}d(|s|)/2 }s∈{0,1}∗ , where d : N → N is onto and non-decreasing. Then, for every
s ∈ {0, 1}n and every x ∈ {0, 1}∗ , we define

h s (x) = h s (x1 )· · · h s (xt 10d(n)−|xt |−1 )

def

where x = x 1 · · · x t , 0 ≤ |xt | < d(n) and |xi | = d(n) for i = 1, ..., t − 1. The index-
selection algorithm of {h s } is identical to the one of {h s }.

Clearly, |h s (x)| = (|x| + 1)/d(n) · d(n)/2, which is approximately |x|/2 (pro-

vided |x| ≥ d(n)). Furthermore, Construction 6.4.24 satisfies Conditions 1 and 2 of
Definition 6.4.18, provided that {h s } satisfies the corresponding conditions of Def-
inition 6.4.19. We thus focus on the hardness to form designated collisions (i.e.,
Condition 3).

Proposition 6.4.25: Suppose that {h s }s∈{0,1}∗ is a (d, d/2)-UOWHF, where d : N → N

is onto, non-decreasing, and sufficiently growing. Then Construction 6.4.22 satisfies
Condition 3 of Definition 6.4.18.

Proof Sketch: Intuitively, a designated collision under h s yields a designated collision

under h s . That is, consider the parsing of each string into blocks of length d(n), as in
Construction 6.4.24. Now if, given x = x1 · · · x t and s, one can find an x  = x1 · · · xt =
x such that h s (x) = h s (x  ), then t  = t and there exists an i such that xi = xi and
h s (xi ) = h s (xi ). Details follow.
The actual proof is by a reducibility argument. Given a probabilistic polynomial-time
algorithm A that forms designated collisions under {h s }, with respect to pre-images
produced by a polynomial-time algorithm A0 , we construct algorithms A0 and A such
that A forms designated collisions under {h s } with respect to pre-images produced
by algorithm A0 . Specifically, algorithm A0 invokes A0 and uses extra randomness
(supplied in its input) to uniformly select one of the d(n)-bit long blocks in the standard
parsing of the output of A0 . That is, the random-tape used by algorithm A0 has the form
(r  , i), and A0 outputs the i-th block in the parsing of the string A0 (r  ). Algorithm A
is derived analogously. That is, given s ∈ {0, 1}n and the coins r = (r  , i) used by A0 ,
algorithm A invokes A on input s and r  , obtains the output x  , and outputs the i-th
block in the standard parsing of x  .
Note that whenever we have a collision under h s (i.e., a pair x = x  such that h s (x) =
h s (x  )), we obtain at least one collision under the corresponding h s (i.e., for some i, the


i-th blocks of x and x  differ, and yet both blocks are mapped by h s to the same image).
Thus, if algorithm A succeeds (in forming designated collisions with respect to {h s })
569
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

with probability ε  (n), then algorithm A succeeds (in forming designated collisions
with respect to {h s }) with probability at least ε (n)/t(n), where t(n) is a bound on the
running time of A (which also upper-bounds the length of the output of A , and so
1/t(n) is a lower bound on the probability that the colliding strings differ at a certain
uniformly selected block). The proposition follows.

Step IV: Obtaining Full-Fledged UOWHFs. The last step on our way consists of
using any quasi-UOWHFs as constructed (in Step III) to obtain full-fledged UOWHFs.
That is, we use quasi-UOWHFs that are applicable to any input length but shrink each
input to half its length (rather than to a fixed length that only depends on the function
description). The resulting construct is a UOWHF (as defined in Definition 6.4.18).
The construction is obtained by composing a sequence of different functions (each
taken from the same quasi-UOWHF); that is, the following construction is analogous
to Construction 6.4.22.

Construction 6.4.26 (a UOWHF): Let {h s : {0, 1}∗ → {0, 1}∗ }s∈{0,1}∗ , such that
|h s (x)| = |x|/2, for every x ∈ {0, 1}2i·|s| where i ∈ N. Then, for every s1 , ..., sn ∈ {0, 1}n
and every t ∈ N and x ∈ {0, 1}2 ·n , we define
t

h s1 ,...,sn (x) = (t, h st (· · · h s2 (h s1 (x))· · ·))

def

def
That is, we let x 0 = x, and xi ← h si (xi−1 ), for i = 1, ..., t.

Tedious details: Strings of lengths that are not of the form 2t · n are padded into
strings of such form in a standard manner. We refer to an index-selection algorithm
√
that, on input 1m , determines n =  m, uniformly selects s1 , ..., sn ∈ {0, 1}n and
def
s0 ∈ {0, 1}m−n , and lets h s0 ,s1 ,...,sn = h s1 ,...,sn .
2

Observe that h s0 ,s1 ,...,sn (x) = h s0 ,s1 ,...,sn (x  ) implies that both equal the pair (t, h st (· · · h s2
(h s1 (x))· · ·)), where t = log2 (|x|/n) = log2 (|x  |/n) . Note that h s0 ,s1 ,...,sn :
{0, 1}∗ → {0, 1}n+log2 n , and that m = |s0 , s1 , ..., sn | < (n + 1)2 .

Proposition 6.4.27: Suppose that {h s }s∈{0,1}∗ satisfies the conditions of Defini-

tion 6.4.18, except that it maps arbitrary input strings to outputs having half the length
(rather than a length determined by |s|). Then Construction 6.4.26 constitutes a col-
lection of UOWHFs.

The proof of Proposition 6.4.27 is omitted because it is almost identical to the proof of
Proposition 6.4.23.

Conclusion. Combining the previous four steps, we obtain a construction of (full-

fledged) UOWHFs (based on any one-way permutation). That is, combining Proposi-
tions 6.4.21, 6.4.23, 6.4.25, and 6.4.27, we obtain:35

35 Actually, there is a minor gap between Constructions 6.4.24 and 6.4.26. In the former we constructed functions
that hash every x into a value of length (|x| + 1)/d(n) · d(n)/2, whereas in the latter we used functions that
hash every x ∈ {0, 1}2i·n into a value of length i · n.

570

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

Theorem 6.4.28: If one-way permutations exist, then universal one-way hash functions
exist.

Note that the only barrier toward constructing UOWHFs based on arbitrary one-way
functions is Proposition 6.4.21, which refers to one-way permutations. Thus, if we
wish to constructs UOWHF based on any one-way function, then we need to present
an alternative construction of a (d, d − 1)-UOWHF (i.e., an alternative to Construc-
tion 6.4.20, which fails in case f is 2-to-1).36 Such a construction is actually known,
and so the following result is known to hold (but its proof it too complex to fit in this
work):

Theorem 6.4.29: Universal one-way hash functions exist if and only if one-way func-
tions exist.

We stress that the difficult direction is the one referred to earlier (i.e., from one-
way functions to UOWHF collections). For the much easier (converse) direction, see
Exercise 19.

6.4.3.3. One-Time Signature Schemes Based on UOWHF

Using universal one-way hash functions, we present an alternative construction of
one-time signature schemes based on length-restricted one-time signature schemes.
Specifically, we replace the hash-and-sign paradigm (i.e., Construction 6.2.6) in which
collision-free hashing functions were used by the following variant (i.e., Construc-
tion 6.4.30) in which universal one-way hash functions are used instead. The difference
between the two constructions is that here, the (description of the) hashing function is
not a part of the signing and verification keys, but is rather selected on the fly by the
signing algorithm (and appears as part of the signature). Furthermore, the description
of the hash function is being authenticated (by the signer) together with the hash value.
It follows that the forging adversary, which is unable to break the length-restricted one-
time signature scheme, must form a designated collision (rather than an arbitrary one).
However, the latter is infeasible, too (by virtue of the UOWHF collection in use). We
comment that the same (new) construction is applicable to length-restricted signature
schemes (rather than to one-time ones): We stress that in the latter case, a new hashing
function is selected at random each time the signing algorithm is applied. In fact, we
present the more general construction.

Construction 6.4.30 (the Hash-and-Sign Paradigm, Revisited): Let ,  : N → N such

that (n) =  (n) + n. Let (G, S, V ) be an -restricted signature scheme as in Defini-

tion 6.2.1, and {hr : {0, 1}∗ → {0, 1} (|r |) }r ∈{0,1}∗ be a collection of functions with an
indexing algorithm I (as in Definition 6.4.18). We construct a general signature scheme,

36 For example, if f (σ, x  ) = (0, f  (x  )), for σ ∈ {0, 1}, then forming designated collisions under Construc-
tion 6.4.20 is easy: Given (0, x  ), one outputs (1, x  ), and indeed a collision is formed (already under f ).

571
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

(G  , S  , V  ), with G  identical to G, as follows:

Signing with S  : On input a signing-key s (in the range of G 1 (1n )) and a document
α ∈ {0, 1}∗ , algorithm S  proceeds in two steps:
1. Algorithm S  invokes I to obtain β1 ← I (1n ).
2. Algorithm S  invokes S to produce β2 ← Ss (β1 , h β1 (α)).
Algorithm S  outputs the signature (β1 , β2 ).
Verification with V  : On input a verifying-key v, a document α ∈ {0, 1}∗ , and an al-
leged signature (β1 , β2 ), algorithm V  invokes V , and outputs Vv ((β1 , h β1 (α)), β2 ).
Recall that secure -restricted one-time signature schemes exist for any polynomial ,
provided that one-way functions exist. Thus, the fact that Construction 6.4.30 requires
(n) > n is not a problem. In applying Construction 6.4.30, one should first choose a

family of UOWHFs {h r : {0, 1}∗ → {0, 1} (|r |) }r ∈{0,1}∗ , then determine (n) =  (n) +
n, and use a corresponding secure -restricted one-time signature scheme.
Let us pause to compare Construction 6.2.6 with Construction 6.4.30. Recall that in
Construction 6.2.6, the function description β1 ← I (1n ) is produced (and fixed as part
of both keys) by the key-generation algorithm. Thus, the function description β1 is triv-
ially authenticated (i.e., by merely being part of the verification-key). Consequently, in
Construction 6.2.6, the S  -signature (of α) equals Ss (h β1 (α)). In contrast, in Construc-
tion 6.4.30 a fresh new (function description) β1 is selected per each signature, and thus
β1 needs to be authenticated. Hence, the S  -signature equals the pair (β1 , Ss (β1 , h β1 (α))).
Since we want to be able to use (length-restricted) one-time signatures, we let the sign-
ing algorithm authenticate both β1 and h β1 (α) via a single signature. (Alternatively,
we could have used two instances of the one-time signature scheme (G, S, V ), one for
signing the function description β1 and the other for signing the hash value h β1 (α).)

Proposition 6.4.31: Suppose that (G, S, V ) is a secure -restricted signature scheme

and that {hr : {0, 1}∗ → {0, 1}(|r |)−|r | }r ∈{0,1}∗ is a collection of UOWHFs. Then
(G  , S  , V  ), as defined in Construction 6.4.30, is a secure (full-fledged) signature
scheme. Furthermore, if (G, S, V ) is only a secure -restricted one-time signature
scheme, then (G  , S  , V  ) is a secure one-time signature scheme.

Proof Sketch: The proof follows the underlying principles of the proof of Proposi-
tion 6.2.7. That is, forgery with respect to (G  , S  , V  ) yields either forgery with respect
to (G, S, V ) or a collision under the hash function, where in the latter case, a desig-
nated collision is formed (in contradiction to the hypothesis regarding the UOWHF).
For the furthermore-part, the observation underlying the proof of Proposition 6.4.7
still holds (i.e., the number of queries made by the forger constructed for (G, S, V )
equals the number of queries made by the forger assumed (toward the contradiction)
for (G  , S  , V  )). Details follow.
Given an adversary A attacking the complex scheme (G  , S  , V  ), we construct an
adversary A that attacks the -restricted scheme, (G, S, V ). The adversary A uses I (the
indexing algorithm of the UOWHF collection) and its oracle Ss in order to emulate the
oracle Ss for A . This is done in a straightforward manner; that is, algorithm A emulates
Ss by using the oracle Ss (exactly as Ss actually does). Specifically, to answer a query
572

www.Ebook777.com
 6.4 CONSTRUCTIONS OF SIGNATURE SCHEMES

q, algorithm A generates a1 ← I (1n ), forwards (a1 , h a1 (q)) to its own oracle (i.e., Ss ),
and answers with (a1 , a2 ), where a2 = Ss (a1 , h a1 (q)). (We stress that A issues a single
Ss -query per each Ss -query made by A .) When A outputs a document-signature pair
relative to the complex scheme (G  , S  , V  ), algorithm A tries to use this pair in order
to form a document-signature pair relative to the -restricted scheme, (G, S, V ). That
is, if A outputs the document-signature pair (α, β), where β = (β1 , β2 ), then A will
def
output the document-signature pair (α2 , β2 ), where α2 = (β1 , h β1 (α)).

Assume that with (non-negligible) probability ε (n), the (probabilistic polynomial-
time) algorithm A succeeds in existentially forging relative to the complex scheme
(G  , S  , V  ). Let (α (i) , β (i) ) denote the i-th query and answer pair made by A , and let
(α, β) be the forged document-signature pair that A outputs (in case of success), where
(i) (i)
β (i) = (β1 , β2 ) and β = (β1 , β2 ). We consider the following two cases regarding the
forging event:
(i)
Case 1: (β1 , h β1 (α)) = (β1 , h β (i) (α (i) )) for all i’s. (That is, the Ss -signed value in the
1
forged signature (i.e., the value (β1 , h β1 (α))) is different from all queries made to
Ss .) In this case, the document-signature pair ((β1 , h β1 (α)), β2 ) constitutes a success
in existential forgery relative to the -restricted scheme (G, S, V ).
(i)
Case 2: (β1 , h β1 (α)) = (β1 , h β (i) (α (i) )) for some i. (That is, the Ss -signed value used
1
in the forged signature equals the i-th query made to Ss , although α = α (i) .) Thus,
(i)
β1 = β1 and h β1 (α) = h β (i) (α (i) ), although α = α (i) . In this case, the pair (α, α (i) )
1
forms a designated collision under h β (i) (and we do not obtain success in existential
1
forgery relative to the -restricted scheme). We stress that A selects α (i) before it
is given the description of the function h β (i) , and thus its ability to later produce
1
α = α (i) such that h β1 (α) = h β (i) (α (i) ) yields a violation of the UOWHF property.
1

Thus, if Case 1 occurs with probability at least ε (n)/2, then A succeeds in its attack
on (G, S, V ) with probability at least ε (n)/2, which contradicts the security of the
-restricted scheme (G, S, V ). On the other hand, if Case 2 occurs with probability
at least ε  (n)/2, then we derive a contradiction to the difficulty of forming designated
collisions with respect to {h r }. Details regarding Case 2 follow.
We start with a sketch of the construction of an algorithm that attempts to form
designated collisions under a randomly selected hash function. Loosely speaking, we
construct an algorithm B  that tries to form designated collisions by emulating the
attack of A on a random instance of (G  , S  , V  ) that B  selects by itself. Thus, B  can
easily answer any signing-query referred to it by A , but in one of these queries (the
index of which is selected at random by B  ), algorithm B  will use a hash function given
to it from the outside (rather than generating such a function at random by itself). In
case A forges a signature while using this specific function-value pair (as in Case 2),
algorithm B  obtains and outputs a designated collision.
We now turn to the actual construction of algorithm B  (which attempts to form
designated collisions under a randomly selected hash function). Recall that such an
algorithm operates in three stages (see discussion in Section 6.4.3.1): First the algorithm
selects a pre-image x 0 , next it is given a description of a function h, and finally it is
573
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

required to output x = x0 such that h(x) = h(x0 ). We stress that the third stage in the
attack is also given the random coins used for producing the pre-image x0 (at the first
stage). Now, on input 1n , algorithm B  proceeds in three stages:

Stage 1: Algorithm B  selects uniformly i ∈ {1, ..., t(n)}, where t(n) bounds the run-
ning time of A (G 1 (1n )) (and thus the number of queries it makes). Next, B  se-
lects (s, v) ← G  (1n ) and emulates the attack of A (v) on Ss , while answering
the queries of Ss as follows. All queries except the i-th one are emulated in the
straightforward manner (i.e., by executing the program of Ss as stated). That is, for
( j)
j = i, the j-th query, denoted α ( j) , is answered by producing β1 ← I (1n ), com-
( j) ( j)
puting β2 ← Ss (β1 , h β ( j) (α ( j) )) (using the knowledge of s), and answering with
1
the pair (β1 , β2 ). The i-th query of A , denoted α (i) , will be used as the designated
( j) ( j)

pre-image. Once α (i) is issued (by A ), algorithm B  completes its first stage (without
answering this query), and the rest of the emulation of A will be conducted by the
third stage of B  .
Stage 2: At this point (i.e., after B  has selected the designated pre-image α (i) ), B 
obtains a description of a random hashing function h r (thus completing its second
operation stage). That is, this stage consists of B  being given r ← I (1n ).
Stage 3: Next, algorithm B  answers the i-th query (i.e., α (i) ) by applying Ss to the
pair (r, h r (α (i) )). Subsequent queries are emulated in the straightforward manner (as
in Stage 1). When A halts, B  checks whether A has output a valid document-
( j)
signature pair (α, β) as in Case 2 (i.e., β1 = β1 and h β1 (α) = h β ( j) (α ( j) ) for some
1
j), and whether the collision formed is indeed on the i-th query (i.e., j = i, which
means that h r (α) = h r (α (i) )). When this happens, B  outputs α (which is different
than α (i) ), and in doing so it has succeeded in forming a designated collision (with
α (i) under h r ).

Now, if Case 2 occurs with probability at least ε 2(n) (and A makes at most t(n) queries),
then B  has succeeded in forming a designated collision with probability at least

1
t(n)
· ε 2(n) , because the actions of A are independent of the random value of i. This
contradicts the hypothesis that {h r } is UOWHF.
As mentioned earlier, the furthermore-part of the proposition follows by observing
that if the forging algorithm A makes at most one query, then the same holds for the
algorithm A constructed in the beginning of the proof. Thus, if (G  , S  , V  ) can be
broken via a single-message attack, then either (G, S, V ) can be broken via a single-
message attack or one can form designated collisions (with respect to {h r }). In both
cases, we reach a contradiction.

Conclusion. Combining the furthermore-part of Proposition 6.4.31, Corollary 6.4.6,

and the fact that UOWHF collections imply one-way functions (see Exercise 19), we
obtain:

Theorem 6.4.32: If there exist universal one-way hash functions, then secure one-time
signature schemes exist, too.

574

www.Ebook777.com
 6.5* SOME ADDITIONAL PROPERTIES

6.4.3.4. Conclusions and Comments

Combining Theorems 6.4.28, 6.4.32, and 6.4.9, we obtain:

Corollary 6.4.33: If one-way permutations exists, then there exist secure signature
schemes.

Like Corollary 6.4.10, Corollary 6.4.33 asserts the existence of secure (public-key) sig-
nature schemes, based on an assumption that does not mention trapdoors. Furthermore,
the assumption made in Corollary 6.4.33 seems weaker than the one made in Corol-
lary 6.4.10. We can further weaken the assumption by using Theorem 6.4.29 (which
was stated without a proof), rather than Theorem 6.4.28. Specifically, combining The-
orems 6.4.29, 6.4.32, and 6.4.9, we establish Theorem 6.4.1. That is, secure signature
schemes exist if and only if one-way functions exist. Furthermore, as in the case of MACs
(see Theorem 6.3.8), the resulting signature schemes have signatures of fixed length.

Comment: The Hash-and-Sign Paradigm, Revisited. We wish to highlight the re-

vised version of the hash-and-sign paradigm that underlies Construction 6.4.30. Similar
to the original instantiation of the hash-and-sign paradigm (i.e., Construction 6.2.6),
Construction 6.4.30 is useful in practice. We warn that using the latter construction
requires verifying that {h r } is a UOWHF (rather than collision-free). The advantage of
Construction 6.4.30 over Construction 6.2.6 is that the former relies on a seemingly
weaker construct; that is, hardness of forming designated collisions (as in UOWHF) is a
seemingly weaker condition than hardness of forming any collision (as in collision-free
hashing). On the other hand, Construction 6.2.6 is simpler and more efficient (e.g., one
need not generate a new hashing function per each signature).

6.5.* Some Additional Properties

We briefly discuss several properties of interest that some signature schemes enjoy.
We first discuss properties that seem unrelated to the original purpose of signature
schemes but are useful toward utilizing a signature scheme as a building block toward
constructing other primitives (e.g., see Section 5.4.4.4). These (related) properties are
having unique valid signatures and being super-secure, where the latter term indi-
cates the infeasibility of finding a different signature even to a document for which a
signature was obtained during the attack. We next turn to properties that offer some
advantages in the originally intended applications of signature schemes. Specifically,
we consider properties that allow for speeding-up the response-time in some settings
(see Sections 6.5.3 and 6.5.4), and a property supporting legitimate revoking of forged
signatures (see Section 6.5.5).

6.5.1. Unique Signatures

Loosely speaking, we say that a signature scheme (G, S, V ) (either a private-key or
a public-key one) has unique signatures if for every possible verification-key v and
every document α there is a unique β such that Vv (α, β) = 1.
575
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

Note that this property is related, but not equivalent, to the question of whether or
not the signing algorithm is deterministic (which is considered in Exercise 1). Indeed,
if the signing algorithm is deterministic, then for every key pair (s, v) and document α,
the result of applying Ss to α is unique (and indeed Vv (α, Ss (α)) = 1). Still, this does
not mean that there is no other β (which is never produced by applying Ss to α) such
that Vv (α, β) = 1. On the other hand, the unique signature property may hold even in
case the signing algorithm is randomized, but (as mentioned earlier) this randomization
can be eliminated anyhow.

Can Secure Signature Schemes Have Unique Signatures? The answer is definitely
affirmative, and in fact we have seen several such schemes in the previous sections.
Specifically, all private-key signature schemes presented in Section 6.3 have unique sig-
natures. Furthermore, every secure private-key signature scheme can be transformed
into one having unique signatures (e.g., by combining deterministic signing as in
Exercise 1 with canonical verification as in Exercise 2). Turning to public-key signature
schemes, we observe that if the one-way function f used in Construction 6.4.4 is 1-1,
then the resulting secure length-restricted one-time (public-key) signature scheme has
unique signatures (because each f -image has a unique pre-image). In addition, Con-
struction 6.2.6 (i.e., the basic hash-and-sign paradigm) preserves the unique signature
property. Let use summarize all these observations:

Theorem 6.5.1 (Secure Schemes with Unique Signatures):

1. Assuming the existence of one-way functions, there exist secure message authenti-
cation schemes having the unique signature property.
2. Assuming the existence of 1-1 one-way functions, there exist secure length-restricted
one-time (public-key) signature schemes having the unique signature property.
3. Assuming the existence of 1-1 one-way functions and collision-free hashing collec-
tions, there exist secure one-time (public-key) signature schemes having the unique
signature property.

In addition, it is known that secure (full-fledged) signature schemes having the unique
signature property can be constructed based on a mild variant on the standard RSA
assumption (see reference in Section 6.6.5). Still, this leaves open the question of
whether or not secure signature schemes having the unique signature property exist if
and only if secure signature schemes exist.

6.5.2. Super-Secure Signature Schemes

In case the signature scheme does not possess the unique signature property, it makes
sense to ask whether, given a message-signature pair, it is feasible to produce a different
signature to the same message. More generally, we may ask whether it is feasible for
a chosen message attack to produce a different signature to any of the messages to
which it has obtained signatures. Such ability may be of concern in some applications
(but, indeed, not in the most natural applications). Combining the new concern with the
standard notion of security, we derive the following notion, which we call super-security.
576

www.Ebook777.com
 6.5* SOME ADDITIONAL PROPERTIES

A signature scheme is called super-secure if it is infeasible for a chosen message attack

to produce a valid message-signature pair that is different from all query-answer pairs
obtained during the attack, regardless of whether or not the message used in the new
pair equals one of the previous queries. (Recall that ordinary security only requires the
infeasibility of producing a valid message-signature pair such that the message part is
different from all queries made during the attack.)

Do Super-Secure Signature Schemes Exist? Indeed, every secure signature scheme

that has unique signatures is super-secure, but the question is whether super-security
may hold for a signature scheme that does not possess the unique signature property.
We answer this question affirmatively.

Theorem 6.5.2 (super-secure signature schemes): Assuming the existence of one-way

functions, there exist super-secure (public-key) signature schemes.

In other words, super-secure signature schemes exist if and only if secure signature
schemes exist. We comment that the signature scheme constructed in the following
proof does not have the unique signature property.

Proof: Starting from (Part 2 of) Theorem 6.5.1, we can use any 1-1 one-way func-
tion to obtain super-secure length-restricted one-time signature schemes. However,
wishing to use arbitrary one-way functions, we will first show that universal one-way
hashing functions can be used (instead of 1-1 one-way functions) in order to obtain
super-secure length-restricted one-time signature schemes. Next, we will show that
super-security is preserved by two transformations presented in Section 6.4: specifi-
cally, the transformation of length-restricted one-time signature schemes into one-time
signature schemes (i.e., Construction 6.4.30), and the transformation of the latter to
(full-fledged) signature schemes (i.e., Construction 6.4.16). Applying these transfor-
mations (to the first scheme), we obtain the desired super-secure signature scheme.
Recall that Construction 6.4.30 also uses universal one-way hashing functions, but the
latter can be constructed using any one-way function (cf. Theorem 6.4.29).37
Claim 6.5.2.1: If there exist universal one-way hashing functions, then for every
polynomially-bounded  : N → N, there exist super-secure -restricted one-time sig-
nature schemes.
Proof Sketch: We modify Construction 6.4.4 by using universal one-way hashing func-
tions (UOWHFs) instead of one-way functions. Specifically, for each pre-image placed
in the signing-key, we select at random and independently a UOWHF, and place its de-
scription both in the signing- and verification-keys. That is, on input 1n , we uniformly
select s10 , s11 , ..., s(n)
0 1
, s(n) ∈ {0, 1}n and UOWHFs h 01 , h 11 , ..., h 0(n) , h 1(n) , and compute
j j j
vi = h i (si ), for i = 1, ..., (n) and j = 0, 1. We let s = ((s10 , s11 ), ..., (s(n)
0 1
, s(n) )),

37 We comment that a simpler proof suffices in case we are willing to use a one-way permutation (rather than
an arbitrary one-way function). In this case, we can start from (Part 2 of ) Theorem 6.5.1 (rather than prove
Claim 6.5.2.1), and use Theorem 6.4.28 (rather than Theorem 6.4.29, which has a more complicated proof ).

577
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

h = ((h 01 , h 11 ), ..., (h 0(n) , h 1(n) )), and v = ((v10 , v11 ), ..., (v(n)
0
, v(n)
1
)), and output the key-
pair (s, v) = ((h, s), (h, v)) (or, actually, we may set (s, v) = (s, (h, v))). Signing and
verification are modified accordingly; that is, the sequence (β1 , ..., β ) is accepted as
a valid signature of the string σ1 · · · σ (with respect to the verification-key v) if and
only if h σi i (βi ) = viσi for every i. In order to show that the resulting scheme is super-
secure under a chosen one-message attack, we adapt the proof of Proposition 6.4.5.
Specifically, fixing such an attacker A, we consider the event in which A violated the
super-security of the scheme. There are two cases to consider:

1. The valid signature formed by A is to the same document for which A has obtained a
different signature (via its single query). In this case, for at least one of the UOWHFs
contained in the verification-key, we obtain a pre-image (of the image also contained
in the verification-key) that is different from the one contained in the signing-key.
Adapting the construction presented in the proof of Proposition 6.4.5, we derive (in
this case) an ability to form designated collisions (in contradiction to the UOWHF
property). We stress that the pre-images contained in the signing-key are selected
independently of the description of the UOWHFs (because both are selected inde-
pendently by the key-generation process). In fact, we obtain a designated collision
for a uniformly selected pre-image.
2. The valid signature formed by A is to a document that is different from the one
for which A has obtained a signature (via its single query). In this case, the proof
of Proposition 6.4.5 yields the ability to invert a randomly selected UOWHF (on
a randomly selected image), which contradicts the UOWHF property (as shown in
Exercise 19).
Thus, in both cases we derive a contradiction, and the claim follows.
Claim 6.5.2.2: When applying the revised hash-and-sign construction (i.e., Construc-
tion 6.4.30) to a super-secure length-restricted signature scheme, the result is a super-
secure signature scheme. In case the length-restricted scheme is only super-secure un-
der a chosen one-message attack, the same holds for the resulting (length-unrestricted)
scheme.
Proof Sketch: We follow the proof of Proposition 6.4.31, and use the same construc-
tion of a forger for the length-restricted scheme (based on the forger for the complex
scheme). Furthermore, we consider the two forgery cases analyzed in the proof of
Proposition 6.4.31:38
(i)
Case 1: (β1 , h β1 (α)) = (β1 , h β (i) (α (i) )) for all i’s. In this case, the analysis is exactly
1
as in the original proof. Note that it does not matter whether or not α = α (i) , since
in both subcases we obtain a valid signature for a new string with respect to the

38 Recall that (α, β) denotes the document-signature pair output by the original forger (i.e., for the complex scheme),
whereas (α (i) , β (i) ) denotes the i-th query-answer pair (to that scheme). The document-signature pair that we
def
output (as a candidate forgery with respect to a length-restricted scheme) is (α2 , β2 ), where α2 = (β1 , h β1 (α))
and β = (β1 , β2 ). Recall that a generic valid document-signature for the complex scheme has the form (α  , β  ),
where β  = (β1 , β2 ) satisfies Vv ((β1 , h β  (α  )), β2 ) = 1.
1

578

www.Ebook777.com
 6.5* SOME ADDITIONAL PROPERTIES

length-restricted signature scheme. Thus, in this case, we derive a violation of the

(ordinary) security of the length-restricted scheme.
(i)
Case 2: (β1 , h β1 (α)) = (β1 , h β (i) (α (i) )) for some i. The case α = α (i) was handled in
1
the original proof (by showing that it yields a designated collision [under h β (i) , which
1
is supposedly a UOWHF]), so here we only handle the case α = α (i) . Now, suppose
(i) (i)
that super-security of the complex scheme was violated; that is, (β1 , β2 ) = (β1 , β2 ).
(i) (i)
Then, by the case hypothesis (which implies β1 = β1 ), it must be that β2 = β2 . This
means that we derive a violation of the super-security of the length-restricted scheme,
(i)
because β2 is a different valid Ss -signature of (β1 , h β1 (α)) = (β1 , h β (i) (α (i) )).
1
(i)
Actually, we have to consider all i’s for which (β1 , h β1 (α)) = (β1 , h β (i) (α (i) ))
1
holds, and observe that violation of super-security for the complex scheme means
(i)
that β2 must be different from each of the corresponding β2 ’s. Alternatively,
(i)
we may first prove that with overwhelmingly high probability, all β1 ’s must be
distinct.

Thus, in both cases we reach a contradiction to the super-security of the length-restricted

signature scheme, which establishes our claim that the resulting (complex) signature
scheme must be super-secure. We stress that, as in Proposition 6.4.31, this proof estab-
lishes that super-security for one-time attacks is preserved, too (because the constructed
forger makes a single query per each query made by the original forger).

Claim 6.5.2.3: Construction 6.4.16, when applied to super-secure one-time signature

schemes, yields super-secure signature schemes.
Proof Sketch: We follow the proof of Proposition 6.4.17, which actually means follow-
ing the proof of Proposition 6.4.15. Specifically, we use almost the same construction
of a forger for the one-time scheme (G, S, V ) (based on the forger for the complex
scheme (G  , S  , V  )). The only difference is in the last step (i.e., the use of the output),
where we consider two forgery cases that are related (but not equal) to the forgery cases
analyzed in the proof of Proposition 6.4.15:39
1. The first case is when the forged signature for the complex scheme (G  , S  , V  )
contains an authentication path (for a leaf) that equals some authentication path
provided by the signing-oracle (as part of the answer to some oracle-query of the
attacker). In this case, the (one-time) verification-key associated with this leaf must
be authentic (i.e., equal to the one used by the signing-oracle), and we derive violation
of the super-security of the instance of (G, S, V ) associated with it. We consider two
subcases (regarding the actual document authenticated via this leaf):
(a) The first subcase is when no oracle-answer has used the instance associated
with this leaf for signing an actual document. (This may happen if the instance

39 Recall that forging a signature for the general scheme requires either using an authentication path supplied
by the (general) signing-oracle or producing an authentication path different from all paths supplied by the
(general) signing-oracle. These are the cases considered here. In contrast, in the proof of Proposition 6.4.15
we considered only the “text part” of these paths, ignoring the question of whether or not the authenticating
(one-time) signatures (provided as part of these paths) are equal.

579
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

associated with the sibling of this leaf was used for signing an actual document.)
In this subcase, as in the proof of Proposition 6.4.15, we obtain (ordinary) forgery
with respect to the instance of (G, S, V ) associated with the leaf (without making
any query to that instance of the one-time scheme).
(b) Otherwise (i.e., the instance associated with this leaf was used for signing an
actual document), the forged document-signature pair differs from the query-
answer pair that used the same leaf. The difference is either in the actual doc-
ument or in the part of the complex-signature that corresponds to the one-time
signature produced at the leaf (because, by the case hypothesis, the authenti-
cation paths are identical). In both subcases this yields violation of the super-
security of the instance of (G, S, V ) associated with that leaf. Specifically, in the
first sub-subcase, we obtain a one-time signature to a different document (i.e.,
violation of ordinary security), whereas in the second sub-subcase, we obtain
a different one-time signature to the same document (i.e., only a violation of
super-security). We stress that in both subcases, the violating signature is ob-
tained after making a single query to the instance of (G, S, V ) associated with
that leaf.
2. We now turn to the second case (i.e., forgery with respect to (G  , S  , V  ) is obtained
by producing an authentication path different from all paths supplied by the signing-
oracle). In this case, we obtain violation of the (one-time) super-security of the
scheme (G, S, V ) associated with one of the internal nodes (specifically the first node
on which the relevant paths differ). The argument is similar (but not identical) to the
one given in the proof of Proposition 6.4.15. Specifically, we consider the maximal
prefix of the authentication path provided by the forger that equals a corresponding
prefix of an authentication path provided by the signing-oracle (as part of its answer).
The extension of this path in the complex-signature provided by the forger either
uses a different pair of (one-time) verification-keys or uses a different (one-time)
signature to the same pair. In the first subcase, we obtain a one-time signature to
a different document (i.e., violation of ordinary security), whereas in the second
subcase, we obtain a different one-time signature to the same document (i.e., only a
violation of super-security). We stress that in both subcases, the violating signature
is obtained after making a single query to the instance of (G, S, V ) associated with
that internal node.
Thus, in both cases we reach a contradiction to the super-security of the one-time
signature scheme, which establishes our claim that the general signature scheme must
be super-secure.
Combining the three claims (and recalling that universal one-way hashing functions
can be constructed using any one-way function [cf. Theorem 6.4.29]), the theorem
follows.

6.5.3. Off-Line/On-Line Signing

Loosely speaking, we say that a signature scheme (G, S, V ) (either a private-key or
a public-key one) has an off-line/on-line signing process if signatures are produced
580

www.Ebook777.com
 6.5* SOME ADDITIONAL PROPERTIES

in two steps, where the first step is independent of the actual message to be signed.
That is, the computation of Ss (α) can be decoupled into two steps, performed by ran-
domized algorithms that are denoted S off and S on , respectively, such that Ss (α) ←
Sson (α, S off (s)). Thus, one may prepare (or precompute) S off (s) before the document
is known (i.e., “off-line”), and produce the actual signature (on-line) once the document
α is presented (by invoking algorithm S on on input (α, S off (s))). This yields improve-
ment in on-line response-time to signing requests, provided that S on is significantly
faster than S itself. This improvement is worthwhile in many natural settings in which
on-line response-time is more important than off-line processing time.
We stress that S off must be randomized (because, otherwise, S off (s) can be incor-
porated in the signing-key). Indeed, one may view algorithm S off as an augmentation
of the key-generation algorithm that produces random extensions of the signing-key
on the fly (i.e., after the verification-key has already been determined). We stress that
algorithm S off is invoked once per each document to be signed, but this invocation can
take place at any time (and even before the document to be signed is even determined).
(In contrast, it may be insecure to reuse the result obtained from S off for two different
signatures.)

Can Secure Signature Schemes Employ Meaningful Off-Line/On-Line Signing

Algorithms? Of course, any algorithm can be vacuously decoupled into two steps, but
we are only interested in meaningful decouplings in which the off-line step takes most
of the computational load. It is interesting to note that schemes based on the refreshing
paradigm (cf. Section 6.4.2.1) lend themselves to such a decoupling. Specifically, in
Construction 6.4.16, only the last step in the signing process depends on the actual
document (and needs to be performed on-line). Furthermore, this last step amounts to
applying the signing algorithm of a one-time signature scheme, which is typically much
faster than all the other steps (which can be performed off-line).40

6.5.4. Incremental Signatures

Loosely speaking, we say that a signature scheme (G, S, V ) (either a private-key or
a public-key one) has an incremental signing process if the signing process can be
sped-up when given a valid signature to a (textually) related document. The actual
definition refers to a set of text-editing operations such as delete word and insert word
(where more powerful operations like cutting a document into two parts and pasting two
documents may be supported, too). Specifically, we require that given a signing-key,
a document-signature pair (α, β), and a sequence of edit operations (i.e., specifying
the operation type and its location), one may modify β into a valid signature β  for the

40 For example, when using the one-time signature scheme suggested in Proposition 6.4.7, producing one-
time signatures amounts to applying a collision-free hashing function and outputting corresponding parts of
the signing-key. This is all that needs to be performed in the on-line step of Construction 6.4.16. In contrast, the
off-line step (of Construction 6.4.16) calls for n applications of a pseudorandom function, n applications of
the key-generation algorithm of the one-time signature scheme, and n applications of the signing algorithm of
the one-time signature scheme.

581
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

modified document α  in time proportional to the number of edit operations (rather than
proportional to |α  |). Indeed, here time is measured in a direct-access model of compu-
tation. Of course, the time saved on the “signing side” should not come at the expense of
a significant increase in verification time. In particular, verification time should depend
only on the length of the final document (and not on the number of edit operations).41
An incremental signing process is beneficial in settings where one needs to sign
many textually related documents (e.g., in simple contracts, much of the text is almost
identical and the few edit changes refer to the party’s specific details, as well as to
specific clauses that may be modified from their standard form in order to meet the
party’s specific needs). In some cases, the privacy of the edit sequence may be of
concern; that is, one may require that the final signature be distributed in a way that
only depends on the final document (rather than depending also on documents that
“contributed” signatures to the process of generating the final signature).

Can Secure Signature Schemes Employ a Meaningful Incremental Signing Pro-

cess? Here, meaningful refers to the set of supported text-modification operations. The
answer is affirmative, and furthermore, these schemes may even protect the privacy of
the edit sequence. In the following, we refer to edit operations that delete/insert fix-
length bit-strings called blocks from/to a document (as well as to the cut-and-paste
operations mentioned previously).

Theorem 6.5.3 (secure schemes with incremental signing process):

1. Assuming the existence of one-way functions, there exist secure message-

authentication schemes having an incremental signing process that supports block
deletion and insertion. Furthermore, the scheme uses a fixed-length authentication
tag.
2. Assuming the existence of one-way functions, there exist secure (private-key and
public-key) signature schemes having an incremental signing process that supports
block deletion and insertion as well as cut and paste.

Furthermore, in both parts, the resulting schemes protect the privacy of the edit
sequence.

Part 1 is proved by using a variant of an efficient message-authentication scheme that is

related to the schemes presented in Section 6.3.1. Part 2 is proved by using an arbitrary
secure (private-key or public-key) signature scheme that produces n-bit long signatures
to O(n)-bit long strings, where n is the security parameter. (Indeed, the scheme need
only be secure in the O(n)-restricted sense.) The document is stored in the leaves of a 2–
3 tree, and the signature essentially consists of the tags of all internal nodes, where each
internal node is tagged by applying the basic signature scheme to the tags of its children.

41 This rules out the naive (unsatisfactory) solution of providing a signature of the original document along with a
signature of the sequence of edit operations. More sophisticated variants of this naive solution (e.g., refreshing
the signature whenever enough edits have occurred) are not ruled out here, but typically they will not satisfy
the privacy requirement discussed in the sequel.

582

www.Ebook777.com
 6.5* SOME ADDITIONAL PROPERTIES

One important observation is that a 2–3 tree supports the said operations while incurring
only a logarithmic (in its size) cost; that is, by modifying only the links of logarithmically
many nodes in the tree. Thus, only the tags of these nodes and their ancestors in the tree
need to be modified in order to form the correspondingly modified signature. (Privacy
of the edit sequence is obtained by randomizing the standard modification procedure for
2–3 trees.) By analogy to Construction 6.2.13 (and Proposition 6.2.14), the incremental
signature scheme is secure.

6.5.5. Fail-Stop Signatures

Loosely speaking, a fail-stop signature scheme is a signature scheme augmented
by a (non-interactive) proof system that allows the legitimate signer to prove to any-
body that a particular (document,signature)-pair was not generated by him/her. Actu-
ally, key-generation involves interaction with an administrating entity (which publi-
cizes the resulting verification-keys), rather than just having the user publicize his/her
verification-key. In addition, we allow memory-dependent signing procedures (as in
Definition 6.4.13).42 The system guarantees the following four properties, where the
first two properties are the standard ones:

1. Proper operation: In case the user is honest, the signatures produced by it will pass
the verification procedure (with respect to the corresponding verification-key).
2. Infeasibility of forgery: In case the user is honest, forgery is infeasible in the standard
sense. That is, every feasible chosen message attack may succeed (in generating a
valid signature to a new message) only with negligible probability.
3. Revocation of forged signatures: In case the user is honest and forgery is commit-
ted, the user can prove that indeed forgery has been committed. That is, for every
chosen message attack (even a computationally unbounded one)43 that produces a
valid signature to a new message, except for with negligible probability, the user
can efficiently convince anyone (which knows the verification-key) that this valid
signature was forged (i.e., produced by somebody else).
4. Infeasibility of revoking unforged signatures: It is infeasible for a user to create
a valid signature and later convince someone that this signature was forged (i.e.,
produced by somebody else). Indeed, it is possible (but not feasible) for a user to
cheat here.
Furthermore, Property 3 (i.e., revocation of forged signatures) holds also in case the
administrating entity participates in the forgery and even if it behaves improperly at the
key-generation stage. (In contrast, the other items hold only if the administrating entity
behaves properly during the key-generation stage.)
To summarize, fail-stop signature schemes allow proving that forgery has occurred,
and so offer an information-theoretic security guarantee to the potential signers (yet the

42 Allowing memory-dependent signing is essential to the existence of secure fail-stop signature schemes; see
Exercise 25.
43 It seems reasonable to restrict even computationally unbounded adversaries to polynomially many signing

requests.

583
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

guarantee to potential signature recipients is only a computational one).44 In contrast,

when following the standard semantics of signature schemes, the potential signers have
only a computational security guarantee, and the signature recipients have an absolute
guarantee: Whenever the verification algorithm accepts a signature, it is by definition
an unrevocable one.

Do Secure Fail-Stop Signature Schemes Exist? Assuming the intractability of either

the Discrete Logarithm Problem or of integer factorization, the answer is affirmative.
Indeed, in fail-stop signature schemes, each document must have super-polynomially
many possible valid signatures (with respect to the publicly known verification-key),
but only a negligible fraction of these will be (properly) produced by the legitimate
signer (who knows a corresponding signing-key, which is not uniquely determined by
the verification-key). Furthermore, any strategy (even an infeasible one) is unlikely to
generate signatures corresponding to the actual signing-key. On the other hand, it is
infeasible given one signing-key to produce valid signatures (i.e., with respect to the
verification-key) that do not correspond to the proper signing with this signing-key.

6.6. Miscellaneous

6.6.1. On Using Signature Schemes

Once defined and constructed, signature schemes may be (and are actually) used as
building blocks toward various goals that are different from the original motivation.
Still, the original motivation (i.e., reliable communication of information) is of great
importance, and in this subsection we discuss several issues regarding the use of signa-
ture schemes toward achieving it. The discussion is analogous to a similar discussion
conducted in Section 5.5.1, but the analogous issues discussed here are even more
severe.

Using Private-Key Schemes: The Key-Exchange Problem. As discussed in Sec-

tion 6.1, using a private-key signature scheme (i.e., a message-authentication scheme)
requires the communicating parties to share a secret key. This key can be generated
by one party and secretly communicated to the other party by an alternative (expen-
sive) secure and reliable channel. Often, a preferable solution consists of employing a
key-exchange (or rather key-generation) protocol, which is executed over the standard
(unreliable) communication channel. We stress that here (unlike in Section 5.5.1) we
must consider active adversaries. Consequently, the focus should be on key-exchange
protocols that are secure against active adversaries and are called unauthenticated key-
exchange protocols (because the messages received over the channel are not necessarily
authentic). Such protocols are too complex to be treated in this section, and the interested
reader is referred to [29, 30, 15].

44 We refer to the natural convention by which a proof of forgery frees the signer of any obligations implied by the
document. In this case, when accepting a valid signature, the recipient is only guaranteed that it is infeasible for
the signer to revoke the signature.

584

www.Ebook777.com
 6.6 MISCELLANEOUS

Using State-Dependent Message-Authentication Schemes. In many communica-

tion settings, it is reasonable to assume that the authentication device may maintain
(and modify) a state (e.g., a counter or a clock). Furthermore, in many applications, a
changing state (e.g., a clock) must be employed anyhow in order to prevent replay of
old messages (i.e., each message is authenticated along with its transmission time). In
such cases, state-dependent schemes as discussed in Section 6.3.2 may be preferable.
(See further discussion in Section 6.3.2 and analogous discussion in Section 5.5.1.)
Using Signature Schemes: Public-Key Infrastructure. The standard use of (public-
key) signature schemes in real-life applications requires a mechanism for providing the
verifiers with the signer’s authentic verification-key. In small systems, one may assume
that each user holds a local record of the verification-keys of all other users. However,
this is not realistic in large-scale systems, and so the verifier must obtain the relevant
verification-key on the fly in a “reliable” way (i.e., typically, certified by some trusted
authority). In most theoretical work, one assumes that the verification-keys are posted
and can be retrieved from a public-file that is maintained by a trusted party (which
makes sure that each user can post only verification-keys bearing its own identity).
Alternatively, such a trusted party may provide each user with a (signed) certificate
stating the authenticity of the user’s verification-key. In practice, maintaining such a
public-file (and/or handling such certificates) is a major problem, and mechanisms that
implement these abstractions are typically referred to by the generic term “public-
key infrastructure” (PKI). For a discussion of the practical problems regarding PKI
deployment, see, e.g., [149, Chap. 13].

6.6.2. On Information-Theoretic Security

In contrast to the bulk of our treatment, which focuses on computationally bounded
adversaries, in this section we consider computationally unbounded adversaries. Specif-
ically, we consider computationally unbounded chosen message attacks, but do bound
(as usual, by an unknown polynomial) the total number of bits in the signing-queries
made by such attackers. We call a (private-key or public-key) signature scheme perfectly
secure (or information-theoretically secure) if even such computationally unbounded
attackers may succeed (in forgery) only with negligible probability.
It is easy to see that no (public-key) signature scheme may be perfectly secure,
not even in a length-restricted one-time sense. The reason is that a computationally
unbounded adversary that is given a verification-key can find (without making any
queries) a corresponding signing-key, which allows it to forge signatures to any message
of its choice.
In contrast, restricted types of message-authentication schemes (i.e., private-key sig-
nature schemes) may be perfectly secure. Specifically, given any polynomial bound on
the total number of messages to be authenticated, one may construct a correspond-
ing state-based perfectly secure message-authentication scheme. In fact, a variant of
Construction 6.3.11 will do, where a truly random one-time pad is used instead of the
pseudorandom sequence generated using the next-step function g. Indeed, this one-
time pad will be part of the key, which in turn must be longer than the total number
of messages to be authenticated. We comment that the use of a state is essential for
585
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

allowing several messages to be authenticated (in a perfectly secure manner). (Proofs

of both statements can be derived following the ideas underlying Exercise 7.)

6.6.3. On Some Popular Schemes

The reader may note that we have avoided the presentation of several popular sig-
nature schemes (i.e., public-key ones). Some of these schemes (e.g., RSA [176] and
DSS [160]) seem to satisfy some weak notions of security (i.e., a notion weaker than
Definition 6.1.2). Variants of these schemes can be proven to be secure in the random
oracle model, provided some standard intractability assumptions hold (cf., e.g., [31]).
For reasons to be outlined, we choose not to present these results here.

On Using Weaker Definitions. We distinguish between weak definitions that make

clear reference to the abilities of the adversary (e.g., one-message attacks, length-
restricted message attacks) and weak notions that make hidden and unspecified as-
sumptions regarding what may be beneficial to the adversary (e.g., “forgery of signa-
tures for meaningful documents”). In our opinion, the fact that the hidden assumptions
often “feel right” makes them even more dangerous, because it means that they are
never seriously considered (and not even formulated). For example, it is often claimed
that existential forgery (see Section 6.1.4) is “merely of theoretical concern,” but these
claims are never supported by any evidence or by a specification of the types of forgery
that are of “real practical concern.” Furthermore, it has been demonstrated that this
“merely theoretical” issue yields a real security breach in some important practical
applications. Still, weak definitions of security may be useful (i.e., suffice for some
applications), provided that they are clearly stated and that one realizes their limitations
(and, in particular, their “non-generality”). However, since the current work focuses
on generally applicable definitions, we choose not to discuss such weaker notions of
security and not to present schemes that can be evaluated only with respect to these
weaker notions.45 The interested reader is referred to [125] for a comprehensive treat-
ment of various (weaker) notions of security (which refer to various types of attacks
and success events).

On the Random Oracle Methodology. The Random Oracle Methodology [92, 28]
consists of two steps: First, one designs an ideal system in which all parties (including
the adversary) have oracle access to a truly random function, and proves this ideal
system to be secure (in which case, one says that the system is secure in the ran-
dom oracle model). Next, one replaces the random oracle with a “good cryptographic
hashing function,” providing all parties (including the adversary) with the succinct de-
scription of this function, and hopes that the resulting (actual) scheme is secure.46 We
warn that this hope has no sound justification. Furthermore, there exist encryption and

45 Needless to say, we did not even consider presenting schemes that are not known to satisfy some robust notion
of security.
46 Recall that, in contrast, the methodology of Section 3.6.3 (which is applied often in the current chapter) refers

to a situation in which the adversary does not have direct oracle access to the random function, and does not
obtain the description of the pseudorandom function used in the latter implementation.

586

www.Ebook777.com
 6.6 MISCELLANEOUS

signature schemes that are secure in the Random Oracle Model, but replacing the ran-
dom function (used in them) by any function ensemble yields a totally insecure scheme
(cf., [54]).

6.6.4. Historical Notes

As in the case of encryption schemes, the rigorous study of the security of private-
key signature schemes (i.e., message-authentication schemes) has lagged behind the
corresponding study of public-key signature schemes. The current section is organized
accordingly.

6.6.4.1. Signature Schemes

The notion of a (public-key) signature scheme was introduced by Diffie and Hell-
man [75], who also suggested implementing it using trapdoor permutations. Con-
crete implementations were suggested by Rivest, Shamir, and Adleman [176] and by
Rabin [171]. However, definitions of security for signature schemes were presented
only a few years afterward. Still, the abstract notion of a signature scheme as well as
the concrete candidate implementations have served as the basis for the development
of the theory presented in the current chapter.
A first rigorous treatment of security notions for signature schemes was suggested by
Goldwasser, Micali, and Yao [127], but their definition is weaker than the one followed
in our text. (Specifically, the adversary’s queries in the definition of [127] are deter-
mined non-adaptively and obliviously of the public-key.) Assuming the intractability of
factoring, they also presented a signature scheme that is secure under their definition.
We mention that the security definition of [127] considers existential forgery, and is
thus stronger than security notions considered before [127].
A comprehensive treatment of security notions for signature schemes, which cul-
minates in the notion used in our text, was presented by Goldwasser, Micali, and
Rivest [125]. Assuming the intractability of factoring, they also presented a signature
scheme that is secure (in the sense of Definition 6.1.2). This was the first time that
a signature scheme was proven secure under a simple intractability assumption such
as the intractability of factoring. Their proof has refuted a folklore (attributed to Ron
Rivest) by which no such “constructive proof” may exist (because the mere existence of
such a proof was believed to yield a forging procedure).47 Whereas the (two) schemes

47 The flaw in this folklore is rooted in implicit (unjustified) assumptions regarding the notion of a “constructive
proof of security” (based on factoring). In particular, it was implicitly assumed that the signature scheme uses
a verification-key that equals a composite number, and that the proof of security reduces the factoring of such a
composite N to forging with respect to the verification-key N . In such a case, the folklore suggested that the re-
duction yields an oracle machine for factoring the verification-key, where the oracle is the corresponding signing-
oracle (associated with N ), and that the factorization of the verification-key allows for efficiently producing
signatures to any message. However, none of these assumptions is justified. In contrast, the verification-key in the
scheme of [125] consists of a pair (N , x), and its security is proven by reducing the factoring of N to forging with
respect to the verification-key (N , r ), where r is randomly selected by the reduction. Furthermore, on input N , the
(factoring) reduction produces a verification-key (N , r ) that typically does not equal the verification-key (N , x)
being attacked, and so being given access to a corresponding signing-oracle does not allow the factoring of N .

587
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

of [127] were inherently memory-dependent, the scheme of [125] has a “memoryless”

variant (cf. [100] and [125]).
Following Goldwasser, Micali, and Rivest [125], research has focused on construct-
ing secure signature schemes under weaker assumptions. In fact, as noted in [125],
their construction of secure signature schemes can be carried out using any collec-
tion of claw-free, trapdoor permutation pairs. The claw-free requirement was removed
in [26], whereas the seemingly more fundamental trapdoor requirement was removed by
Naor and Yung [163]. Finally, Rompel showed that one may use arbitrary one-way func-
tions rather than one-way permutations [178], and thus established Theorem 6.4.1. The
progress briefly summarized here was enabled by the use of many important ideas and
paradigms; some of them were introduced in that body of work and some were “only”
revisited and properly formalized. Specifically, we refer to the introduction of the re-
freshing paradigm in [125], the use of authentication trees (cf., [151, 152], and [125]),
the use of the hash-and-sign paradigm (rigorously analyzed in [70]), the introduction of
Universal One-Way Hash Functions (and the adaptation of the hash-and-sign paradigm
to them) in [163], and the use of one-time signature schemes (cf., [170]).
We comment that our presentation of the construction of signature schemes is dif-
ferent from the one given in any of these cited papers. Specifically, the main part of
Section 6.4 (i.e., Sections 6.4.1 and 6.4.2) is based on a variant of the signature scheme
of [163], in which collision-free hashing (cf. [70]) is used instead of universal one-way
hashing (cf. [163]).

6.6.4.2. Message-Authentication Schemes

Message authentication schemes were first discussed in the information-theoretic set-
ting, where a one-time pad was used. Such schemes were first suggested in [99], and
further developed in [188]. The one-time pad can be implemented by a pseudoran-
dom function (or an on-line pseudorandom generator), yielding only computational
security, as we have done in Section 6.3.2. Specifically, Construction 6.3.11 is based
on [139, 140]. In contrast, in Section 6.3.1 we have followed a different paradigm that
amounts to applying a pseudorandom function to the message (or its hashed value),
rather than using a pseudorandom function (or an on-line pseudorandom generator) to
implement a one-time pad. This alternative paradigm is due to [111], and is followed
in works such as [25, 22, 13]. Indeed, following this paradigm, one may focus on
constructing generalized pseudorandom function ensembles (as in Definition 3.6.12),
based on ordinary pseudorandom functions (as in Definition 3.6.4). See comments
on alternative presentations at the end of Sections 6.3.1.2 and 6.3.1.3, as well as in
Section C.2 of Appendix C.

6.6.4.3. Additional Topics

Collision-free hashing was first defined in [70]. Construction 6.2.8 is also due to [70],
with underlying principles that can be traced to [125]. Construction 6.2.11 is due to [71].
Construction 6.2.13 is due to [153].
Unique signatures and super-security have been used in several works, but they
were not treated explicitly before. The notion of off-line/on-line signature schemes
588

www.Ebook777.com
 6.6 MISCELLANEOUS

was introduced (and first instantiated) in [85]. The notion of incremental crypto-
graphic schemes (and, in particular, incremental signature schemes) was introduced
and instantiated in [18, 19]. In particular, the incremental MAC of [19] (i.e., Part 1 of
Theorem 6.5.3) builds on the message-authentication scheme of [22], and the incre-
mental signature scheme that protects the privacy of the edit sequence is due to [158]
(building upon [19]). Fail-stop signatures were defined and constructed in [167].

6.6.5. Suggestions for Further Reading

As mentioned, the work of Goldwasser, Micali, and Rivest contains a comprehensive
treatment of security notions for signature schemes [125]. Their treatment refers to
two parameters: (1) the type of attack, and (2) the type of forgery that is deemed
successful. The most severe type of attack allows the adversary to adaptively select
the documents to be signed (as in Definition 6.1.2). The most liberal notion of forgery
refers to producing a signature to any document for which a signature was not obtained
during the attack (again, as in Definition 6.1.2). Thus, the notion of security presented
in Definition 6.1.2 is the strongest among the notions discussed in [125]. Still, in some
applications, weaker notions of security may suffice. We stress that one may still benefit
from the definitional part of [125], but the constructive part of [125] should be ignored
because it is superseded by later work (on which our presentation is based).
Pfitzmann’s book [168] contains a comprehensive discussion of many aspects in-
volved in the integration of signature schemes in real-life systems. In addition, her
book surveys variants and augmentations of the notion of signature schemes, viewing
the one treated in the current book as “ordinary.” The focus is on fail-stop signature
schemes [168, Chap. 7–11], but much attention is given to the presentation of a gen-
eral framework [168, Chap. 5] and to a review of other “non-ordinary” schemes [168,
Secs. 2.7 and 6.1].
As hinted in Section 6.6.4.2, our treatment of the construction of message-
authentication schemes is merely the tip of the iceberg. The interested reader is
referred to [186, 139, 140, 35] for details on the “one-time pad” approach, and
to [25, 22, 13, 14, 20, 2] for alternative approaches. Constructions and discussion
of AXU hashing functions (which are stronger than generalized hashing functions) can
be found in [139, 140].
The constructions of universal one-way hash functions presented in Section 6.4.3
use any one-way permutation, and do so in a generic way. The number of applications
of the one-way permutation in these constructions is linearly related to the difference
between the number of input and output bits in the hash function. In [98], it is shown that
as far as generic (black-box) constructions go, this is essentially the best performance
that one can hope for.
In continuation of the discussion in Section 6.4.2.4 (regarding the construction of
signature schemes based on authentication trees), we refer the reader to [81, 67], in
which specific implementations (of a generalization) of Constructions 6.4.14 and 6.4.16
are presented. Specifically, these works utilize an authentication-tree of large degree
(rather than binary trees as in Section 6.4.2.2).
589
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

In continuation of the discussion in Section 6.5.1, we mention that signature schemes

having unique signatures are related (but not equivalent) to verifiable pseudorandom
functions (as defined and constructed in [155]). In particular, the construction in [155]
does yield signature schemes having unique signatures, and thus the latter exist under
a quite standard assumption (regarding RSA). We comment that signature schemes
having unique signatures are stronger than invariant signature schemes (as defined
in [128] and studied in [21, 128]).

6.6.6. Open Problems

The known construction of signature schemes from arbitrary one-way functions [178]
is merely a feasibility result. It is indeed an important open problem to provide an
alternative construction that may be practical and still utilize an arbitrary one-way
function. We believe that providing such a construction may require the discovery of
important new paradigms.

6.6.7. Exercises
Exercise 1: Deterministic signing and verification algorithms:

1. Using a pseudorandom function ensemble, show how to transform any (private-

key or public-key) signature scheme into one employing a deterministic signing
algorithm.
2. Using a pseudorandom function ensemble, show how to transform any message-
authentication scheme into one employing deterministic signing and verification
algorithms.
3. Verify that all signature schemes presented in the current chapter employ a deter-
ministic verification algorithm.
4. (By Boaz Barak:) Show that any length-restricted signature scheme can be easily
transformed into one employing a deterministic verification algorithm.

Guideline (for Part 1): Augment the signing-key with a description of a pseudo-
random function, and apply this function to the string to be signed in order to extract
the randomness used by the original signing algorithm.

Guideline (for Part 2): Analogous to Part 1. (Highlight your use of the private-key
hypothesis.) Alternatively, see Exercise 2.

Guideline (for Part 4): First transform the signature scheme into one in which
all valid signatures are of a length that is bounded by a polynomial in the security
parameter (and the length of the messages). Let (n) denote the length of the docu-
ments and m(n) denote the length of the corresponding signatures. Next, amplify the
verification algorithm such that its error probability is smaller than 2−((n)+m(n)+n) .
Finally, incorporate the coin tosses of the verification algorithm in the verification-
key, making the former deterministic.

590

www.Ebook777.com
 6.6 MISCELLANEOUS

Exercise 2: Canonical verification in the private-key version: Show that, without loss
of generality, the verification algorithm of a private-key signature scheme may con-
sist of comparing the alleged signature to one produced by the verification algo-
rithm itself; that is, the verification algorithm uses a verification-key that equals the
signing-key and produces signatures exactly as the signing algorithm.
Why does this claim fail with respect to public-key schemes?

Guideline: Use Part 1 of Exercise 1, and conclude that on a fixed input, the signing
algorithm always produces the same output. Use the fact that (by Exercise 7.3) the
existence of message-authentication schemes implies the existence of pseudoran-
dom functions, which are used in Part 1 of Exercise 1.

Exercise 3: Augmented attacks in the private-key case: In continuation of the discus-

sion in Section 6.1.5.1, consider the definition of an augmented attack (on a private-
key signature scheme) in which the adversary is allowed verification-queries.

1. Show that in case the private-key signature scheme has unique valid signatures,
it is secure against augmented attacks if and only if it is secure against ordinary
attacks (as in Definition 6.1.2).
2. Assuming the existence of secure private-key signature schemes (as in
Definition 6.1.2), present such a secure scheme that is insecure under augmented
attacks.

Guideline (Part 1): Analyze the emulation outlined in the proof of Proposi-
tion 6.1.3. Specifically, ignoring the redundant verification-queries (for which the
answer is determined by previous answers), consider the probability that the em-
ulation has gambled correctly on all the verification-queries up to (and including)
the first such query that should be answered affirmatively.

Guideline (Part 2): Given any secure MAC, (G, S, V ), assume without loss of
generality that in the key-pairs output by G, the verification-key equals the signing-
key. Consider the scheme (G  , S  , V  ) (with G  = G), where Ss (α) = (Ss (α), 0),
Vv (α, (β, 0)) = Vv (α, β), and Vv (α, (β, i, σ )) = 1 if both Vv (α, β) = 1 and the i-th
bit of v is σ . Prove that (G  , S  , V ) is secure under ordinary attacks, and present an
augmented attack that totally breaks it (i.e., obtains the signing-key s = v).

Exercise 4: The signature may reveal the document: Both for private-key and public-
key signature schemes, show that if such secure schemes exist, then there exist
secure signature schemes in which any valid signature to a message allows for
efficient recovery of the entire message.
Exercise 5: On the triviality of some length-restricted signature schemes:

1. Show that for logarithmically bounded , secure -restricted private-key signa-

ture schemes (i.e., message-authentication schemes) can be trivially constructed
(without relying on any assumption).
2. In contrast, show that the existence of a secure -restricted public-key signature
scheme, even for  ≡ 1, implies the existence of one-way functions.
591
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

(n)
Guideline (Part 1): On input 1n , the key-generator uniformly selects s ∈ {0, 1}2 ·n ,
and outputs the key pair (s, s). View s = s1 · · · s2(n) , where each si is an n-bit
long string, and consider any fixed ordering of the 2(n) strings of length (n).
The signature to α ∈ {0, 1}(n) is defined as si , where i is the index of α in the latter
ordering.

Guideline (Part 2): Let (G, S, V ) be a 1-restricted public-key signature scheme.

Define f (1n , r ) = v if, on input 1n and coins r , algorithm G generates a key-pair
of the form (·, v). Assuming that algorithm A inverts f with probability ε(n), we
construct a forger that attacks (G, S, V ) as follows. On input a verification key v,
the forger invokes A on input v. With probability ε(n), the forger obtains r such
that f (1n , r ) = v. In such a case, the forger obtains a matching signing-key s (i.e.,
(s, v) is output by G(1n ) on coins r ), and so can produce valid signatures to any
string of its choice.

Exercise 6: Failure of Construction 6.2.3 in case (n) = O(log n): Show that if Con-
struction 6.2.3 is used with a logarithmically bounded , then the resulting scheme
is insecure.
Guideline: Note that by asking for polynomially many signatures, the adversary
may obtain two Ss -signatures that use the same (random) identifier. Specifically,
consider making the queries αα, for all possible α ∈ {0, 1}(n) , and note that if
αα and α α  are Ss -signed using the same identifier, then we can derive a valid
Ss -signature to αα  .

Exercise 7: Secure MACs imply one-way functions: Prove that the existence of se-
cure message-authentication schemes implies the existence of one-way functions.
Specifically, let (G, S, V ) be as in the hypothesis.

1. To simplify the following two items, show that, without loss of generality, G(1n )
uses n coins and outputs a signing-key of length n.
2. Assume first that S is a deterministic signing algorithm. Prove that
def
f (r, α1 , ..., αm ) = (Ss (α1 ), ..., Ss (αm ), α1 , ..., αm ) is a one-way function, where
s = G 1 (r ) is the signing-key generated with coins r , all αi ’s are of length n = |r |,
and m = (n).
3. Extend the proof to handle randomized signing algorithms, thus establishing the
main result.
Guideline (Parts 2 and 3): Note that with high probability (over the choice of the
αi ’s), the m signatures (i.e., Ss (αi )’s) determine a set R such that for every r  ∈ R,
it holds that SG 1 (r  ) (α) = Ss (α) for most α ∈ {0, 1}n . (Note that G 1 (r  ) does not
necessarily equal s.) Show that this implies that the ability to invert f yields the
ability to forge (under a chosen message attack). (Hint: Use m random signing-
queries to produce a random image of f , and use the obtained pre-image under
f , which contains an adequate signing-key, to forge a signature to a new random
message.) The extension to randomized signing is obtained by augmenting the pre-
image of the one-way function with the coins used by the m invocations of the
signing algorithm.

592

www.Ebook777.com
 6.6 MISCELLANEOUS

Exercise 8: General pseudorandom functions yield general secure MACs: Using

a pseudorandom function ensemble of the form { f s : {0, 1}∗ → {0, 1}|s| }s∈{0,1}∗ ,
construct a general secure message-authentication scheme (rather than a length-
restricted one).

Guideline: The construction is identical to Construction 6.3.1, except that here we

use a general pseudorandom function ensemble rather than the one used there. The
proof of security is analogous to the proof of Proposition 6.3.2.

Exercise 9: Consider a generalization of Construction 6.3.5 in which the pseudo-

random function is replaced by an arbitrary secure MAC such that on input a
signing-key (r, s), a document α ∈ {0, 1}(n) is signed by applying the MAC (with
signing-key s) to h r (α). Show that, for some secure MAC and some collections
of hash functions with negligible collision probability, the suggested scheme is
insecure.

Guideline: Use the fact that the MAC may reveal the first part of its argument,
whereas the hashing function may yield an output value in which the second part
is fixed. Furthermore, it may be easy to infer the hashing function from sufficiently
many input–output pairs, and it may be easy to find a random pre-image of a
given hash function on a given image. Present constructions that satisfy all these
conditions, and show how combining them yields the desired result.

Exercise 10: Easily obtaining pseudorandom functions from certain MACs (advanced
exercise, based on [162]): Let (G, S, V ) be a secure message-authentication scheme,
and suppose that S is deterministic. Furthermore, suppose that |G 1 (1n )| = n and that
def
for every s, x ∈ {0, 1}n it holds that |Ss (x)| = (n) = |Ss (1n )|. Consider the Boolean
function ensemble { f s1 ,s2 : {0, 1}|s1 | → {0, 1}}s1 ,s2 , where s1 is selected according to
G 1 (1n ) and s2 ∈ {0, 1}(n) is uniformly distributed, such that f s1 ,s2 (α) is defined to
equal the inner product mod 2 of Ss1 (α) and s2 . Prove that this function ensemble
is pseudorandom (as defined in Definition 3.6.9 for the case d(n + (n)) = n and
r (n) = 1).

Guideline: Consider hybrid experiments such that in the i-th hybrid the first i
queries are answered by a truly random Boolean function and the rest of the queries
are answered by a uniformly distributed f s1 ,s2 . (Note that it seems important to use
this non-standard order of random versus pseudorandom answers.) Show that distin-
guishability of the i-th and i + 1st hybrids implies that a probabilistic polynomial-
time oracle machine can have a non-negligible advantage in the following game. In
the game, the machine is first asked to select α; next f s1 ,s2 is uniformly selected,
and the machine is given s2 as well as oracle access to Ss1 (but is not allowed the
query α) and is asked to guess f s1 ,s2 (α) (or, equivalently, to distinguish f s1 ,s2 (α)
from a truly random bit).48 At this point, one may apply the proof of Theorem 2.5.2,

48 Note that the particular order (of random versus pseudorandom answers in the hybrids) allows this oracle
machine to generate the (corresponding) hybrid while playing this game properly. That is, the player answers

593
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

and deduce that the said oracle machine can be modified to construct Ss1 (α) with
non-negligible probability (when given oracle access to Ss1 but not being allowed
the query α), in contradiction to the security of the MAC.

Exercise 11: Prove that without loss of generality, one can always assume that a chosen
message attack makes at least one query. (This holds for general signature schemes
as well as for length-restricted and/or one-time ones.)
Guideline: Given an adversary A that outputs a message-signature pair (α  , β  )
without making any query, modify it such that it makes an arbitrary query α ∈

{0, 1}|α | \ {α  } just before producing that output.

Exercise 12: On perfectly secure one-time message-authentication (MAC) schemes:

By perfect (or information-theoretic) security we mean that even computationally
unbounded chosen message attacks may succeed (in forgery) only with negligible
probability.
Define perfect (or information-theoretic) security for one-time MACs and length-
restricted one-time MACs. (Be sure to bound the length of documents (e.g., by some
super-polynomial function) also in the unrestricted case; see Part 3 of the current
exercise, as well as Exercise 21.)
Prove the following, without relying on any (intractability) assumptions (which are
useless anyhow in the information-theoretic context):
1. For any polynomially bounded and polynomial-time computable function  :
N → N, perfectly secure -restricted one-time MACs can be trivially constructed.
2. Using a suitable AXU family of hashing functions, present a construction of a
perfectly secure one-time MAC. Furthermore, present such a MAC in which the
authentication-tags have fixed length (i.e., depending on the length of the key but
not on the length of the message being authenticated).
3. Show that any perfectly secure one-time MAC that utilizes fixed-length
authentication-tags and a deterministic signing algorithm yields a generalized
hashing ensemble with negligible collision probability. Specifically, for any poly-
nomial p, this ensemble has a ( p, 1/ p)-collision property.

Guideline: For Part 1, combine the ideas underlying Exercise 5.1 and Construc-
tion 6.4.4. For Part 2, use the ideas underlying Construction 6.3.11 and the proof of
Proposition 6.3.12. For Part 3, given a MAC as in the claim, consider the functions
def
h s (x) = Ss (x), where s is selected as in the key-generation algorithm.

Exercise 13: Secure one-time (public-key) signatures imply one-way functions: In con-
trast to Exercise 12, prove that the existence of secure one-time signature schemes
implies the existence of one-way functions. Furthermore, prove that this holds even

the first i queries at random, sets α to equal the i + 1st query, uses the tested bit value as the corresponding
answer, and uses s2 and the oracle Ss1 to answer the subsequent queries. It is also important that the game be
defined such that s2 is given only after the machine has selected α; see [162].

594

www.Ebook777.com
 6.6 MISCELLANEOUS

for 1-restricted signature schemes that are secure (only) under attacks that make no
signing-queries.

Guideline: See guideline for Item 2 in Exercise 5.

Exercise 14: Prove that the existence of collision-free hashing collections implies the
existence of one-way functions.

Guideline: Given a collision-free hashing collection, {h r : {0, 1}∗ →

{0, 1}(|r |) }r ∈{0,1}∗ , consider the function f (r, x) = (r, h r (x)), where (say) |x| =
(|r |) + |r |. Prove that f is a one-way function, by assuming toward the contra-
diction that f can be efficiently inverted with non-negligible probability, and de-
riving an efficient algorithm that forms collisions on random h r ’s. Given r , form a
collision under the function h r , by uniformly selecting x ∈ {0, 1}(|r |)+|r | , and feed-
ing the inverting algorithm with input (r, h r (x)). Observe that with non-negligible
probability, a pre-image is obtained, and that only with exponentially vanishing
probability this pre-image is (r, x) itself. Thus, with non-negligible probability, we
obtain a pre-image (r, x  ) = (r, x) such that h r (x  ) = h r (x).

Exercise 15: Modify Construction 6.2.8 so as to allow the computation of the hash-
value of an input string while processing the input in an on-line fashion; that is,
the implementation of the hashing process should process the input x in a bit-by-bit
manner, while storing only the current bit and a small amount of state information
(i.e., the number of bits encountered so far and an element of Ds ).

def y y y
Guideline: All that is needed is to redefine h (s,r ) (x) = f s t f s t−1 · · · f s 1 (r ), where
y1 · · · yt is a suffix-free encoding of x; that is, for any x = x  , the coding of x is not
a suffix of the coding of x  .

Exercise 16: Secure MACs that hide the message: In contrast to Exercise 4, show that
if secure message-authentication schemes exist, then there exist such schemes in
which it is infeasible (for a party not knowing the key) to extract from the sig-
nature any partial information about the message (except for the message length).
(Indeed, privacy of the message is formulated as the definition of semantic security
of encryption schemes; see Chapter 5.)

Guideline: Combine a message-authentication scheme with an adequate private-

key encryption scheme. Refer to such issues as the type of security required of the
encryption scheme and why the hypothesis yields the existence of the ingredients
used in the construction.

Exercise 17: In continuation of Exercise 16, show that if there exist collision-free
hashing functions, then there exist message-authentication schemes in which it is
infeasible (for a party not knowing the key) to extract from the signature any partial
information about the message including the message length. How come we can

595
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

hide the message length in this context, whereas we cannot do this in the context of
encryption schemes?

Guideline: Combine a message-authentication scheme having fixed-length signa-

tures with an adequate private-key encryption scheme. Again, refer to issues as in
Exercise 16.

Exercise 18: Alterntaive formulation of state-based MACs (by Boaz Barak): For
S = (S  , S  ) and V = (V  , V  ), consider the following reformulation of Item 2 of
Definition 6.3.9: For every pair (s (0) , v (0) ) in the range of G(1n ), every sequence
of messages α (i) ’s, and every i, it holds that V  (v (i−1) , α (i) , S  (s (i−1) , α (i) )) = 1,
 ( j−1) |α ( j) |
where s ( j) = S  (s ( j−1) , 1|α | ) and v ( j) = V  (v ( j−1) , 1|α | , 1|S (e ,1 )| ) for j =
( j) ( j)

1, ..., i − 1. Prove the equivalence of the two formulations.

Exercise 19: Prove that the existence of collections of UOWHF implies the existence
of one-way functions. Furthermore, show that uniformly chosen functions in any
collection of UOWHFs are hard to invert (in the sense of Definition 2.4.3).

Guideline: Note that the guidelines provided in Exercise 14 can be modified to fit

the current context. Specifically, the suggested collision-forming algorithm is given
uniformly distributed r and x, and invokes the inverter on input (r, h r (x)), hoping
to obtain a designated collision with x under h r . Note that the furthermore-clause
is implicit in the proof.

Exercise 20: Assuming the existence of one-way functions, show that there exists a
collection of universal one-way hashing functions that is not collision-free.

Guideline: Given a collection of universal one-way hashing functions, { f s :

{0, 1}∗ → {0, 1}|s| }, consider the collection F  = { f s : {0, 1}∗ → {0, 1}|s| } defined
so that f s (x) = (0, f s (x)) if the |s|-bit long prefix of x is different from s, and
f s (sx  ) = (1, s) otherwise. Clearly, F  is not collision-free. Show that F  is a col-
lection of universal one-way hashing functions.

Exercise 21: Show that for every finite family of functions H , there exists x = y such
that h(x) = h(y) for every h ∈ H . Furthermore, show that for H = {h : {0, 1}∗ →
{0, 1}m }, this holds even for |x|, |y| ≤ m · |H |.

Guideline: Consider the mapping x → (h 1 (x), ..., h t (x)), where H = {h i }i=1

t
.
m t
Since the number of possible images is at most (2 ) , we get a collision as soon as
we consider more than 2mt pre-images.

Exercise 22: Constructions of Hashing Families with Bounded Collision Probability:

In continuation of Exercise 22.2 in Chapter 3, consider the set of functions Sm
associated with -by-m Toeplitz matrix; that is, h T (x) = T x, where T = (Ti, j ) is a
Toeplitz matrix (i.e., Ti, j = Ti+1, j+1 for all i, j). Show that this family has collision
probability 2−m . (Note that each -by-m Toeplitz matrix is specified using  + m − 1
bits.)

596

www.Ebook777.com
 6.6 MISCELLANEOUS

Guideline: Note that we have eliminated the shifting vector b used in Exercise 22.2
of Chapter 3, but this does not affect the relevant analysis.

Exercise 23: Constructions of Generalized Hashing Families with Bounded Collision

Property: (See definition in Section 6.3.1.3.)
1. Using a variant of the tree-hashing scheme of Construction 6.2.13, construct a
generalized
√
hashing ensemble with a ( f, 1/ f )-collision property, where f (n) =
2 n/2 . (Hint: Use a different hashing function at each level of the tree.)
2. (By Hugo Krawczyk): Show that the tree-hashing scheme of Construction 6.2.13,
where the same hashing function is used in all levels of the tree, fails in the current
context. That is, there exists a hashing ensemble {hr : {0, 1}2m(|r |) → {0, 1}m(|r |) }r
with negligible collision probability, such that applying Construction 6.2.13 to it
(even with depth two) yields an ensemble with high collision probability.
3. As in Part 2, show that the block-chaining method of Construction 6.2.11 fails in
the current context (even for three blocks).

Guideline (Part 1): Let {h r : {0, 1}2m(|r |) → {0, 1}m(|r |) }r be a hashing ensemble
with collision probability cp. Recall that such ensembles with m(n) = n/3 and
cp(n) = 2−m(n) can be constructed (see Exercise 22). Then, consider the function
ensemble {h r1 ,...,rm(n) : {0, 1}∗ → {0, 1}2m(n) }n∈N , where all ri ’s are of length n, such
that h r1 ,...,rm(n) (x) is defined as follows:
def
1. As in Construction 6.2.13, break x into t = 2 log2 (|x|/m(n)) consecutive blocks,
denoted x 1 , ..., xt , and let d = log2 t.
def
2. Let i = 1, ..., t, and let yd,i = xi . For j = d − 1, ..., 1, 0 and i = 1, ..., 2 j , let
y j,i = h r j (y j+1,2i−1 y j+1,2i ). The hash value equals (y0,1 , |x|).
def
The above functions have description length N = m(n) · n and map strings of
length at most 2m(n) to strings of length 2m(n). It is easy to bound the collision
probability (for strings of equal length) by the probability of collision occuring
in each of the levels of the tree. In fact, for x1 · · · x t = x1 · · · xt such that xi = xi ,
it suffices to bound the sum of the probabilities that y j, i/2d− j = y j, i/2d− j holds
(given that y j+1, i/2d−( j+1) = y j+1, i/2d−( j+1) ) for j = d − 1, ..., 1, 0. Thus, this gen-
eralized hashing ensemble has a (, )-collision property, where (N ) = 2m(n) and
(N ) = m(n) · cp(n). We stress that the collision probability of the tree-hashing
scheme grows linearly with the depth of the tree (rather than linearly with its size).
Recalling that we may use m(n) = n/3 and cp(n) = 2−m(n) , we obtain (using N =
n 2 /3 = 3m(n)2 ) (N ) = 2(N /3) > 2(N /4) and (N ) < (N /(N )) < 2−(N /4) (as
1/2 1/2 1/2

desired).

Guideline (Part 2): Given a hashing family as in the hypothesis, modify it into
  
{h r,s : {0, 1}2m → {0, 1}m }r,s , where s ∈ {0, 1}m , such that h r,s (02m ) = s, h r,s (sv) =

0m for all v ∈ {0, 1}m , and h r,s (w) = h r (w) for each other w ∈ {0, 1}2m . Note that
the new family maintains the collision probability of the original one up to an
additive term of O(2−m ). On the other hand, for every w ∈ {0, 1}2m , it holds

597
 Free ebooks ==> www.Ebook777.com
DIGITAL SIGNATURES AND MESSAGE AUTHENTICATION

   
that TreeHashr,s (02m w) = h r,s (h r,s (02m ) h r,s (w)) = h r,s (s v) = 0m , where v =

hr,s (w).


Guideline (Part 3): For h r,s as in Part 2 and every v ∈ {0, 1}m , it holds that
  
ChainHashr,s (0 v) = h r,s (h r,s (02m ) v) = h r,s
2m
(sv) = 0m .

Exercise 24: On the additional properties required in Proposition 6.4.21: In contin-

uation of Exercise 23 of Chapter 3, show that the function family Snn−1 presented
there satisfies the following two properties:
1. All but a negligible fraction of the functions in Snn−1 are 2-to-1.
2. There exists a probabilistic polynomial-time algorithm that, given y1 , y2 ∈ {0, 1}n
and z 1 , z 2 ∈ {0, 1}n−1 , outputs a uniformly distributed element of {s ∈ Snn−1 :
h s (yi ) = z i ∀i ∈ {1, 2}}.

Guideline: Recall that each function in Snn−1 is described by a pair of elements of

the finite field GF(2n ), where the pair (a, b) describes the function h a,b that maps
x ∈ GF(2n ) to the (n − 1)-bit prefix of the n-bit representation of ax + b, where
the arithmetic is of the field GF(2n ). The first condition follows by observing that
the function h a,b is 2-to-1 if and only if a = 0. The second condition follows by
observing that h a,b (yi ) = zi if and only if ayi + b = vi for some vi that is a single-
bit extension of zi . Thus, generating a pair (a, b), such that h a,b (yi ) = z i for both
i’s, amounts to selecting random single-bit extensions vi ’s, and (assuming y1 = y2 )
solving the system {ayi + b = vi }i=1,2 (for the variables a and b).

Exercise 25: Fail-stop signatures require a memory-dependent signing process: In

continuation of Section 6.5.5, prove that a secure fail-stop signature scheme must
employ a memory-dependent signing process (as in Definition 6.4.13).

Guideline: Suppose toward the contradiction that there exists a secure memoryless
fail-stop signature scheme. For every signing-key s ∈ {0, 1}n , consider the ran-
domized process Ps in which one first selects uniformly x ∈ {0, 1}n , produces a
(random) signature y ← Ss (x), and outputs the pair (x, y). Show that, given poly-
nomially many samples of Ps , one can find (in exponential time) a string s  ∈ {0, 1}n
such that with probability at least 0.99, the statistical distance between Ps and Ps  is
at most 0.01. Thus, a computationally unbounded adversary making polynomially
many signing queries can find a signing-key that typically produces the same sig-
natures as the true signer. It follows that either these signatures cannot be revoked
or that the user may also revoke its own signatures.

598

www.Ebook777.com
 CHAPTER SEVEN

General Cryptographic Protocols

The design of secure protocols that implement arbitrarily desired functionalities is a

major part of modern cryptography. Taking the opposite perspective, the design of any
cryptographic scheme may be viewed as the design of a secure protocol for implement-
ing a suitable functionality. Still, we believe that it makes sense to differentiate between
basic cryptographic primitives (which involve little interaction) like encryption and
signature schemes, on the one hand, and general cryptographic protocols, on the other
hand.
In this chapter we consider general results concerning secure multi-party computa-
tions, where the two-party case is an important special case. In a nutshell, these results
assert that one can construct protocols for securely computing any desirable multi-party
functionality (see the following terminology). Indeed, what is striking about these re-
sults is their generality, and we believe that the wonder is not diminished by the (various
alternative) conditions under which these results hold.
Our focus on the general study of secure multi-party computation (rather than on
protocols for solving specific problems) is natural in the context of the theoretical
treatment of the subject matter. We wish to highlight the importance of this general
study to practice. Firstly, this study clarifies fundamental issues regarding security in
a multi-party environment. Secondly, it draws the lines between what is possible in
principle and what is not. Thirdly, it develops general techniques for designing secure
protocols. And last, sometimes it may even yield schemes (or modules) that may be
incorporated in practical systems. Thus, we believe that the current chapter is both of
theoretical and practical importance.

Terminology. The notion of a (multi-party) functionality is central to the current chap-

ter. By an m-ary functionality we mean a random process that maps m inputs to m
outputs, where functions mapping m inputs to m outputs are a special case (also re-
ferred to as deterministic functionalities). Thus, functionalities are randomized exten-
sions of ordinary functions. One may think of a functionality F as being a probability
distribution over (corresponding) functions (i.e., F equals the function f (i) with prob-
ability pi ). Alternatively, we think of F(x 1 , ..., x m ) as selecting at random a string r ,
599
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

and outputting F
(r, x 1 , ..., xm ), where F
is a function mapping m + 1 inputs to m

outputs.

Teaching Tip. The contents of the current chapter are quite complex. We suggest
covering in class only the overview section (i.e., Section 7.1), and consider the rest of
this chapter to be advanced material. Furthermore, we assume that the reader is familiar
with the material in all the previous chapters. This familiarity is important, not only
because we use some of the notions and results presented in these chapters but also
because we use similar proof techniques (and do so while assuming that this is not the
reader’s first encounter with these techniques).

Organization. In addition to the overview section (i.e., Section 7.1), the current chapter
consists of two main parts:

The first part (i.e., Sections 7.2–7.4) consists of a detailed treatment of general secure
two-party protocols. Our ultimate goal in this part is to design two-party protocols
that withstand any feasible adversarial behavior. We proceed in two steps. First, we
consider a benign type of adversary, called semi-honest, and construct protocols that
are secure with respect to such an adversary (cf. Section 7.3). Next, we show how to
force parties to behave in a semi-honest manner (cf. Section 7.4). That is, we show
how to transform any protocol, secure in the semi-honest model, into a protocol
that is secure against any feasible adversarial behavior. But before presenting these
constructions, we present the relevant definitions (cf. Section 7.2).
The second part (i.e., Sections 7.5 and 7.6) deals with general secure multi-party pro-
tocols. Specifically, in Section 7.5 we extend the treatment presented in the first part
to multi-party protocols, whereas in Section 7.6 we consider the “private channel
model” and present alternative constructions for it.
Although it is possible to skip some of the earlier sections of this chapter before reading
a later section, we recommend not doing so. In particular, we recommend reading the
overview section (i.e., Section 7.1) before reading any later section.

7.1. Overview

A general framework for casting (m-party) cryptographic (protocol) problems consists

of specifying a random process that maps m inputs to m outputs. The inputs to the
process are to be thought of as local inputs of m parties, and the m outputs are their
corresponding (desired) local outputs. The random process describes the desired func-
tionality. That is, if the m parties were to trust each other (or trust some external party),
then they could each send their local input to the trusted party, who would compute
the outcome of the process and send to each party the corresponding output. A pivotal
question in the area of cryptographic protocols is the extent to which this (imaginary)
trusted party can be “emulated” by the mutually distrustful parties themselves. (See
illustration in Figure 7.1.)
600

www.Ebook777.com
 7.1 OVERVIEW

REAL MODEL IDEAL MODEL

Figure 7.1: Secure protocols emulate a trusted party: an illustration.

The results mentioned previously and surveyed later describe a variety of models in
which such an “emulation” is possible. The models vary by the underlying assumptions
regarding the communication channels, the numerous parameters relating to the extent
of adversarial behavior, and the desired level of emulation of the trusted party (i.e.,
level of “security”). We stress that unless stated differently, the two-party case is an
important special case of the treatment of the multi-party setting (i.e., we consider any
m ≥ 2).

7.1.1. The Definitional Approach and Some Models

Before describing the abovementioned results, we further discuss the notion of “emu-
lating a trusted party,” which underlies the definitional approach to secure multi-party
computation. The approach can be traced back to the definition of zero-knowledge (see
Section 4.3 of Volume 1), and even to the definition of semantic security (see Sec-
tion 5.2.1). The underlying paradigm (called the simulation paradigm) is that a scheme
is secure if whatever a feasible adversary can obtain after attacking it is also feasibly
attainable in an “ideal setting.” In the case of zero-knowledge, this amounts to say-
ing that whatever a (feasible) verifier can obtain after interacting with the prover on a
prescribed valid assertion can be (feasibly) computed from the assertion itself. In the
case of multi-party computation, we compare the effect of adversaries that participate
in the execution of the actual protocol to the effect of adversaries that participate in
an imaginary execution of a trivial (ideal) protocol for computing the desired func-
tionality with the help of a trusted party. If whatever adversaries can feasibly obtain
in the former real setting can also be feasibly obtained in the latter ideal setting, then
the protocol “emulates the ideal setting” (i.e., “emulates a trusted party”), and so is
deemed secure. This means that properties that are satisfied in the ideal setting are
also satisfied by a secure protocol that is executed in the real setting. For example,
security typically implies the preservation of the privacy of the parties’ local inputs
(beyond whatever is revealed by the local outputs provided to the adversary), and
601
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

the correctness of the honest parties’ local outputs (i.e., their consistency with the
functionality).
The approach outlined here can be applied in a variety of models, and is used to
define the goals of security in these models.1 We first discuss some of the parameters
used in defining various models, and next demonstrate the application of this approach
to a couple of important cases (cf. Sections 7.1.1.2 and 7.1.1.3).

7.1.1.1. Some Parameters Used in Defining Security Models

The following parameters are described in terms of the actual (or real) computation.
In some cases, the corresponding definition of security is obtained by some restric-
tions or provisions applied to the ideal model. In all cases, the desired notion of se-
curity is defined by requiring that for any adequate adversary in the real model, there
exists a corresponding adversary in the corresponding ideal model that obtains es-
sentially the same impact on the computation of the functionality (as the real-model
adversary).
r Set-up assumptions: Unless differently stated, we make no set-up assumptions (ex-
cept for the obvious assumption that all parties have copies of the protocol’s program).
However, in some cases it is assumed that each party knows some information (e.g.,
a verification-key) corresponding to each of the other parties (or, one may assume
the existence of a public-key infrastructure). Another assumption, made more rarely,
is that all parties have access to some common (trusted) random string.
r The communication channels: Here we refer to issues like the privacy and reliablity
of data sent over the channels, as well as to the availability and the communication
features of the channels.
The standard assumption in the area is that the adversary can tap all communication
channels (between honest parties); that is, the channels per se do not provide privacy
(i.e., privacy of the data sent over them). In contrast, one may postulate that the ad-
versary cannot obtain messages sent between honest parties, yielding the so-called
private-channel model. This postulate may be justified in some settings. Further-
more, it may be viewed as a useful abstraction that provides a clean model for study
and development of secure protocols. In this respect, it is important to mention that
in a variety of settings of the other parameters, the private-channel model can be
emulated by ordinary (i.e., “tapped” point-to-point) channels.
The standard assumption in the area is that the adversary cannot omit, modify, dupli-
cate, or generate messages sent over the communication channels (between honest

1 A few technical comments are in place. Firstly, we assume that the inputs of all parties are of the same length.
We comment that as long as the lengths of the inputs are polynomially related, this convention can be enforced
by padding. On the other hand, some length restriction is essential for the security results, because (in general)
it is impossible to hide all information regarding the length of the inputs to a protocol. Secondly, we assume that
the desired functionality is computable in probabilistic polynomial-time, because we wish the secure protocol to
run in probabilistic polynomial-time (and a protocol cannot be more efficient than the corresponding centralized
algorithm). Clearly, the results can be extended to functionalities that are computable within any given (time-
constructible) time bound, using adequate padding.

602

www.Ebook777.com
 7.1 OVERVIEW

parties); that is, the channels are postulated to be reliable (in the sense that they guar-
antee the authenticity of the data sent over them). Furthermore, one may postulate
the existence of a broadcast channel. Again, these assumptions can be justified in
some settings and emulated in others.
Most work in the area assumes that communication is synchronous and that point-
to-point channels exist between every pair of processors. However, one may also
consider asynchronous communication and arbitrary networks of point-to-point
channels.
r Computational limitations: Typically, we consider computationally bounded adver-
saries (e.g., probabilistic polynomial-time adversaries). However, the private-channel
model also allows for (meaningful) consideration of computationally unbounded ad-
versaries.
We stress that, also in the latter case, security should be defined by requiring
that for every real adversary, whatever the adversary can compute after partici-
pating in the execution of the actual protocol be computable within comparable
time (e.g., in polynomially related time) by an imaginary adversary participating
in an imaginary execution of the trivial ideal protocol (for computing the desired
functionality with the help of a trusted party). Thus, results in the computationally
unbounded–adversary model trivially imply results for computationally bounded
adversaries.
r Restricted adversarial behavior: The most general type of an adversary considered
in the literature is one that may corrupt parties to the protocol while the execution
goes on, and decide which parties to corrupt based on partial information it has
gathered so far. A somewhat more restricted model, which seems adequate in many
settings, postulates that the set of dishonest parties is fixed (arbitrarily) before the
execution starts (but this set is, of course, not known to the honest parties). The latter
model is called non-adaptive as opposed to the adaptive adversary mentioned
first.
An orthogonal parameter of restriction refers to whether a dishonest party takes
active steps to disrupt the execution of the protocol (i.e., sends messages that dif-
fer from those specified by the protocol), or merely gathers information (which it
may later share with the other dishonest parties). The latter adversary has been
given a variety of names, such as semi-honest, passive, and honest-but-curious.
This restricted model may be justified in certain settings, and certainly provides a
useful methodological locus (cf. Section 7.1.3). In the following, we refer to the
adversary of the unrestricted model as active; another commonly used name is
malicious.
r Restricted notions of security: One example is the willingness to tolerate “unfair”
protocols in which the execution can be suspended (at any time) by a dishonest
party, provided that it is detected doing so. We stress that in case the execution is
suspended, the dishonest party does not obtain more information than it could have
obtained if the execution were not suspended. What may happen is that some honest
parties will not obtain their desired outputs (although other parties did obtain their
corresponding outputs), but will rather detect that the execution was suspended. We

603
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

will say that this restricted notion of security allows abort (or allows premature
suspension of the execution).
r Upper bounds on the number of dishonest parties: In some models, secure multi-party
computation is possible only if a strict majority of the parties are honest.2 Some-
times even a special majority (e.g., 2/3) is required. General “resilient adversary-
structures” have been considered, too (i.e., security is guaranteed in the case that the
set of dishonest parties equals one of the sets specified in a predetermined family
of sets).
r Mobile adversary: In most works, once a party is said to be dishonest it remains
so throughout the execution. More generally, one may consider transient adversarial
behavior (e.g., an adversary seizes control of some site and later withdraws from
it). This model, which will not be further discussed in this work, allows for the
construction of protocols that remain secure, even in case the adversary may seize
control of all sites during the execution (but never control concurrently, say, more than
10 percent of the sites). We comment that schemes secure in this model were later
termed “proactive.”

In the rest of this chapter we will consider a few specific settings of these parameters.
Specifically, we will focus on non-adaptive, active, and computationally bounded ad-
versaries, and will not assume the existence of private channels. In Section 7.1.1.2 we
consider this setting while restricting the dishonest parties to a strict minority, whereas
in Section 7.1.1.3 we consider a restricted notion of security for two-party protocols
that allows “unfair suspension” of execution (or “allows abort”).

7.1.1.2. Example: Multi-Party Protocols with Honest Majority

We consider a non-adaptive, active, computationally bounded adversary, and do not
assume the existence of private channels. Our aim is to define multi-party protocols
that remain secure provided that the honest parties are in the majority. (The reason for
requiring an honest majority will be discussed at the end of this subsection.) For more
details about this model, see Section 7.5.1.
Consider any multi-party protocol. We first observe that each party may change its
local input before even entering the execution of the protocol. Furthermore, this is also
unavoidable when the parties utilize a trusted party Consequently, such an effect of the
adversary on the real execution (i.e., modification of its own input prior to entering the
actual execution) is not considered a breach of security. In general, whatever cannot
be avoided (even) when the parties utilize a trusted party is not considered a breach
of security. We wish secure protocols (in the real model) to suffer only from whatever
is also unavoidable when the parties utilize a trusted party. Thus, the basic paradigm
underlying the definitions of secure multi-party computations amounts to saying that
the only situations that may occur in the real execution of a secure protocol are those that
can also occur in a corresponding ideal model (where the parties may employ a trusted

2 Indeed, requiring an honest majority in the two-party case yields a meaningless model.

604

www.Ebook777.com
 7.1 OVERVIEW

party). In other words, the “effective malfunctioning” of parties in secure protocols is

restricted to what is postulated in the corresponding ideal model.
When defining secure multi-party protocols (with honest majority), we need to pin-
point what cannot be avoided in the ideal model (i.e., when the parties utilize a trusted
party). This is easy, because the ideal model is very simple. Since we are interested in
executions in which the majority of parties are honest, we consider an ideal model in
which any minority group (of the parties) may collude as follows:
1. Firstly, the dishonest minority parties share their original inputs and decide together
on replaced inputs to be sent to the trusted party. (The other parties send their
respective original inputs to the trusted party. We stress that the communication
between the honest parties and the trusted party is not seen by the dishonest colluding
minority parties.)
2. Upon receiving inputs from all parties, the trusted party determines the corresponding
outputs and sends them to the corresponding parties.
3. Upon receiving the “output message” from the trusted party, each honest party outputs
it locally, whereas the dishonest colluding minority parties may determine their
outputs based on all they know (i.e., their initial inputs and their received outputs).
Note that such behavior of the minority group is unavoidable in any execution of any
protocol (even in the presence of trusted parties). This is the reason that the ideal model
was so defined. Now, a secure multi-party computation with honest majority is required
to emulate this ideal model. That is, the effect of any feasible adversary that controls
a minority of the parties in a real execution of the actual protocol can be essentially
simulated by a (different) feasible adversary that controls the same parties in the ideal
model.

Definition 7.1.1 (secure protocols – a sketch): Let f be an m-ary functionality and

be an m-party protocol operating in the real model.
r For a real-model adversary A, controlling some minority of the parties (and tapping
all communication channels), and an m-sequence x, we denote by real
, A (x) the
sequence of m outputs resulting from the execution of
on input x under attack of
the adversary A.
r For an ideal-model adversary A
, controlling some minority of the parties, and an
m-sequence x, we denote by ideal f, A
(x) the sequence of m outputs resulting from
the ideal process described previously, on input x under attack of the adversary A
.
We say that
securely implements f with honest majority if for every feasible real-
model adversary A, controlling some minority of the parties, there exists a feasible ideal-
model adversary A
, controlling the same parties, so that the probability ensembles
{real
, A (x)}x and {ideal f, A
(x)}x are computationally indistinguishable (as in Part 2
of Definition 3.2.7 in Volume 1).

Thus, security means that the effect of each minority group in a real execution of a secure
protocol is “essentially restricted” to replacing its own local inputs (independently of
the local inputs of the majority parties) before the protocol starts, and replacing its
605
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

own local outputs (depending only on its local inputs and outputs) after the protocol
terminates. (We stress that in the real execution, the minority parties do obtain additional
pieces of information; yet in a secure protocol they gain nothing from these additional
pieces of information.)
The fact that Definition 7.1.1 refers to a model without private channels is reflected
in the set of possible ensembles {real
, A (x)}x that is determined by the (sketchy)
definition of the real-model adversary (which is allowed to tap all the communication
channels). When defining security in the private-channel model, the real-model adver-
sary is not allowed to tap channels between honest parties, which in turn restricts the
set of possible ensembles {real
, A (x)}x . Thus, the difference between the two models
is only reflected in the definition of the real-model adversary. On the other hand, when
we wish to define security with respect to passive adversaries, both the scope of the
real-model adversaries and the scope of the ideal-model adversaries change. In the real-
model execution, all parties follow the protocol, but the adversary may alter the output
of the dishonest parties arbitrarily, depending on all their intermediate internal states
(during the execution). In the corresponding ideal-model, the adversary is not allowed
to modify the inputs of dishonest parties (in Step 1), but is allowed to modify their
outputs (in Step 3).
We comment that a definition analogous to Definition 7.1.1 can also be presented
in case the dishonest parties are not in the minority. In fact, such a definition seems
more natural, but the problem is that such a definition cannot be satisfied. That is,
most natural functionalities do not have a protocol for computing them securely in case
at least half of the parties are dishonest and employ an adequate (active) adversarial
strategy. This follows from an impossibility result regarding two-party computation,
which essentially asserts that there is no way to prevent a party from prematurely
suspending the execution. On the other hand, secure multi-party computation with
dishonest majority is possible if (and only if) premature suspension of the execution is
not considered a breach of security.

7.1.1.3. Another Example: Two-Party Protocols Allowing Abort

In light of the last paragraph, we now consider multi-party computations in which
premature suspension of the execution is not considered a breach of security. For con-
creteness, we focus here on the special case of two-party computations.3 For more
details about this model, see Section 7.2.3.
Intuitively, in any two-party protocol, each party may suspend the execution at any
point in time, and furthermore, it may do so as soon as it learns the desired output.
Thus, in case the output of each parties depends on both inputs, it is always possible
for one of the parties to obtain the desired output while preventing the other party from
fully determining its own output. The same phenomenon occurs even in case the two
parties just wish to generate a common random value. Thus, when considering active
adversaries in the two-party setting, we do not consider such premature suspension of
the execution a breach of security. Consequently, we consider an ideal model where

3 As in Section 7.1.1.2, we consider a non-adaptive, active, computationally bounded adversary.

606

www.Ebook777.com
 7.1 OVERVIEW

each of the two parties may “shut down” the trusted (third) party at any point in time.
In particular, this may happen after the trusted party has supplied the outcome of the
computation to one party but before it has supplied it to the second. That is, an execution
in the ideal model proceeds as follows:

1. Each party sends its input to the trusted party, where the dishonest party may replace
its input or send no input at all (which may be viewed as aborting).
2. Upon receiving inputs from both parties, the trusted party determines the corre-
sponding outputs and sends the first output to the first party.
3. In case the first party is dishonest, it may instruct the trusted party to halt; otherwise
it always instructs the trusted party to proceed. If instructed to proceed, the trusted
party sends the second output to the second party.
4. Upon receiving the output message from the trusted party, the honest party outputs
it locally, whereas the dishonest party may determine its output based on all it knows
(i.e., its initial input and its received output).

A secure two-party computation allowing abort is required to emulate this ideal

model. That is, as in Definition 7.1.1, security is defined by requiring that for every
feasible real-model adversary A, there exists a feasible ideal-model adversary A
, con-
trolling the same party, so that the probability ensembles representing the corresponding
(real and ideal) executions are computationally indistinguishable. This means that each
party’s “effective malfunctioning” in a secure protocol is restricted to supplying an ini-
tial input of its choice and aborting the computation at any point in time. (Needless to
say, the choice of the initial input of each party may not depend on the input of the other
party.)
We mention that an alternative way of dealing with the problem of premature sus-
pension of execution (i.e., abort) is to restrict attention to single-output functionalities,
that is, functionalities in which only one party is supposed to obtain an output. The defi-
nition of secure computation of such functionalities can be identical to Definition 7.1.1,
with the exception that no restriction is made on the set of dishonest parties (and, in
particular, one may consider a single dishonest party in the case of two-party protocols).
For further details, see Section 7.2.3.2.

7.1.2. Some Known Results

We briefly mention some of the models for which general secure multi-party com-
putation is known to be attainable; that is, models in which one can construct secure
multi-party protocols for computing any desired functionality.

7.1.2.1. The Main Results Presented in This Chapter

We start with results that refer to secure two-party protocols, as well as to secure multi-
party protocols in the standard model (where the adversary may tap the communication
lines).
607
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Theorem 7.1.2 (the main feasibility results – a sketch): Assuming the existence of
enhanced trapdoor permutations (as in Definition C.1.1 in Appendix C), general secure
multi-party computation is possible in the following three models:
1. Passive adversary, for any number of dishonest parties.
2. Active adversary that may control only a strict minority of the parties.
3. Active adversary, controlling any number of bad parties, provided that suspension
of execution is not considered a violation of security.
In all these cases, the adversary is computationally bounded and non-adaptive. On the
other hand, the adversary may tap the communication lines between honest parties (i.e.,
we do not assume the existence of private channels). The results for active adversaries
assume a broadcast channel.

Recall that a broadcast channel can be implemented (while tolerating any num-
ber of bad parties) using a signature scheme and assuming a public-key infrastruc-
ture (i.e., each party knows the verification-key corresponding to each of the other
parties).4
Most of the current chapter will be devoted to proving Theorem 7.1.2. In Sections 7.3
and 7.4 we prove Theorem 7.1.2 for the special case of two parties: In that case, Part 2
is not relevant, Part 1 is proved in Section 7.3, and Part 3 is proved in Section 7.4. The
general case (i.e., of multi-party computation) is treated in Section 7.5.

7.1.2.2. Other Results

We next list some other models in which general secure multi-party computation is
attainable:
r Making no computational assumptions and allowing computationally unbounded
adversaries, but assuming the existence of private channels, general secure multi-
party computation is possible in the following models:
1. Passive adversary that may control only a (strict) minority of the parties.
2. Active adversary that may control only less than one third of the parties. (Fault-
tolerance can be increased to a regular minority if a broadcast channel exists.)
In both cases the adversary may be adaptive. For details, see Sections 7.6 and 7.7.1.2.
r General secure multi-party computation is possible against an active adaptive, and
mobile adversary that may control a small constant fraction of the parties at any point
in time. This result makes no computational assumptions, allows computationally
unbounded adversaries, but assumes the existence of private channels.

4 Note that the implementation of a broadcast channel can be cast as a cryptographic protocol problem (i.e., for
the functionality (v, λ, ..., λ) → (v, v, ..., v), where v ∈ {0, 1}∗ and λ denotes the empty string). Thus, it is not
surprising that the results regarding active adversaries assume the existence of either such a channel or a setting
in which such a channel can be implemented (e.g., either that less than a third of the parties are faulty or that a
public-key infrastructure exists). (This reasoning fails if the definition of secure protocols is relaxed such that
it does not imply agreement; see [122].)

608

www.Ebook777.com
 7.1 OVERVIEW

r Assuming the intractability of inverting RSA (or of the DLP), general secure
multi-party computation is possible in a model allowing an adaptive and active
computationally bounded adversary that may control only less than one third of the
parties. We stress that this result does not assume the existence of private channels.
Results for asynchronous communication and arbitrary networks of point-to-point chan-
nels are also known. For further details, see Section 7.7.5.

7.1.2.3. An Extension and Efficiency Considerations

Secure Reactive Computation. All the aforementioned results extend (easily) to a

reactive model of computation in which each party interacts with a high-level process
(or application). The high-level process adaptively supplies each party with a sequence
of inputs, one at a time, and expects to receive corresponding outputs from the parties.
That is, a reactive system goes through (a possibly unbounded number of) iterations of
the following type:
r Parties are given inputs for the current iteration.
r Depending on the current inputs, the parties are supposed to compute outputs for the
current iteration. That is, the outputs in iteration j are determined by the inputs of
the j-th iteration.
A more general formulation allows the outputs of each iteration to depend also on a
global state, which is possibly updated at each iteration. The global state may include all
inputs and outputs of previous iterations, and may only be partially known to individual
parties. (In a secure reactive computation, such a global state may be maintained by all
parties in a “secret sharing” manner.) For further discussion, see Section 7.7.1.3.

Efficiency considerations. One important efficiency measure regarding protocols is

the number of communication rounds in their execution. The results mentioned earlier
were originally obtained using protocols that use an unbounded number of rounds. In
some cases, subsequent works obtained secure constant-round protocols. Other im-
portant efficiency considerations include the total number of bits sent in the execution
of a protocol and the local computation time. The communication and computation
complexities of the aforementioned protocols are related to the computational com-
plexity of the desired functionalities, but alternative relations (e.g., referring to the
communication complexity of the latter) may be possible.

7.1.3. Construction Paradigms

We briefly sketch three paradigms used in the construction of secure multi-party pro-
tocols. We focus on the construction of secure protocols for the model of computa-
tionally bounded and non-adaptive adversaries. These constructions proceed in two
steps: First, a secure protocol is presented for the model of passive adversaries (for
any number of dishonest parties), and next, such a protocol is “compiled” into a
609
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

protocol that is secure in one of the two models of active adversaries (i.e., either
in a model allowing the adversary to control only a minority of the parties or in a
model in which premature suspension of the execution is not considered a violation of
security).
Recall that in the model of passive adversaries, all parties follow the prescribed
protocol, but at termination, the adversary may alter the output of the dishonest parties
depending on all their intermediate internal states (during the execution). In the fol-
lowing, we refer to protocols that are secure in the model of passive (resp., general or
active) adversaries by the term passively secure (resp., actively secure).

7.1.3.1. From Passively Secure Protocols to Actively Secure Ones

We show how to transform any passively secure protocol into a corresponding ac-
tively secure protocol. The communication model in both protocols consists of a single
broadcast channel. Note that the messages of the original (passively secure) protocol
may be assumed to be sent over a broadcast channel, because the adversary may see
them anyhow (by tapping the point-to-point channels). As for the resulting actively se-
cure protocol, the broadcast channel it uses can be implemented via an (authenticated)
Byzantine Agreement protocol (cf. Section 7.5.3.2), thus providing an emulation of
this model on the standard point-to-point model (in which a broadcast channel does not
exist). We mention that authenticated Byzantine Agreement is typically implemented
using a signature scheme (and assuming that each party knows the verification-key
corresponding to each of the other parties).
Turning to the transformation itself, the main idea is to use zero-knowledge proofs
in order to force parties to behave in a way that is consistent with the (passively secure)
protocol. Actually, we need to confine each party to a unique consistent behavior (i.e.,
according to some fixed local input and a sequence of coin tosses), and to guarantee
that a party cannot fix its input (and/or its coins) in a way that depends on the inputs of
honest parties. Thus, some preliminary steps have to be taken before the step-by-step
emulation of the original protocol can take place. Specifically, the compiled protocol
(which, like the original protocol, is executed over a broadcast channel) proceeds as
follows:

1. Prior to the emulation of the original protocol, each party commits to its input (using
a commitment scheme). In addition, using a zero-knowledge proof-of-knowledge
(cf. Section 4.7 of Volume 1), each party also proves that it knows its own in-
put, that is, that it can properly decommit to the commitment it sent. (These
zero-knowledge proofs-of-knowledge are conducted sequentially to prevent dis-
honest parties from setting their inputs in a way that depends on inputs of honest
parties.)
2. Next, all parties jointly generate a sequence of random bits for each party such that
only this party knows the outcome of the random sequence generated for it, but
everybody gets a commitment to this outcome. These sequences will be used as the
random-inputs (i.e., sequence of coin tosses) for the original protocol. Each bit in
the random-sequence generated for Party X is determined as the exclusive-or of the
610

www.Ebook777.com
 7.1 OVERVIEW

outcomes of instances of an (augmented) coin-tossing protocol that Party X plays

with each of the other parties.
3. In addition, when compiling (the passively secure protocol to an actively secure
protocol) for the model that allows the adversary to control only a minority of the
parties, each party shares its input and random-input with all other parties using
a Verifiable Secret Sharing protocol (cf. Section 7.5.5). This will guarantee that if
some party prematurely suspends the execution, then all the parties can together
reconstruct all its secrets and carry on the execution while playing its role.
4. After all these steps have been completed, we turn to the main step in which the new
protocol emulates the original one. In each step, each party augments the message
determined by the original protocol with a zero-knowledge that asserts that the mes-
sage was indeed computed correctly. Recall that the next message (as determined
by the original protocol) is a function of the sender’s own input, its random-input,
and the messages it has received so far (where the latter are known to everybody
because they were sent over a broadcast channel). Furthermore, the sender’s input
is determined by its commitment (as sent in Step 1), and its random-input is simi-
larly determined (in Step 2). Thus, the next message (as determined by the original
protocol) is a function of publicly known strings (i.e., the said commitments as well
as the other messages sent over the broadcast channel). Moreover, the assertion that
the next message was indeed computed correctly is an NP-assertion, and the sender
knows a corresponding NP-witness (i.e., its own input and random-input, as well as
the corresponding decommitment information). Thus, the sender can prove (to each
of the other parties) in zero-knowledge that the message it is sending was indeed
computed according to the original protocol.
A detailed description is provided in Section 7.4 (see also Section 7.5.4).

7.1.3.2. Passively Secure Computation with “Scrambled Circuits”

The following technique refers mainly to two-party computation. Suppose that two
parties, each having a private input, wish to obtain the value of a predetermined two-
argument function evaluated at their corresponding inputs, that is, we consider only
functionalities of the form (x, y) → ( f (x, y), f (x, y)). Further suppose that the two
parties hold a circuit that computes the value of the function on inputs of the adequate
length. The idea is to have one party construct a “scrambled” form of the circuit so
that the other party can propagate encrypted values through the “scrambled gates” and
obtain the output in the clear (while all intermediate values remain secret). Note that the
roles of the two parties are not symmetric, and recall that we are describing a protocol
that is secure (only) with respect to passive adversaries. An implementation of this idea
proceeds as follows:
r Constructing a “scrambled” circuit: The first party constructs a scrambled form
of the original circuit. The scrambled circuit consists of pairs of encrypted secrets
that correspond to the wires of the original circuit and gadgets that correspond to
the gates of the original circuit. The secrets associated with the wires entering a
gate are used (in the gadget that corresponds to this gate) as keys in the encryption
611
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

of the secrets associated with the wire exiting this gate. Furthermore, there is a
random correspondence between each pair of secrets and the Boolean values (of the
corresponding wire). That is, wire w is assigned a pair of secrets, denoted (sw
, sw

),
and there is a random 1-1 mapping, denoted νw , between this pair and the pair of
Boolean values (i.e., {νw (sw
), νw (sw

)} = {0, 1}).
Each gadget is constructed such that knowledge of a secret that corresponds to
each wire entering the corresponding gate (in the circuit) yields a secret corre-
sponding to the wire that exits this gate. Furthermore, the reconstruction of se-
crets using each gadget respects the functionality of the corresponding gate. For
example, if one knows the secret that corresponds to the 1-value of one entry-wire
and the secret that corresponds to the 0-value of the other entry-wire, and the gate
is an or-gate, then one obtains the secret that corresponds to the 1-value of the
exit-wire.
Specifically, each gadget consists of four templates that are presented in a random
order, where each template corresponds to one of the four possible values of the
two entry-wires. A template may be merely a double encryption of the secret that
corresponds to the appropriate output value, where the double encryption uses as
keys the two secrets that correspond to the input values. That is, suppose a gate
computing f : {0, 1}2 → {0, 1} has input wires w 1 and w 2 , and output wire w 3 .
Then, each of the four templates of this gate has the form E sw 1 (E sw 2 (sw 3 )), where
f (νw 1 (sw 1 ), νw 2 (sw 2 )) = νw 3 (sw 3 ).
r Sending the “scrambled” circuit: The first party sends the scrambled circuit to the
second party. In addition, the first party sends to the second party the secrets that
correspond to its own (i.e., the first party’s) input bits (but not the values of these bits).
The first party also reveals the correspondence between the pair of secrets associated
with each output (i.e., circuit-output wire) and the Boolean values.5 We stress that
the random correspondence between the pair of secrets associated with each other
wire and the Boolean values is kept secret (by the first party).
r Oblivious Transfer of adequate secrets: Next, the first party uses a (1-out-of-2) Obliv-
ious Transfer protocol in order to hand the second party the secrets corresponding
to the second party’s input bits (without the first party learning anything about these
bits).
Loosely speaking, a 1-out-of-k Oblivious Transfer is a protocol enabling one party to
obtain one of k secrets held by another party, without the second party learning which
secret was obtained by the first party. That is, we refer to the two-party functionality

(i, (s1 , ..., sk )) → (si , λ) (7.1)

where λ denotes the empty string.

r Locally evaluating the “scrambled” circuit: Finally, the second party “evaluates” the
scrambled circuit gate-by-gate, starting from the top (circuit-input) gates (for which
it knows one secret per each wire) and ending at the bottom (circuit-output) gates

5 This can be done by providing, for each output wire, a succinct 2-partition (of all strings) that separates the two
secrets associated with this wire.

612

www.Ebook777.com
 7.1 OVERVIEW

(for which, by construction, the correspondence of secrets to values is known). Thus,

the second party obtains the output value of the circuit (but nothing else), and sends
it to the first party.
For further details, see Section 7.7.5.

7.1.3.3. Passively Secure Computation with Shares

For any m ≥ 2, suppose that m parties, each having a private input, wish to obtain the
value of a predetermined m-argument function evaluated at their sequence of inputs.
Further suppose that the parties hold a circuit that computes the value of the function
on inputs of the adequate length, and that the circuit contains only and- and not-
gates. Again, the idea is to propagate information from the top (circuit-input) gates
to the bottom (circuit-output) gates, but this time the information is different, and the
propagation is done jointly by all parties. The idea is to share the value of each wire
in the circuit such that all shares yield the value, whereas lacking even one of the
shares keeps the value totally undetermined. That is, we use a simple secret-sharing
scheme such that a bit b is shared by a random sequence of m bits that sum up to
b mod 2. First, each party shares each of its input bits with all parties (by sending
each party a random value and setting its own share accordingly).6 Next, all parties
jointly scan the circuit from its input wires to the output wires, processing each gate as
follows:
r When encountering a gate, the parties already hold shares of the values of the wires
entering the gate, and their aim is to obtain shares of the value of the wire exiting
the gate.
r For a not-gate, propagating shares through the gate is easy: The first party just flips
the value of its share, and all other parties maintain their shares.
r For an and-gate, propagating shares through the gate requires a secure (i.e., passively
secure) multi-party protocol. Since an and-gate corresponds to multiplication mod-
ulo 2, the parties need to securely compute the following randomized functionality
(in which the xi ’s denote shares of one entry-wire, the yi ’s denote shares of the second
entry-wire, the z i ’s denote shares of the exit-wire, and the shares indexed by i belong
to Party i):
((x 1 , y1 ), ..., (x m , ym )) → (z 1 , ..., z 2 ) (7.2)
where

m

m

m
zi = xi · yi (7.3)
i=1 i=1 i=1

That is, the z i ’s are random subject to Eq. (7.3).

At the end, each party holds a share of each output wire. The desired output is ob-
tained by letting each party send its share of each output wire to all parties. Thus,

6 For simplicity, we may assume the private-channel model, in which case a value sent to an honest party cannot
be read by the adversary.

613
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

securely evaluating the entire (arbitrary) circuit “reduces” to securely conducting a

specific (very simple) multi-party computation. But things get even simpler: the key
observation is that
   

m
m
m
 
xi · yi = xi yi + xi y j + x j yi (7.4)
i=1 i=1 i=1 1≤i< j≤m

Thus, the m-ary functionality of Eq. (7.2) and Eq. (7.3) can be computed as follows
(where all arithmetic operations are mod 2):
def
1. Each Party i locally computes z i,i = xi yi .
2. Next, each pair of parties (i.e., Parties i and j) securely compute random shares of
xi y j + x j yi . That is, Parties i and j (holding (xi , yi ) and (x j , y j ), respectively), need
to securely compute the randomized two-party functionality ((xi , yi ), (x j , y j )) →
(z i, j , z j,i ), where the z’s are random subject to z i, j + z j,i = xi y j + yi x j . The latter
(simple) two-party computation can be securely implemented using (a 1-out-of-4)
Oblivious Transfer. Specifically, Party i uniformly selects z i, j ∈ {0, 1}, and defines
its four secrets as follows:

Index of Corresponding Value of the secret

the secret value of (x j , y j ) (output of Party j)
1 (0, 0) z i, j
2 (0, 1) z i, j + xi
3 (1, 0) z i, j + yi
4 (1, 1) z i, j + xi + yi

Party j sets its input to 2x j + y j + 1, and obtains the secret z i, j + xi y j + yi x j .

(Indeed, for “small” B, any two-party functionality f : A × B → {λ} × {0, 1} can
be securely implemented by a single invocation of a 1-out-of-|B| Oblivious Transfer,
where the first party defines its |B| secrets in correspondence to the |B| possible
values of the input to the second party.)
3. Finally, for every i = 1, ..., m, the sum mj=1 z i, j yields the desired share of Party i.
A detailed description is provided in Section 7.3 (see also Section 7.5.2).

A related construction. We mention that an analogous construction has been subse-

quently used in the private-channel model and withstands computationally unbounded
active (resp., passive) adversaries that control less than one third (resp., a minority)
of the parties. The basic idea is to use a more sophisticated secret-sharing scheme,
specifically, via low-degree polynomials. That is, the Boolean circuit is viewed as an
arithmetic circuit over a finite field having more than m elements, and a secret element
s in the field is shared by selecting uniformly a polynomial of degree d = (m − 1)/3
(resp., degree d = (m − 1)/2 ) having a free-term equal to s, and handing each party
the value of this polynomial evaluated at a different (fixed) point (e.g., party i is
given the value at point i). Addition is emulated by (local) pointwise addition of the
614

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

(secret-sharing) polynomials representing the two inputs (using the fact that for poly-
nomials p and q, and any field element e [and in particular e = 0, 1, ..., m], it holds
that p(e) + q(e) = ( p + q)(e)). The emulation of multiplication is more involved and
requires interaction (because the product of polynomials yields a polynomial of higher
degree, and thus the polynomial representing the output cannot be the product of the
polynomials representing the two inputs). Indeed, the aim of the interaction is to turn
the shares of the product polynomial into shares of a degree d polynomial that has the
same free-term as the product polynomial (which is of degree 2d). This can be done
using the fact that the coefficients of a polynomial are a linear combination of its values
at sufficiently many arguments (and the other way around), and the fact that one can
privately compute any linear combination (of secret values). For further details, see
Section 7.6.

7.2.* The Two-Party Case: Definitions

In this section we define security for two models of adversaries for two-party proto-
cols. In both models, the adversary is non-adaptive and computationally bounded (i.e.,
restricted to probabilistic polynomial-time with [non-uniform] auxiliary inputs). In the
first model, presented in Section 7.2.2, we consider a restricted adversary called semi-
honest, whereas the general case of malicious adversary is considered in Section 7.2.3.
In addition to being of independent interest, the semi-honest model will play a major
role in the constructions of protocols for the malicious model (see Sections 7.3 and 7.4).

7.2.1. The Syntactic Framework

A two-party protocol problem is cast by specifying a random process that maps pairs
of inputs (one input per each party) to pairs of outputs (one per each party). We refer to
such a process as the desired functionality, denoted f : {0, 1}∗ × {0, 1}∗ → {0, 1}∗ ×
{0, 1}∗ . That is, for every pair of inputs (x, y), the desired output pair is a random
variable, f (x, y), ranging over pairs of strings. The first party, holding input x, wishes
to obtain the first element in f (x, y); whereas the second party, holding input y, wishes
to obtain the second element in f (x, y). A few interesting special cases are highlighted
next:
r Symmetric deterministic functionalities: This is the simplest general case often con-
sidered in the literature. In this case, for some predetermined function, g, both parties
wish to obtain the value of g evaluated at the input pair. That is, the functionality
def
they wish to (securely) compute is f (x, y) = (g(x, y), g(x, y)). For example, they
may be interested in determining whether their local inputs are equal (i.e., g(x, y) =
1 iff x = y) or whether their local inputs viewed as sets are disjoint (i.e., g(x, y) =
1 iff for every i either x i = 0 or yi = 0).
r Input-oblivious randomized functionalities: Whereas input-oblivious deterministic
functionalities are trivial, some input-oblivious randomized functionalities are very
interesting. Suppose, for example, that the two parties wish to toss a fair coin (i.e.,
615
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

such that no party can “influence the outcome” by itself). This task can be cast
by requiring that for every input pair (x, y), the output pair f (x, y) is uniformly
distributed over {(0, 0), (1, 1)}.
r Asymmetric functionalities: The general case of asymmetric functionalities is cap-
def
tured by functionalities of the form f (x, y) = ( f
(x, y), λ), where f
: {0, 1}∗ ×
∗ ∗
{0, 1} → {0, 1} is a randomized process and λ denotes the empty string. A spe-
cial case of interest is when one party wishes to obtain some predetermined partial
information regarding the secret input of the other party, where the latter secret
is verifiable with respect to the input of the first party. This task is captured by a
def def
functionality f such that f (x, y) = (R(y), λ) if V (x, y) = 1 and f (x, y) = (⊥, λ)
otherwise, where R represents the partial information to be revealed and V represents
the verification procedure.7
We stress that whenever we consider a protocol for securely computing f , it is implicitly
assumed that the protocol correctly computes f when both parties follow the prescribed
program. That is, the joint output distribution of the protocol, played by honest parties,
on input pair (x, y), equals the distribution of f (x, y).

Notation. We let λ denote the empty string and ⊥ denote a special error symbol.
That is, whereas λ ∈ {0, 1}∗ (and |λ| = 0), we postulate that ⊥ ∈ {0, 1}∗ (and is thus
distinguishable from any string in {0, 1}∗ ).

7.2.1.1. Simplifying Conventions

To simplify the exposition, we make the following three assumptions:

1. The protocol problem has to be solved only for inputs of the same length (i.e.,
|x| = |y|).
2. The functionality is computable in time polynomial in the length of the inputs.
3. Security is measured in terms of the length of the inputs.

As discussed next, these conventions (or assumptions) can be greatly relaxed, yet each
represents an essential issue that must be addressed.
We start with the first convention (or assumption). Observe that making no restriction
on the relationship among the lengths of the two inputs disallows the existence of
secure protocols for computing any “non-degenerate” functionality. The reason is that
the program of each party (in a protocol for computing the desired functionality) must
either depend only on the length of the party’s input or obtain information on the
counterpart’s input length. In case information of the latter type is not implied by the
output value, a secure protocol “cannot afford” to give it away.8 By using adequate

7 One may also consider the “non-verifiable” case (i.e., V ≡ 1), but in this case, nothing can prevent the second
party from acting as if its input is different from its “actual” secret input.
8 The situation is analogous to the definition of secure encryption, where it is required that the message length
be polynomially related to the key length. Actually, things become even worse in the current setting because of
the possible malicious behavior of parties.

616

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

padding, any “natural” functionality can be cast as one satisfying the equal-length
convention.9
We now turn to the second convention. Certainly, the total running time of a secure
two-party protocol for computing the functionality cannot be smaller than the time re-
quired to compute the functionality (in the ordinary sense). Arguing as in the case of
input lengths, one can see that we need an a priori bound on the complexity of the func-
tionality. A more general approach would be to let such a bound be given explicitly to
both parties as an auxiliary input. In such a case, the protocol can be required to run for
a time that is bounded by a fixed polynomial in this auxiliary parameter (i.e., the time-
complexity bound of f ). Assuming that a good upper bound of the complexity of f is
time-constructible, and using standard padding techniques, we can reduce this general
case to the special case discussed previously: That is, given a general functionality, g,
and a time-bound t : N → N, we introduce the functionality

def g(x, y) if i = j = t(|x|) = t(|y|)
f ((x, 1 ), (y, 1 )) =
i j
(⊥, ⊥) otherwise

where ⊥ is a special error symbol. Now, the problem of securely computing g reduces
to the problem of securely computing f , which in turn is polynomial-time computable.
Finally, we turn to the third convention. Indeed, a more general convention would
be to have an explicit security parameter that determines the security of the protocol.
This general alternative is essential for allowing “secure” computation of finite func-
tionalities (i.e., functionalities defined on finite input domains). We may accommodate
the general convention using the special case, postulated previously, as follows. Sup-
pose that we want to compute the functionality f , on input pair (x, y) with security
(polynomial in) the parameter s. Then we introduce the functionality

f
((x, 1s ), (y, 1s )) = f (x, y) ,

def

and consider secure protocols for computing f
. Indeed, this reduction corresponds to
the realistic setting where the parties first agree on the desired level of security, and
only then proceed to compute the function (using this level of security).

Partial functionalities. The first convention postulates that we are actually not con-
sidering mapping from the set of all pairs of bit strings, but rather mappings from a
certain (general) set of pairs of strings (i.e., ∪n∈N {0, 1}n × {0, 1}n ). Taking this conven-
tion one step further, one may consider functionalities that are defined only over a set
R ⊆ ∪n∈N {0, 1}n × {0, 1}n . Clearly, securely computing such a functionality f
can be
reduced to computing any of its extensions to ∪n∈N {0, 1}n × {0, 1}n (e.g., computing
def def
f such that f (x, y) = f
(x, y) for (x, y) ∈ R and f (x, y) = (⊥, ⊥) otherwise). With
one exception (to be discussed explicitly), our exposition only refers to functionalities
that are defined over the set of all pairs of strings of equal length.

9 In the sequel, we sometimes take the liberty of presenting functionalities in a form that violates the equal-length
convention (e.g., in the case of Oblivious Transfer). Indeed, these formulations can be easily modified to fit the
equal-length convention.

617
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

An alternative set of conventions. An alternative way of addressing all three concerns

discussed previously is to introduce an explicit security parameter, denoted n, and
consider the following sequence of functionalities  f n n∈N . Each f n is defined over the
set of all pairs of bit strings, but typically one considers only the value of f n on strings
of poly(n) length. In particular, for a functionality f as in our main conventions, one
def def
may consider f n (x, y) = f (x, y) if |x| = |y| = n and f n (x, y) = (⊥, ⊥) otherwise.
When following the alternative convention, one typically postulates that there exists a
poly(n)-time algorithm for computing f n (for a generic n), and security is also evaluated
with respect to the parameter n. We stress that in this case, the protocol’s running time
and its security guarantee are only related to the parameter n, and are independent of
the length of the input.10

7.2.1.2. Computational Indistinguishability: Conventions and Notation

As in Definition 7.1.1, we will often talk of the computational indistinguishability of
probability ensembles indexed by strings (as in Part 2 of Definition 3.2.7). Whenever
we do so, we refer to computational indistinguishability by (non-uniform) families
def
of polynomial-size circuits. That is, we say that the ensembles, X = {X w }w∈S and
def c
Y = {Yw }w∈S , are computationally indistinguishable, denoted X ≡ Y , if the following
holds:

For every polynomial-size circuit family, {Cn }n∈N , every positive polynomial
p(·), every sufficiently large n, and every w ∈ S ∩ {0, 1}n ,

1
|Pr [Cn (X w ) = 1] − Pr [Cn (Yw ) = 1] | < (7.5)
p(n)

Note that an infinite sequence of w’s may be incorporated in the family; hence, the
definition is not strengthened by providing the circuit Cn with w as an additional input.11
Recall that computational indistinguishability is a relaxation of statistical indistin-
def def
guishability, where here, the ensembles X = {X w }w∈S and Y = {Yw }w∈S are statisti-
s
cally indistinguishable, denoted X ≡ Y , if for every positive polynomial p(·), every
sufficiently large n and every w ∈ S ∩ {0, 1}n ,

1
|Pr [X w = α] − Pr [Yw = α]| < (7.6)
α∈{0,1}∗
p(n)

In case the differences are all equal to zero, we say that the ensembles are identically
distributed (and denote this by X ≡ Y ).

10 Consequently, the value of f n (x, y) may depend only on poly(n)-long prefixes of x and y.
11 Indeed, here we capitalize on the non-uniformity of the class of potential distinguishers. In case one considers the
class of uniform (probabilistic polynomial-time) distinguishers, it is necessary to provide these distinguishers
with the distribution’s index (i.e., w); see (Part 2 of) Definition 3.2.2.

618

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

7.2.1.3. Representation of Parties’ Strategies

In Chapter 4, the parties’ strategies for executing a given protocol (e.g., a proof system)
were represented by interactive Turing machines. In this chapter, we prefer an equiva-
lent formulation, which is less formal and less cumbersome. Specifically, the parties’
strategies are presented as functions mapping the party’s current view of the interactive
execution to the next message to be sent. Recall that the party’s view consists of its
initial input, an auxiliary initial input (which is relevant only for modeling adversarial
strategies), its random-tape, and the sequence of messages it has received so far. A
strategy will be called feasible if it is implementable in probabilistic polynomial-time
(i.e., the function associated with it is computable in polynomial-time).
As in Chapter 4, it is typically important to allow the adversaries to obtain (non-
uniformly generated) auxiliary inputs (cf. Section 4.3.3). Recall that auxiliary inputs
play a key role in guaranteeing that zero-knowledge is closed under sequential compo-
sition (see Section 4.3.4). Similarly, auxiliary inputs to the adversaries will play a key
role in composition theorems for secure protocols, which are pivotal to our exposition
and very important in general. Nevertheless, for the sake of simplicity, we often omit
the auxiliary inputs from our notations and discussions (especially in places where they
do not play an “active” role).
Recall that considering auxiliary inputs (as well as ordinary inputs) without intro-
ducing any restrictions (other than on their length) means that we are actually presenting
a treatment in terms of non-uniform complexity. Thus, all our assumptions will refer
to non-uniform complexity.

7.2.2. The Semi-Honest Model

Loosely speaking, a semi-honest party is one who follows the protocol properly with
the exception that it keeps a record of all its intermediate computations. Actually, it
suffices to keep the internal coin tosses and all messages received from the other party. In
particular, a semi-honest party tosses fair coins (as instructed by its program) and sends
messages according to its specified program (i.e., as a function of its input, outcome of
coin tosses, and incoming messages). Note that a semi-honest party corresponds to the
“honest verifier” in the definitions of zero-knowledge (cf. Section 4.3.1.7).
In addition to the methodological role of semi-honest parties in our exposition, they
do constitute a model of independent interest. In particular, deviation from the specified
program, which may be invoked inside a complex software application, is more difficult
than merely recording the contents of some communication registers. Furthermore,
records of these registers may be available through some standard activities of the
operating system. Thus, whereas general malicious behavior may be infeasible for many
users, semi-honest behavior may be feasible for them (and one cannot assume that they
just behave in a totally honest way). Consequently, in many settings, one may assume
that although the users may wish to cheat, they actually behave in a semi-honest way.
(We mention that the “augmented semi-honest” model, introduced in Section 7.4.4.1,
may be more appealing and adequate for more settings.)
619
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

In the following, we present two equivalent formulations of security in the semi-

honest model. The first formulation capitalizes on the simplicity of the current model
and defines security in it by a straightforward extension of the definition of zero-
knowledge. The second formulation applies the general methodology outlined in Sec-
tion 7.1.1. Indeed, both formulations follow the simulation paradigm, but the first does
so by extending the definition of zero-knowledge, whereas the second does so by de-
generating the general “real-vs.-ideal” methodology.

7.2.2.1. The Simple Formulation of Privacy

Loosely speaking, a protocol privately computes f if whatever can be obtained from a
party’s view of a (semi-honest) execution could be essentially obtained from the input
and output available to that party. This extends the formulation of (honest-verifier)
zero-knowledge by providing the simulator also with the (proper) output. The essence
of the definition is captured by the simpler special case of deterministic functionalities,
highlighted in the first item.

Definition 7.2.1 (privacy with respect to semi-honest behavior): Let f : {0, 1}∗ ×
{0, 1}∗ → {0, 1}∗ × {0, 1}∗ be a functionality, and f 1 (x, y) (resp., f 2 (x, y)) denote
the first (resp., second) element of f (x, y). Let
be a two-party protocol for
computing f .12 The view of the first (resp., second) party during an execution of

on (x, y), denoted view


1 (x, y) (resp., view2 (x, y)), is (x, r, m 1 , ..., m t ) (resp.,
(y, r, m 1 , ..., m t )), where r represents the outcome of the first (resp., second) party’s
internal coin tosses, and m i represents the i-th message it has received. The output
of the first (resp., second) party after an execution of
on (x, y), denoted output
1 (x, y)
(resp., output
2 (x, y)), is implicit in the party’s own view of the execution, and




output (x, y) = (output1 (x, y), output2 (x, y)).
r (deterministic case) For a deterministic functionality f , we say that
privately
computes f if there exist probabilistic polynomial-time algorithms, denoted S1 and
S2 , such that
c
{S1 (x, f 1 (x, y))}x, y∈{0,1}∗ ≡ {view

1 (x, y)} x, y∈{0,1}∗ (7.7)
c
{S2 (y, f 2 (x, y))}x, y∈{0,1}∗ ≡ {view

2 (x, y)} x, y∈{0,1}∗ (7.8)
c
where |x| = |y|. (Recall that ≡ denotes computational indistinguishability by (non-
uniform) families of polynomial-size circuits.)
r (general case) We say that
privately computes f if there exist probabilistic
polynomial-time algorithms, denoted S1 and S2 , such that
c
{(S1 (x, f 1 (x, y)), f (x, y))}x, y ≡ {(view


1 (x, y), output (x, y))} x, y (7.9)
c
{(S2 (y, f 2 (x, y)), f (x, y))}x, y ≡ {(view


2 (x, y), output (x, y))} x, y (7.10)

12 By saying that
computes (rather than privately computes) f , we mean that the output distribution of the
protocol (when played by honest or semi-honest parties) on input pair (x, y) is distributed identically to f (x, y).

620

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

We stress that view

1 (x, y), view2 (x, y), output1 (x, y), and output2 (x, y) are
related random variables, defined as a function of the same random execution. In
particular, outputi
(x, y) is fully determined by viewi
(x, y).
Consider first the deterministic case: Eq. (7.7) (resp., Eq. (7.8)) asserts that the view of
the first (resp., second) party, on each possible pair of inputs, can be efficiently simulated
based solely on its own input and output. Thus, all that this party learns from the full tran-
script of the proper execution is effectively implied by its own output from this execution
(and its own input to it). In other words, all that the party learns from the (semi-honest)
execution is essentially implied by the output itself. Next, note that the formulation for
the deterministic case coincides with the general formulation as applied to deterministic
functionalities (because, in any protocol
that computes a deterministic functionality
f , it must hold that output
(x, y) = f (x, y), for each pair of inputs (x, y)).13
In contrast to the deterministic case, augmenting the view of the semi-honest party by
the output of the other party is essential when randomized functionalities are concerned.
Note that in this case, for any protocol
that computes a randomized functionality
f , it does not necessarily hold that output
(x, y) = f (x, y), because each of the two
objects is a random variable. Indeed, these two random variables must be identically (or
similarly) distributed, but this does not suffice for asserting, for example, that Eq. (7.7)
implies Eq. (7.9). Two disturbing counterexamples follow:

1. Consider the functionality (1n , 1n ) → (r, λ), where r is uniformly distributed in

{0, 1}n , and a protocol in which Party 1 uniformly selects r ∈ {0, 1}n , sends it to
Party 2, and outputs r . Clearly, this protocol computes the said functionality; alas,
intuitively we should not consider this computation private (because Party 2 learns
the output of Party 1 although Party 2 is not supposed to learn anything about that
output). However, a simulator S2 (1n ) that outputs a uniformly chosen r ∈ {0, 1}n
satisfies Eq. (7.8) (but does not satisfy Eq. (7.10)).
The point is that Eq. (7.9) and Eq. (7.10) refer to the relation between a party’s
output and the other party’s view in the same execution of the protocol, and require
that this relation be maintained in the simulation. As is vividly demonstrated in
the aforementioned example, this relation is at the heart of the notion of security:
We should simulate a view (of the semi-honest party) that fits the actual output of the
honest party, and not merely simulate a view that fits the distribution of the possible
output of the honest party.
2. Furthermore, Eq. (7.9) and Eq. (7.10) require that the party’s simulated view fit its
actual output (which is given to the simulator). To demonstrate the issue at hand,
consider the foregoing functionality, and a protocol in which Party 1 uniformly selects
s ∈ {0, 1}n , and outputs r ← F(s), where F is a one-way permutation. Again, this
protocol computes the previous functionality, but it is not clear whether or not we
may consider this computation private (because Party 1 learns the pre-image of its
own output under F, something it could not have obtained if a trusted party were to

13 Recall that the input pairs (x, y) serve as indices to the distributions in the two ensembles under consideration,
and as such they are always given (or incorporated) in the potential distinguisher; see Section 7.2.1.2.

621
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

give it the output). Note that a simulator S1 (1n , r ) that uniformly selects s ∈ {0, 1}n
and outputs (s, F(s)) satisfies Eq. (7.7) (but does not satisfy Eq. (7.9)).
We comment that the current issue is less acute than the first one (i.e., the one raised
in Item 1). Indeed, consider the following alternative to both Eq. (7.7) and Eq. (7.9):
c
{(S1 (x, f 1 (x, y)), f 2 (x, y))}x, y ≡ {(view


1 (x, y), output2 (x, y))}x, y (7.11)

Note that Eq. (7.11) addresses the problem raised in Item 1, but not the problem raised
in the current item. But is the current problem a real one? Note that the only difference
between Eq. (7.9) and Eq. (7.11) is that the former forces the simulated view to fit the
output given to the simulator, whereas this is not guaranteed in Eq. (7.11). Indeed,
in Eq. (7.11) the view simulated for Party 1 may not fit the output given to the
simulator, but the simulated view does fit the output given to the honest Party 2. Is
the former fact of real importance or is it the case that all that matters is the relation
of the simulated view to the honest party’s view? We are not sure, but (following a
general principle) when in doubt, we prefer to be more careful and adopt the more
stringent definition. Furthermore, the stronger definition simplifies the proof of the
Composition Theorem for the semi-honest model (i.e., Theorem 7.3.3).

What about Auxiliary Inputs? Auxiliary inputs are implicit in Definition 7.2.1. They
are represented by the fact that the definition asks for computational indistinguisha-
bility by non-uniform families of polynomial-size circuits (rather than computational
indistinguishability by probabilistic polynomial-time algorithms). In other words, in-
distinguishability also holds with respect to probabilistic polynomial-time machines
that obtain (non-uniform) auxiliary inputs.

Private Computation of Partial Functionalities. For functionalities that are defined

only for inputs pairs in some set R ⊂ {0, 1}∗ × {0, 1}∗ (see Section 7.2.1.1), private
computation is defined as in Definition 7.2.1, except that the ensembles are indexed by
pairs in R.

7.2.2.2. The Alternative Formulation

It is instructive to recast the previous definition in terms of the general (“real-vs.-ideal”)
framework discussed in Section 7.1.1 (and used extensively in the case of arbitrary
malicious behavior). In this framework, one first considers an ideal model in which the
(two) parties are joined by a (third) trusted party, and the computation is performed via
this trusted party. Next, one considers the real model in which a real (two-party) protocol
is executed (and there exist no trusted third parties). A protocol in the real model is said
to be secure with respect to certain adversarial behavior if the possible real executions
with such an adversary can be “simulated” in the corresponding ideal model. The notion
of simulation used here is different from the one used in Section 7.2.2.1: The simulation
is not of the view of one party via a traditional algorithm, but rather a simulation of the
joint view of both parties by the execution of an ideal-model protocol.
According to the general methodology (framework), we should first specify the
ideal-model protocol. In the case of semi-honest adversaries, the ideal model consists
622

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

of each party sending its input to the trusted party (via a secure private channel), and
the third party computing the corresponding output pair and sending each output to the
corresponding party. The only adversarial behavior allowed here is for one of the parties
to determine its own output based on its input and the output it has received (from the
trusted party).14 This adversarial behavior represents the attempt to learn something
from the party’s view of a proper execution (which, in the ideal model, consists only of
its local input and output). The other (i.e., honest) party merely outputs the output that
it has received (from the trusted party).
Next, we turn to the real model. Here, there is a real two-party protocol and the
adversarial behavior is restricted to be semi-honest. That is, the protocol is executed
properly, but one party may produce its output based on (an arbitrary polynomial-
time computation applied to) its view of the execution (as defined earlier). We
stress that the only adversarial behavior allowed here is for one of the parties to
determine its own output based on its entire view of the proper execution of the
protocol.
Finally, we define security in the semi-honest model. A secure protocol for the real
(semi-honest) model is such that for every feasible semi-honest behavior of one of the
parties, we can simulate the joint outcome (of their real computation) by an execution in
the ideal model (where also one party is semi-honest and the other is honest). Actually,
we need to augment the definition to account for a priori information available to semi-
honest parties before the protocol starts. This is done by supplying these parties with
auxiliary inputs.
Note that in both (ideal and real) models, the (semi-honest) adversarial behavior
takes place only after the proper execution of the corresponding protocol. Thus, in the
ideal model, this behavior is captured by a computation applied to the local input–output
pair, whereas in the real model, this behavior is captured by a computation applied to
the party’s local view (of the execution).

Definition 7.2.2 (security in the semi-honest model): Let f : {0, 1}∗ × {0, 1}∗ →
{0, 1}∗ × {0, 1}∗ be a functionality, where f 1 (x, y) (resp., f 2 (x, y)) denotes the first
(resp., second) element of f (x, y), and let
be a two-party protocol for computing f .
r Let B = (B1 , B2 ) be a pair of probabilistic polynomial-time algorithms representing
parties’ strategies for the ideal model. Such a pair is admissible (in the ideal model)
if for at least one Bi we have Bi (u, v, z) = v, where u denotes the party’s local input,
v its local output, and z its auxiliary input. The joint execution of f under B in the
ideal model on input pair (x, y) and auxiliary input z, denoted ideal f, B(z) (x, y), is
defined as ( f (x, y), B1 (x, f 1 (x, y), z), B2 (y, f 2 (x, y), z)).
(That is, if Bi is honest, then it just outputs the value fi (x, y) obtained from the
trusted party, which is implicit in this definition. Thus, our peculiar choice to feed
both parties with the same auxiliary input is immaterial, because the honest party
ignores its auxiliary input.)

14 We stress that unlike in the malicious model, discussed in Section 7.2.3, here the dishonest (or rather semi-honest)
party is not allowed to modify its input (but must hand its actual input to the trusted party).

623
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

r Let A = (A1 , A2 ) be a pair of probabilistic polynomial-time algorithms representing

parties’ strategies for the real model. Such a pair is admissible (in the real model)
if for at least one i ∈ {1, 2} we have Ai (view, aux) = out for every view and aux,
where out is the output implicit in view. The joint execution of
under A in the real
model on input pair (x, y) and auxiliary input z, denoted real
, A(z) (x, y), is defined
as (output
(x, y), A1 (view



1 (x, y), z), A2 (view2 (x, y), z)), where output (x, y)


and the viewi (x, y)’s refer to the same execution and are defined as in Defini-
tion 7.2.1.
(Again, if Ai is honest, then it just outputs the value f i (x, y) obtained from the
execution of
, and we may feed both parties with the same auxiliary input.)
Protocol
is said to securely compute f in the semi-honest model (secure with
respect to f and semi-honest behavior) if for every probabilistic polynomial-time
pair of algorithms A = (A1 , A2 ) that is admissible for the real model, there exists a
probabilistic polynomial-time pair of algorithms B = (B1 , B2 ) that is admissible for
the ideal model such that
c
{ideal f, B(z) (x, y)}x, y,z ≡ {real
, A(z) (x, y)}x, y,z (7.12)
where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).

Observe that the definition of the joint execution in the real model prohibits both
parties (honest and semi-honest) from deviating from the strategies specified by
.
The difference between honest and semi-honest parties is merely in their actions on the
corresponding local views of the execution: An honest party outputs only the output part
of the view (as specified by
), whereas a semi-honest party may output an arbitrary
(feasibly computable) function of the view. Note that including the output f (x, y) (resp.,
output
(x, y)) in ideal f, B(z) (x, y) (resp., in real
, A(z) (x, y)) is meaningful only in
the case of a randomized functionality f , and is done in order to match the formulation
in Definition 7.2.1. We stress that the issue is the inclusion of the output of the dishonest
party (see Item 2 in the discussion that follows Definition 7.2.1).
We comment that, as will become clear in the proof of Proposition 7.2.3, omitting
the auxiliary input does not weaken Definition 7.2.2. Intuitively, since the adversary is
passive, the only affect of the auxiliary input is that it appears as part of the adversary’s
view. However, since Eq. (7.12) refers to the non-uniform formulation of computational
indistinguishability, augmenting the ensembles by auxiliary inputs has no affect.

7.2.2.3. Equivalence of the Two Formulations

It is not hard to see that Definitions 7.2.1 and 7.2.2 are equivalent. That is,

Proposition 7.2.3: Let
be a protocol for computing f . Then,
privately computes

f if and only if
securely computes f in the semi-honest model.

Proof Sketch: We first show that Definition 7.2.2 implies Definition 7.2.1. Suppose
that
securely computes f in the semi-honest model (i.e., satisfies Definition 7.2.2).
624

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

Without loss of generality, we show how to simulate the first party’s view. Toward this
end, we define the following admissible pair A = (A1 , A2 ) for the real model: A1 is
merely the identity transformation (i.e., it outputs the view given to it), whereas A2
(which represents an honest strategy for Party 2) produces an output as determined
by the view given to it. We stress that we consider an adversary A1 that does not
get an auxiliary input (or alternatively ignores it). Furthermore, the adversary merely
outputs the view given to it (and leaves the possible processing of this view to the
potential distinguisher). Let B = (B1 , B2 ) be the ideal-model adversary guaranteed
by Definition 7.2.2. We claim that (using) B1 (in the role of S1 ) satisfies Eq. (7.9),
rather than only Eq. (7.7). Loosely speaking, the claim holds because Definition 7.2.2
guarantees that the relation between the view of Party 1 and the outputs of both parties
in a real execution is preserved in the ideal model. Specifically, since A1 is a passive
adversary (and
computes f ), the output of Party 1 in a real execution equals the
value that is determined in the view (of Party 1), which in turn fits the functionality.
Now, Definition 7.2.2 implies that the same relation between the (simulated) view of
Party 1 and the outputs must hold in the ideal model. It follows that using B1 in role of
S1 guarantees that the simulated view fits the output given to the simulator (as well as
the output not given to it).
We now show that Definition 7.2.1 implies Definition 7.2.2. Suppose that
privately
computes f , and let S1 and S2 be as guaranteed in Definition 7.2.1. Let A = (A1 , A2 ) be
an admissible pair for the real-model adversaries. Without loss of generality, we assume
that A2 merely maps the view (of the second party) to the corresponding output (i.e.,
f 2 (x, y)); that is, Party 2 is honest (and Party 1 is semi-honest). Then, we define an ideal-
def def
model pair B = (B1 , B2 ) such that B1 (x, v, z) = A1 (S1 (x, v), z) and B2 (y, v, z) = v.
(Note that B is indeed admissible with respect to the ideal model.) The following holds
(for any infinite sequence of (x, y, z)’s):

real
, A(z) (x, y) = (output
(x, y), A1 (view

1 (x, y), z), A 2 (view2 (x, y), z))

= (output
(x, y), A1 (view

1 (x, y), z), output2 (x, y))
c
≡ ( f (x, y), A1 (S1 (x, f 1 (x, y)), z), f 2 (x, y))
= ( f (x, y), B1 (x, f 1 (x, y), z), B2 (y, f 2 (x, y), z))
= ideal f, B(z) (x, y)
c
where the computational indistinguishability (i.e., ≡) is due to the guarantee re-
garding S1 (in its general form); that is, Eq. (7.9). Indeed, the latter only guaran-
c
tees (view


1 (x, y), output (x, y)) ≡ (S1 (x, f 1 (x, y)), f (x, y)), but by incorporating
A1 and z in the potential distinguisher, the soft-equality follows.

Conclusion. This proof demonstrates that the alternative formulation of Definition 7.2.2
is merely a cumbersome form of the simpler Definition 7.2.1. We stress that the rea-
son we have presented the cumbersome form is the fact that it follows the general
framework of definitions of security that is used for the malicious adversarial behav-
ior. In the rest of this chapter, whenever we deal with the semi-honest model (for
625
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

two-party computation), we will use Definition 7.2.1. Furthermore, since much of the
text focuses on deterministic functionalities, we will be able to use the simpler case of
Definition 7.2.1.

7.2.3. The Malicious Model

We now turn to consider arbitrary feasible deviation from the specified program of a
two-party protocol. A few preliminary comments are in place. Firstly, there is no way to
force parties to participate in the protocol. That is, a possible malicious behavior may
consist of not starting the execution at all, or, more generally, suspending (or aborting)
the execution at any desired point in time. In particular, a party can abort at the first
moment when it obtains the desired result of the computed functionality. We stress that
our model of communication does not allow conditioning of the receipt of a message
by one party on the concurrent sending of a proper message by this party. Thus, no two-
party protocol can prevent one of the parties from aborting when obtaining the desired
result and before its counterpart also obtains the desired result. In other words, it can
be shown that perfect fairness – in the sense of both parties obtaining the outcome of
the computation concurrently – is not achievable in a two-party computation. We thus
give up on such fairness altogether. (We comment that partial fairness is achievable;
see Section 7.7.1.1).
Secondly, observe that when considering malicious adversaries, it is not clear what
their input to the protocol is. That is, a malicious party can enter the protocol with
arbitrary input, which may not equal its “true” local input. There is no way for a
protocol to tell the “true” local input from the one claimed by a party (or, in other
words, to prevent a malicious party from modifying its input). (We stress that these
phenomena did not occur in the semi-honest model, for the obvious reason that parties
were postulated not to deviate from the protocol.)
In view of this discussion, there are three things we cannot hope to avoid (no matter
what protocol we use):
1. Parties refusing to participate in the protocol (when the protocol is first invoked).
2. Parties substituting their local input (and entering the protocol with an input other
than the one provided to them).
3. Parties aborting the protocol prematurely (e.g., before sending their last message).
Thus, we shall consider a two-party protocol to be secure if the adversary’s behavior in it
is essentially restricted to these three actions. Following the real-vs.-ideal methodology
(of Section 7.1.1), this means that we should define an ideal model that corresponds to
these possible actions, and define security such that the execution of a secure protocol
in the real model can be simulated by the ideal model.

7.2.3.1. The Actual Definition

We start with a straightforward implementation of the previous discussion. An alterna-
tive approach, which is simpler but partial, is presented in Section 7.2.3.2. (Specifically,
the alternative approach is directly applicable only to single-output functionalities, in
626

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

which case the complications introduced by aborting do not arise. The interested reader
may proceed directly to Section 7.2.3.2, which is mostly self-contained.)

The Ideal Model. We first translate the previous discussion into a definition of an ideal
model. That is, we will allow in the ideal model whatever cannot possibly be prevented
in any real execution. An alternative way of looking at things is that we assume that the
the two parties have at their disposal a trusted third party, but even such a party cannot
prevent certain malicious behavior. Specifically, we allow a malicious party in the ideal
model to refuse to participate in the protocol or to substitute its local input. (Clearly,
neither can be prevented by a trusted third party.) In addition, we postulate that the
first party has the option of “stopping” the trusted party just after obtaining its part of
the output, and before the trusted party sends the other output part to the second party.
Such an option is not given to the second party.15 Thus, an execution in the ideal model
proceeds as follows (where all actions of both the honest and the malicious parties must
be feasible to implement):

Inputs: Each party obtains an input, denoted u.

Sending inputs to the trusted party: An honest party always sends u to the trusted party.
A malicious party may, depending on u (as well as on an auxiliary input and its coin
tosses), either abort or send some u
∈ {0, 1}|u| to the trusted party.16
The trusted party answers the first party: In case it has obtained an input pair, (x, y),
the trusted party (for computing f ) first replies to the first party with f 1 (x, y).
Otherwise (i.e., in case it receives only one input), the trusted party replies to both
parties with a special symbol, denoted ⊥.
The trusted party answers the second party: In case the first party is malicious, it may,
depending on its input and the trusted party’s answer, decide to stop the trusted
party. In this case, the trusted party sends ⊥ to the second party. Otherwise (i.e., if
not stopped), the trusted party sends f 2 (x, y) to the second party.
Outputs: An honest party always outputs the message it has obtained from the trusted
party. A malicious party may output an arbitrary (polynomial-time computable)
function of its initial input (auxiliary input and random-tape) and the message it has
obtained from the trusted party.
In fact, without loss of generality, we may assume that both parties send inputs to the
trusted party (rather than allowing the malicious party not to enter the protocol). This
assumption can be justified by letting the trusted party use some default value (or a

15 This asymmetry is due to the non-concurrent nature of communication in the model. Since we postulate that
the trusted party sends the answer first to the first party, the first party (but not the second) has the option of
stopping the trust party after obtaining its part of the output. The second party can only stop the trust party
before obtaining its output, but this is the same as refusing to participate. See further discussion at the end of
the current subsection.
16 We comment that restricting the ideal-model adversary (to replacing u by u
of the same length) only strengthens

the definition of security. This restriction is essential to our formulation, because (by our convention) the
functionality f is defined only for pairs of strings of equal length.

627
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

special abort symbol) in case it does not get an input from one of the parties.17 Thus, the
ideal model (computation) is captured by the following definition, where the algorithms
B1 and B2 represent all possible actions in the model.18 In particular, B1 (x, z, r ) (resp.,
B2 (y, z, r )) represents the input handed to the trusted party by Party 1 (resp., Party 2)
having local input x (resp., y) and auxiliary input z and using random-tape r . Indeed, if
Party 1 (resp., Party 2) is honest, then B1 (x, z, r ) = x (resp., B2 (y, z, r ) = y). Likewise,
B1 (x, z, r, v) = ⊥ represents a decision of Party 1 to stop the trusted party, on input
x (auxiliary input z and random-tape r ), after receiving the (output) value v from the
trusted party. In this case, B1 (x, z, r, v, ⊥) represents the party’s local output. Otherwise
(i.e., B1 (x, z, r, v) = ⊥), we let B1 (x, z, r, v) itself represent the party’s local output.
The local output of Party 2 is always represented by B2 (y, z, r, v), where y is the party’s
local input (z is the auxiliary input, r is the random-tape) and v is the value received from
the trusted party. Indeed, if Party 1 (resp., Party 2) is honest, then B1 (x, z, r, v) = v
(resp., B2 (y, z, r, v) = v).

Definition 7.2.4 (malicious adversaries, the ideal model): Let f : {0, 1}∗ × {0, 1}∗ →
{0, 1}∗ × {0, 1}∗ be a functionality, where f 1 (x, y) (resp., f 2 (x, y)) denotes the first
(resp., second) element of f (x, y). Let B = (B1 , B2 ) be a pair of probabilistic
polynomial-time algorithms representing strategies in the ideal model. Such a pair
is admissible (in the ideal malicious model) if for at least one i ∈ {1, 2}, called hon-
est, we have Bi (u, z, r ) = u and Bi (u, z, r, v) = v, for every possible value of u, z, r ,
and v. Furthermore, |Bi (u, z, r )| = |u| must hold for both i’s. The joint execution of
f under B in the ideal model (on input pair (x, y) and auxiliary input z), denoted
ideal f, B(z) (x, y), is defined by uniformly selecting a random-tape r for the adversary,
def
and letting ideal f, B(z) (x, y) = ϒ(x, y, z, r ), where ϒ(x, y, z, r ) is defined as follows:
r In case Party 1 is honest, ϒ(x, y, z, r ) equals

def
( f 1 (x, y
) , B2 (y, z, r, f 2 (x, y
))), where y
= B2 (y, z, r ). (7.13)
r In case Party 2 is honest, ϒ(x, y, z, r ) equals

(B1 (x, z, r, f 1 (x
, y), ⊥) , ⊥) if B1 (x, z, r, f 1 (x
, y)) = ⊥ (7.14)

(B1 (x, z, r, f 1 (x
, y)) , f 2 (x
, y)) otherwise (7.15)
def
where, in both cases, x
= B1 (x, z, r ).

17 Both options (i.e., default value or a special abort symbol) are useful, and the choice depends on the protocol
designer. In case a special abort symbol is used, the functionality should be modified accordingly, such that if
one of the inputs equals the special abort symbol, then the output is a special abort symbol.
18 As in Definition 7.2.2, we make the peculiar choice of feeding both B ’s with the same auxiliary input z (and the
i
same random-tape r ). However, again, the honest strategy ignores this auxiliary input, which is only used by the
malicious strategy. Note that unlike in previous definitions, we make the random-tape (of the adversary) explicit
in the notation, the reason being that the same strategy is used to describe two different actions of the adversary
(rather than a single action, as in Definition 7.2.2). Since these actions may be probabilistically related, it is
important that they be determined based on the same random-tape.

628

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

Eq. (7.14) and Eq. (7.15) refer to the case in which Party 2 is honest (and Party 1 may
be malicious). Specifically, Eq. (7.14) represents the sub-case where Party 1 invokes
the trusted party with a possibly substituted input, denoted B1 (x, z, r ), and aborts while
stopping the trusted party right after obtaining the output, f 1 (B1 (x, z, r ), y). In this
sub-case, Party 2 obtains no output (from the trusted party). Eq. (7.15) represents the
sub-case where Party 1 invokes the trusted party with a possibly substituted input, and
allows the trusted party to answer Party 2. In this sub-case, Party 2 obtains and outputs
f 2 (B1 (x, z, r ), y). In both sub-cases, the trusted party computes f (B1 (x, z, r ), y), and
Party 1 outputs a string that depends on both x, z, r and f 1 (B1 (x, z, r ), y). Likewise,
Eq. (7.13) represents possible malicious behavior of Party 2; however, in accordance
with the previous discussion, the trusted party first supplies output to Party 1, and so
Party 2 does not have a “real” aborting option (analogous to Eq. (7.14)).

Execution in the Real Model. We next consider the real model in which a real (two-
party) protocol is executed (and there exist no trusted third parties). In this case, a
malicious party may follow an arbitrary feasible strategy, that is, any strategy imple-
mentable by a probabilistic polynomial-time algorithm (which gets an auxiliary input).
In particular, the malicious party may abort the execution at any point in time, and when
this happens prematurely, the other party is left with no output. In analogy to the ideal
case, we use algorithms to define strategies in a protocol, where these strategies (or
algorithms implementing them) map partial execution histories to the next message.

Definition 7.2.5 (malicious adversaries, the real model): Let f be as in Definition 7.2.4,
and
be a two-party protocol for computing f . Let A = (A1 , A2 ) be a pair of prob-
abilistic polynomial-time algorithms representing strategies in the real model. Such a
pair is admissible (with respect to
) (for the real malicious model) if at least one Ai
coincides with the strategy specified by
. (In particular, this Ai ignores the auxiliary
input.) The joint execution of
under A in the real model (on input pair (x, y) and
auxiliary input z), denoted real
, A(z) (x, y), is defined as the output pair resulting from
the interaction between A1 (x, z) and A2 (y, z). (Recall that the honest Ai ignores the
auxiliary input z, and so our peculiar choice of providing both Ai ’s with the same z is
immaterial.)

In some places (in Section 7.4), we will assume that the algorithms representing the
real-model adversaries (i.e., the algorithm Ai that does not follow
) are deterministic.
This is justified by observing that one may just (consider and) fix the “best” possible
choice of coins for a randomized adversary and incorporate this choice in the auxiliary
input of a deterministic adversary (cf. Section 1.3.3 of Volume 1).

Security as Emulation of Real Execution in the Ideal Model. Having defined the
ideal and real models, we obtain the corresponding definition of security. Loosely
speaking, the definition asserts that a secure two-party protocol (in the real model)
emulates the ideal model (in which a trusted party exists). This is formulated by saying
that admissible adversaries in the ideal model are able to simulate (in the ideal model)
the execution of a secure real-model protocol under any admissible adversaries.
629
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Definition 7.2.6 (security in the malicious model): Let f and
be as in Defini-

tion 7.2.5. Protocol
is said to securely compute f (in the malicious model) if for
every probabilistic polynomial-time pair of algorithms A = (A1 , A2 ) that is admissible
for the real model (of Definition 7.2.5), there exists a probabilistic polynomial-time pair
of algorithms B = (B1 , B2 ) that is admissible for the ideal model (of Definition 7.2.4)
such that
c
{ideal f, B(z) (x, y)}x, y,z ≡ {real
, A(z) (x, y)}x, y,z
c
where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|). (Recall that ≡ de-
notes computational indistinguishability by (non-uniform) families of polynomial-size
circuits.) When the context is clear, we sometimes refer to
as a secure implementa-
tion of f .

One important property that Definition 7.2.6 implies is privacy with respect to ma-
licious adversaries. That is, all that an adversary can learn by participating in the
protocol, while using an arbitrary (feasible) strategy, can be essentially inferred from
the corresponding output alone. Another property that is implied by Definition 7.2.6 is
correctness, which means that the output of the honest party must be consistent with
an input pair in which the element corresponding to the honest party equals the party’s
actual input. Furthermore, the element corresponding to the adversary must be chosen
obliviously of the honest party’s input. We stress that both properties are easily implied
by Definition 7.2.6, but the latter is not implied by combining the two properties. For
further discussion, see Exercise 3.
We wish to highlight another property that is implied by Definition 7.2.6: Loosely
speaking, this definition implies that at the end of the (real) execution of a secure pro-
tocol, each party “knows” the value of the corresponding input for which the output
is obtained.19 That is, when a malicious Party 1 obtains the output v, it knows an x

(which does not necessarily equal to its initial local input x) such that v = f 1 (x
, y) for
some y (i.e., the local input of the honest Party 2). This “knowledge” is implied by the
equivalence to the ideal model, in which the party explicitly hands the (possibly modi-
fied) input to the trusted party. For example, say Party 1 uses the malicious strategy A1 .
Then the output values (in real
, A (x, y)) correspond to the input pair (B1 (x), y),
where B1 is the ideal-model adversary derived from the real-model adversarial
strategy A1 .
We comment that although Definition 7.2.6 does not talk about transforming ad-
missible A’s to admissible B’s, we will often use such phrases. Furthermore, although
the definition does not even guarantee that such a transformation is effective (i.e.,
computable), the transformations used in this work are all polynomial-time com-
putable. Moreover, these transformations consist of generic programs for Bi that use

19 One concrete case where this property plays a central role is in the input-commitment functionality (of Sec-
tion 7.4.3.6). Specifically, if a secure implementation of this functionality is first used in order to let Party 1
commit to its input, and next, Party 2 uses it in order to commit to its own input, then this property implies that
Party 2 cannot just copy the “commitment” made by Party 1 (unless Party 2 knows the input of Party 1).

630

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

subroutine (or oracle) calls to the corresponding Ai . Consequently, we sometimes

describe these transformations without referring to the auxiliary input, and the de-
scription can be completed by having Bi pass its auxiliary input to Ai (in each of its
invocations).

Further Discussion. As explained earlier, it is unavoidable that one party can abort the
real execution after it (fully) learns its output but before the other party (fully) learns
its own output. However, the convention by which this ability is designated to Party 1
(rather than to Party 2) is quite arbitrary. More general conventions (and corresponding
definitions of security) may be more appealing, but the current one seems simplest
and suffices for the rest of our exposition.20 An unrelated is issue is that unlike in
the treatment of the semi-honest model (cf. Definitions 7.2.1 and 7.2.2), we did not
explicitly include the output f (x, y) (resp., output
(x, y)) in ideal f, B(z) (x, y) (resp.,
in real
, A(z) (x, y)). Note that such an augmentation would not make much sense in the
current (malicious) context. Furthermore, recall that this issue is meaningful only in the
case of a randomized functionality f , and that its concrete motivation was to simplify
the proof of the composition theorem for the semi-honest model (which is irrelevant
here). Finally, referring to a third unrelated issue, we comment that the definitional
treatment can be extended to partial functionalities.

Remark 7.2.7 (security for partial functionalities): For functionalities that are defined
only for input pairs in some set R ⊂ {0, 1}∗ × {0, 1}∗ (see Section 7.2.1.1), security is
defined as in Definition 7.2.6 with the following two exceptions:

1. When defining the ideal model, the adversary is allowed to modify its input arbitrarily
as long as the modified input pair is in R.
2. The ensembles considered are indexed by triplets (x, y, z) that satisfy (x, y) ∈ R as
well as |x| = |y| and |z| = poly(|x|).

7.2.3.2. An Alternative Approach

A simpler definition of security may be used in the special case of single-output func-
tionalities (i.e., functionalities in which only one party obtains an output). Assume,
without loss of generality, that only the first party obtains an output (from the func-
tionality f ); that is, f (x, y) = ( f 1 (x, y), λ).21 In this case, we need not be concerned

20 One alternative convention is to associate with each protocol a binary value indicating which of the two parties is
allowed to meaningfully abort. This convention yields a more general (or less restrictive) definition of security,
where Definition 7.2.6 is obtained as a special case (in which this value is always required to equal 1). Yet the
protocols presented in this work are shown to be secure under the more restrictive definition.
21 Actually, the treatment of the case in which only the second party obtains an output (i.e., f (x, y) = (λ, f (x, y)))
2
is slightly different. However, also in this case, the event in which the first party aborts after obtaining its (empty)
output can be discarded. In this case, this event (of obtaining an a priori fixed output) is essentially equivalent
to the party aborting before obtaining output, which in turn can be viewed as replacing its input by a special
symbol.

631
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

with what happens after the first party obtains its output (because the second party has
no output), and thus the complications arising from the issue of aborting the execution
can be eliminated. Consequently, computation in the ideal model takes the following
form:

Inputs: Each party obtains an input, denoted u.

Sending inputs to the trusted party: An honest party always sends u to the trusted party.
A malicious party may, depending on u (as well as on an auxiliary input and its coin
tosses), either abort or send some u
∈ {0, 1}|u| to the trusted party. However, without
loss of generality, aborting at this stage may be treated as supplying the trusted party
with a special symbol.

The answer of the trusted party: Upon obtaining an input pair, (x, y), the trusted party
(for computing f ) replies to the first party with f 1 (x, y). Without loss of generality,
the trusted party only answers the first party, because the second party has no output
(or, alternatively, should always output λ).

Outputs: An honest party always outputs the message it has obtained from the trusted
party. A malicious party may output an arbitrary (polynomial-time computable)
function of its initial input (auxiliary input and its coin tosses) and the message it
has obtained from the trusted party.

Thus, the ideal model (computation) is captured by the following definition, where
the algorithms B1 and B2 represent all possible actions in the model. In particular,
B1 (x, z, r ) (resp., B2 (y, z, r )) represents the input handed to the trusted party by Party 1
(resp., Party 2) having local input x (resp., y), auxiliary input z, and random-tape r .
Indeed, if Party 1 (resp., Party 2) is honest, then B1 (x, z, r ) = x (resp., B2 (y, z, r ) = y).
Likewise, B1 (x, z, r, v) represents the output of Party 1, when having local input x
(auxiliary input z and random-tape r ) and receiving the value v from the trusted party,
whereas the output of Party 2 is represented by B2 (y, z, r, λ). Indeed, if Party 1 (resp.,
Party 2) is honest, then B1 (x, z, r, v) = v (resp., B2 (y, z, r, λ) = λ).

Definition 7.2.8 (the ideal model): Let f : {0, 1}∗ × {0, 1}∗ → {0, 1}∗ × {λ} be a
single-output functionality such that f (x, y) = ( f 1 (x, y), λ). Let B = (B1 , B2 ) be a
pair of probabilistic polynomial-time algorithms representing strategies in the ideal
model. Such a pair is admissible (in the ideal malicious model) if for at least one
i ∈ {1, 2}, called honest, we have Bi (u, z, r ) = u and Bi (u, z, r, v) = v for all pos-
sible u, z, r , and v. Furthermore, |Bi (u, z, r )| = |u| must hold for both i’s. The joint
execution of f under B in the ideal model (on input pair (x, y) and auxiliary input
z), denoted ideal f, B(z) (x, y), is defined by uniformly selecting a random-tape r for the
def
adversary, and letting ideal f, B(z) (x, y) = ϒ(x, y, z, r ), where

def
ϒ(x, y, z, r ) = (B1 (x, z, r, f 1 (B1 (x, z, r ), B2 (y, z, r ))) , B2 (y, z, r, λ)) (7.16)

632

www.Ebook777.com
 7.2* THE TWO-PARTY CASE: DEFINITIONS

def
That is, ideal f, B(z) (x, y) = (B1 (x, z, r, v), B2 (y, z, r, λ)), where v ← f 1 (B1 (x, z, r ),
B2 (y, z, r )) and r is uniformly distributed among the set of strings of adequate length.22
We next consider the real model in which a real (two-party) protocol is executed (and
there exist no trusted third parties). In this case, a malicious party may follow an arbitrary
feasible strategy, that is, any strategy implementable by a probabilistic polynomial-time
algorithm. The definition is identical to Definition 7.2.5, and is reproduced here (for
the reader’s convenience).
Definition 7.2.9 (the real model): Let f be as in Definition 7.2.8, and
be a two-party
protocol for computing f . Let A = (A1 , A2 ) be a pair of probabilistic polynomial-time
algorithms representing strategies in the real model. Such a pair is admissible (with
respect to
) (for the real malicious model) if at least one Ai coincides with the strategy
specified by
. The joint execution of
under A in the real model (on input pair
(x, y) and auxiliary input z), denoted real
, A(z) (x, y), is defined as the output pair
resulting from the interaction between A1 (x, z) and A2 (y, z). (Note that the honest Ai
ignores the auxiliary input z.)
Having defined the ideal and real models, we obtain the corresponding definition of
security. Loosely speaking, the definition asserts that a secure two-party protocol (in
the real model) emulates the ideal model (in which a trusted party exists). This is
formulated by saying that admissible adversaries in the ideal model are able to simulate
(in the ideal model) the execution of a secure real-model protocol under any admissible
adversaries. The definition is analogous to Definition 7.2.6.
Definition 7.2.10 (security): Let f and
be as in Definition 7.2.9. Protocol
is said
to securely compute f (in the malicious model) if for every probabilistic polynomial-
time pair of algorithms A = (A1 , A2 ) that is admissible for the real model (of Defini-
tion 7.2.9), there exists a probabilistic polynomial-time pair of algorithms B = (B1 , B2 )
that is admissible for the ideal model (of Definition 7.2.8) such that
c
{ideal f, B(z) (x, y)}x, y,z ≡ {real
, A(z) (x, y)}x, y,z

where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).

Clearly, as far as single-output functionalities are concerned, Definitions 7.2.6

and 7.2.10 are equivalent (because in this case, the ideal-model definitions coincide).
It is also clear from the previous discussions that the two definitions are not equiv-
alent in general (i.e., with respect to two-output functionalities). Still, it is possible
to securely implement any (two-output) functionality by using a protocol for securely
computing a (related) single-output functionality. That is, the ability to construct secure
protocols under Definition 7.2.10 yields the ability to construct secure protocols under
Definition 7.2.6.

22 Recall that if Bi is honest, then it passes its input to the trusted party and outputs its response. Thus, our peculiar
choice to feed both parties with the same auxiliary input and same random-tape is immaterial, because the
honest party ignores both.

633
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Proposition 7.2.11: Suppose that there exist one-way functions and that any single-
output functionality can be securely computed as per Definition 7.2.10. Then any func-
tionality can be securely computed as per Definition 7.2.6.

Proof Sketch: Suppose that the parties wish to securely compute the (two-output)
functionality (x, y) → ( f 1 (x, y), f 2 (x, y)). The first idea that comes to mind is to
first let the parties (securely) compute the first output (i.e., by securely computing
(x, y) → ( f 1 (x, y), λ)) and next let them (securely) compute the second output (i.e.,
by securely computing (x, y) → (λ, f 2 (x, y))). This solution is insecure, because a
malicious party may enter different inputs in the two invocations (not to mention that
the approach will fail for randomized functionalities even if both parties are honest).
Instead, we are going to let the first party obtain its output as well as an (authenticated
and) encrypted version of the second party’s output, which it will send to the second
party (which will be able to decrypt and verify the value). That is, we will use private-
key encryption and authentication schemes, which exist under the first hypothesis, as
follows. First, the second party generates an encryption/decryption-key, denoted e, and
a signing/verification-key, denoted s. Next, the two parties securely compute the ran-
domized functionality ((x, (y, e, s)) → (( f 1 (x, y), c, t) , λ), where c is the ciphertext
obtained by encrypting the plaintext v = f 2 (x, y) under the encryption-key e, and t is
an authentication-tag of c under the signing-key s. Finally, the first party sends (c, t) to
the second party, which verifies that c is properly signed and (if so) recovers f 2 (x, y)
from it.

7.3.* Privately Computing (Two-Party) Functionalities

Recall that our ultimate goal is to design (two-party) protocols that withstand any
feasible adversarial behavior. We proceed in two steps. In this section, we show how
to construct protocols for privately computing any functionality, that is, protocols that
are secure with respect to the semi-honest model. In Section 7.4, we will show how to
compile these protocols into ones that are secure also in the malicious model.
Throughout the current section, we assume that the desired (two-party) functionality
(along with the desired input length) is represented by a Boolean circuit. We show how to
transform this circuit into a two-party protocol for evaluating the circuit on a given pair
of local inputs. The transformation follows the outline provided in in Section 7.1.3.3.23
The circuit-evaluation protocol, to be presented in Section 7.3.4, scans the circuit
from the input wires to the output wires, processing a single gate in each basic step.
When entering each basic step, the parties hold shares of the values of the input wires of
the gate, and when the step is completed, they hold shares of the output wire of the gate.
The shares held by each party yield no information about the corresponding values, but
combining the two shares of any value allows for reconstructing the value. Each basic
step is performed without yielding any additional information; that is, the generation
of shares for all wires (and in particular for the circuit’s output wires) is performed in

23 Indeed, the current section is essentially a detailed version of Section 7.1.3.3.

634

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

a private manner. Put in other words, we will show that privately evaluating the circuit
“reduces” to privately evaluating single gates on values shared by both parties.
Our presentation is modular, where the modularity is supported by an appropriate
notion of a reduction. Thus, we first define such notion, and show that indeed it is
suitable to our goals; that is, combining a reduction of (the private computation of) g
to (the private computation of) f and a protocol for privately computing f yields a
protocol for privately computing g. Applying this notion of a reduction, we reduce the
private computation of general functionalities to the private computation of determin-
istic functionalities, and thus focus on the latter.
We next consider, without loss of generality, the evaluation of Boolean circuits with
and and xor gates of fan-in 2.24 Actually, we find it more convenient to consider the
corresponding arithmetic circuits over GF(2), where multiplication corresponds to and
and addition to xor. A value v is shared by the two parties in the natural manner (i.e.,
the sum of the shares equals v mod 2). We show how to propagate shares of values
through any given gate (operation). Propagation through an addition gate is trivial, and
we concentrate on propagation through a multiplication gate. The generic case is that
the first party holds (a1 , b1 ) and the second party holds (a2 , b2 ), where a1 + a2 is the
value of one input wire and b1 + b2 is the value of the other input wire. What we want
is to provide each party with a random share of the value of the output wire, that is, a
share of the value (a1 + a2 ) · (b1 + b2 ). In other words, we are interested in privately
computing the following randomized functionality
((a1 , b1 ), (a2 , b2 )) → (c1 , c2 ) (7.17)
where c1 + c2 = (a1 + a2 ) · (b1 + b2 ). (7.18)
That is, (c1 , c2 ) ought to be uniformly distributed among the pairs satisfying c1 + c2 =
(a1 + a2 ) · (b1 + b2 ). As shown in Section 7.3.3, this functionality can be privately
computed by reduction to a variant of Oblivious Transfer (OT). This variant is defined
in Section 7.3.2, where it is shown that this variant can be privately implemented
assuming the existence of (enhanced) trapdoor one-way permutations. We stress that
the specific functionalities mentioned here are relatively simple (e.g., they have a finite
domain). Thus, Section 7.3.4 reduces the private computation of arbitrary (complex)
functionalities to the construction of protocols for privately computing a specific simple
functionality (e.g., the one of Eq. (7.17) and Eq. (7.18)).
The actual presentation proceeds bottom-up. We first define reductions between (two-
party) protocol problems (in the semi-honest model). Next, we define and implement
OT, and show how to use OT for privately computing a single multiplication gate. Finally,
we show how to use the latter protocol to derive a protocol for privately evaluating the
entire circuit.
Teaching Tip. Some readers may prefer to see a concrete protocol (and its privacy
analysis) before coping with the abstract notion of a privacy reduction (and a corre-
sponding composition theorem). We advise such readers to read Section 7.3.2 before
reading Section 7.3.1.

24 Indeed, negation can be emulated by xoring the given bit with the constant true.

635
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

7.3.1. Privacy Reductions and a Composition Theorem

It is time to define what we mean by saying that the private computation of one func-
tionality reduces to the private computation of another functionality. Our definition
is a natural extension of the standard notion of a reduction as used in the context of
ordinary (i.e., one-party) computation. Recall that standard reductions are defined in
terms of oracle machines. Thus, we need to consider two-party protocols with oracle
access. Here, the oracle is invoked by both parties, each supplying it with one input (or
query), and it responds with a pair of answers, one per each party. We stress that the
answer-pair depends on the (entire) query-pair.

Definition 7.3.1 (protocols with oracle access): An oracle-aided protocol is an or-

dinary protocol augmented by pairs of oracle-tapes, one pair per each party, and
oracle-call steps defined as follows. Each of the parties may send a special oracle
request message, to the other party. Such a message is typically sent after this party
writes a string, called its query, on its own write-only oracle-tape. In response, the
other party also writes a string, called its query, on its own oracle-tape and responds
to the requesting party with an oracle call message. At this point, the oracle is in-
voked and the result is that a string, not necessarily the same, is written by the oracle
on the read-only oracle-tape of each party. This pair of strings is called the oracle
answer.

We stress that the syntax of Definition 7.3.1 allows (only) sequential oracle calls (but
not parallel ones). We call the reader’s attention to the second item in Definition 7.3.2
that requires that the oracle-aided protocol privately compute the functionality, rather
than merely computes it.

Definition 7.3.2 (privacy reductions):

r An oracle-aided protocol is said to be using the oracle-functionality f if the
oracle answers are according to f . That is, when the oracle is invoked, such that the
requesting party writes the query q1 and responding party writes the query q2 , the
answer-pair is distributed as f (q1 , q2 ), where the requesting party gets the first part
(i.e., f 1 (q1 , q2 )).25
We require that the length of each query be polynomially related to the length of the
initial input.26
r An oracle-aided protocol using the oracle-functionality f is said to privately com-
pute g if there exist polynomial-time algorithms, denoted S1 and S2 , satisfying
Eq. (7.9) and Eq. (7.10), respectively, where the corresponding views of the exe-
cution of the oracle-aided protocol are defined in the natural manner.

25 The identity of the requesting party may be determined by the two parties (according to interaction prior to the
request). In particular, as in all protocols used in this work, the identity of the requesting party may be fixed a
priori.
26 This requirement guarantees that the security of the oracle calls be related to the security of the high-level

protocol.

636

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

r An oracle-aided protocol is said to privately reduce g to f if it privately computes

g when using the oracle-functionality f . In such a case, we say that g is privately
reducible to f ,
We are now ready to state a composition theorem for the semi-honest model.

Theorem 7.3.3 (Composition Theorem for the semi-honest model): Suppose that g is
privately reducible to f and that there exists a protocol for privately computing f .
Then there exists a protocol for privately computing g.

Theorem 7.3.3 can be generalized to assert that if g is privately reducible to f , and f

is privately reducible to e, then g is privately reducible to e. See Exercise 5.
Proof Sketch: Let
g| f be a oracle-aided protocol that privately reduces g to f , and let

f be a protocol that privately computes f . We construct a protocol
for computing
g in the natural manner; that is, starting with
g| f , we replace each invocation of the
oracle (i.e., of f ) by an execution of the protocol
f . Clearly,
computes g. We need
to show that
privately computes g.
g| f f
For each i = 1, 2, let Si and Si be the corresponding simulators for the view of
Party i (i.e., in
and
, respectively). We construct a simulator Si , for the view of
g| f f
g| f
Party i in
, in the natural manner. That is, we first run Si and obtain the (simulated)
view of Party i in
g| f . This (simulated) view includes queries made by Party i and
corresponding answers. (Recall, we have only the part of Party i in the query-answer
f
pair.) Invoking Si on each such “partial query-answer,” we fill in the view of Party i
f
for each of these invocations of
f . (Note that we rely on the fact that the simulator Si
outputs a view that fits the output given to it; see Item 2 in the discussion that follows
Definition 7.2.1.)
A minor technicality: There is a minor inaccuracy in this description, which pre-
supposes that Party i is the party that plays the i-th party in
f (i.e., Party 1 is the
party in
g| f that requests all oracle calls to f ). But, in general, it may be that, in
some invocations of
f , Party 2 plays the first party in
f (i.e., Party 1 is the party
in
g| f that requests this particular oracle call). In this case, we should simulate the
execution of
f by using the simulator that simulates the view of the corresponding
party in
f (rather than the corresponding party in
).
Advanced comment: Note that we capitalize on the fact that in the semi-honest
model, the execution of the steps of
g| f (inside
) is independent of the actual
executions of
f (and depends only on the outcomes of
f ). This fact allows us
to first simulate a transcript of
g| f , and next generate simulated transcripts of
f .
In contrast, in the malicious model, the adversary’s actions in
g| f may depend on
the transcript of previous executions of
f , and thus this simulation strategy will
not work in the malicious model (and a more complex simulation strategy will be
used).

It is left to show that Si indeed generates a distribution that (augmented by the value
of g) is indistinguishable from the view of Party i (augmented by the output of both
parties) in actual executions of
. Toward this end, we introduce a hybrid distribution,
denoted Hi . This hybrid distribution represents the view of Party i (and the output of
637
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

both parties) in an execution of
g| f that is augmented by corresponding invocations of

f
Si . That is, for each query-answer pair, (q, a), viewed by Party i, we augment its view
f
with Si (q, a). In other words, Hi represents the execution of
, with the exception that
the invocations of
f are replaced by simulated transcripts.
Comment: We stress that since g may be a randomized functionality, we should
establish that the protocol satisfies the general form of Definition 7.2.1, rather than
its simplified form. That is, we consider the joint distribution consisting of the view
of Party i and the output of both parties (rather than merely the former). This fact
merely makes the phrases more cumbersome, and the essence of the argument may
be better captured by assuming that g is deterministic and using the special (simpler)
form of Definition 7.2.1. Likewise, in case f is randomized, we have to rely on the
general form of Definition 7.2.1 in order to show that the distributions represented
by Hi and
are computationally indistinguishable.
f g| f
Using the guarantees regarding Si (resp., Si ), we show that the distributions cor-
responding to Hi and
(resp., Hi and Si ) are computationally indistinguishable.
Specifically:
1. The distributions represented by Hi and
are computationally indistinguishable:
The reason is that these distributions differ only in that the invocations of
f in
f

are replaced in Hi by Si -simulated transcripts. Thus, the hypothesis regarding
f
Si implies that the two distributions are computationally indistinguishable (where
indistinguishability is measured with respect to the length of the queries, and holds
also when measured with respect to the length of the initial inputs).27 Specifically,
one may consider hybrids of
and Hi such that in the j-th hybrid, the first j
invocations of
f are real and the rest are simulated. Then distinguishability of
f
neighboring hybrids contradicts the hypothesis regarding Si (by incorporating a
possible transcript of the rest of the execution into the distinguisher).
2. The distributions represented by Hi and Si are computationally indistinguishable:
g| f
The reason is that these distributions are obtained, respectively, from
g| f and Si ,
f
by augmenting the latter with invocations of Si . Thus, indistinguishability follows
g| f
by the hypothesis regarding Si . Specifically, distinguishing Hi and Si implies dis-
g| f f
tinguishing
g| f and Si (by incorporating the program Si into the distinguisher).
The theorem follows.

Application: Reducing Private Computation of General Functionalities to Deter-

ministic Ones. Given a general functionality g, we first write it in a way that makes
the randomization explicit. That is, we let g(r, (x, y)) denote the value of g(x, y) when
using coin tosses r ∈ {0, 1}poly(|x|) ; that is, g(x, y) is the randomized process consisting
of uniformly selecting r ∈ {0, 1}poly(|x|) , and deterministically computing g(r, (x, y)).
Next, we privately reduce g to a deterministic f , where f is defined as follows:
def
f ((x1 , r1 ), (x2 , r2 )) = g(r1 ⊕ r2 , (x1 , x2 )) (7.19)

27 Here we use the hypothesis (made in the first item of Definition 7.3.2) that the length of each query is polynomially
related to the length of the initial input.

638

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

Applying Theorem 7.3.3 (while using a straightforward privacy reduction of g to f ),

we conclude that the existence of a protocol for privately computing the deterministic
functionality f implies the existence of a protocol for privately computing the ran-
domized functionality g. For sake of future reference, we explicitly state the privacy
reduction of g to f (i.e, the oracle-aided protocol for g given f ).

Proposition 7.3.4 (privately reducing a randomized functionality to a deterministic

one): Let g be a randomized functionality, and f be as defined in Eq. (7.19). Then the
following oracle-aided protocol privately reduces g to f .

Inputs: Party i gets input xi ∈ {0, 1}n .

Step 1: Party i uniformly selects ri ∈ {0, 1}poly(|xi |) .
Step 2 – Reduction: Party i invokes the oracle with query (xi , ri ), and records the oracle
response.
Outputs: Each party outputs the oracle’s response.
We comment that this construction is also applicable in the case of malicious adver-
saries; see Section 7.4.2.

Proof: Clearly, this protocol, denoted
, computes g. To show that
privately computes
g, we need to present a simulator for each party view. The simulator for Party i, denoted
Si , is the obvious one. On input (xi , vi ), where xi is the local input to Party i and vi
is its local output, the simulator uniformly selects ri ∈ {0, 1}m , and outputs (x i , ri , vi ),
where m = poly(|xi |). The main observation underlying the analysis of this simulator
is that for every fixed x1 , x2 and r ∈ {0, 1}m , we have v = g(r, (x1 , x2 )) if and only if
v = f ((x 1 , r1 ), (x2 , r2 )), for every pair (r1 , r2 ) satisfying r1 ⊕ r2 = r . Now, let ζi be a
random variable representing the random choice of Party i in Step 1, and ζi
denote the
corresponding choice made by the simulator Si . Then, referring to the general form of
Definition 7.2.1 (as we should, since g is a randomized functionality), we show that for
every fixed x1 , x2 , ri and v = (v1 , v2 ), it holds that

viewi
(x1 , x2 ) = (xi , ri , vi )
Pr = Pr[(ζi = ri ) ∧ ( f ((x1 , ζ1 ), (x2 , ζ2 )) = v)]
∧ output
(x1 , x2 ) = (v1 , v2 )
|{r3−i : f ((x1 , r1 ), (x2 , r2 )) = v}|
= Pr[ζi = ri ] ·
2m
|{r : g(r, (x 1 , x 2 )) = v}|
= 2−m ·
2m
= Pr[ζi
= ri ] · Pr[g(x1 , x2 ) = v]
= Pr[(ζi
= ri ) ∧ (g(x1 , x 2 ) = v)]

S (x , g (x , x )) = (xi , ri , vi )
= Pr i i i 1 2
∧ g(x1 , x 2 ) = (v1 , v2 )
where the equalities are justified as follows: the 1st by definition of
, the 2nd by
independence of the ζi ’s, the 3rd by definition of ζi and f , the 4th by definition of
ζi
and g, the 5th by independence of ζi
and g, and the 6th by definition of Si . Thus,
639
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

the simulated view (and output) is distributed identically to the view (and output) in a
real execution. The claim (which only requires these ensembles to be computationally
indistinguishable) follows.

7.3.2. The OTk1 Protocol: Definition and Construction

The (following version of the) Oblivious Transfer functionality is a main ingredient
of our construction. Let k be a fixed integer (k = 4 will do for our purpose), and let
σ1 , σ2 , ..., σk ∈ {0, 1} and i ∈ {1, ..., k}. Then, the (single-output) functionality 1-out-
of-k Oblivious Transfer, denoted OTk1 , is defined as

OTk1 ((σ1 , σ2 , ..., σk ), i) = (λ, σi ) (7.20)

Indeed, 1-out-of-k Oblivious Transfer, is asymmetric. Traditionally, the first player,

holding input (σ1 , σ2 , ..., σk ) is called the sender, whereas the second player, holding
the input i ∈ {1, ..., k} is called the receiver. Intuitively, the goal is to transfer the i-th bit
to the receiver, without letting the receiver obtain knowledge of any other bit and without
letting the sender obtain knowledge of the identity of the bit required by the receiver.
Using any enhanced trapdoor permutation, { f α : Dα → Dα }α∈I , we present a proto-
col for privately computing OTk1 . The following description refers to the algorithms
guaranteed by such a collection (see Definition 2.4.5 in Volume 1 and Definition C.1.1
in Appendix C) and to a hard-core predicate b for such a collection (see Section 2.5
of Volume 1). We denote the sender (i.e., the first party) by S and the receiver (i.e.,
the second party) by R. As discussed in Section 7.2.1, since we are dealing with a
finite functionality, we want the security to be stated in terms of an auxiliary security
parameter, n, presented to both parties in unary.

Construction 7.3.5 (Oblivious Transfer protocol for semi-honest model):

Inputs: The sender has input (σ1 , σ2 , ..., σk ) ∈ {0, 1}k , the receiver has input i ∈
{1, 2, ..., k}, and both parties have the auxiliary security parameter 1n .
Step S1: The sender uniformly selects an index-trapdoor pair, (α, t), by running the
generation algorithm, G, on input 1n . That is, it uniformly selects a random-tape, r ,
for G and sets (α, t) = G(1n , r ). It sends the index α to the receiver.
Step R1: The receiver uniformly and independently selects x1 , ..., xk ∈ Dα , sets yi =
f α (xi ) and y j = x j for every j = i, and sends (y1 , y2 , ..., yk ) to the sender. That is:

1. It uniformly and independently selects x1 , ..., xk ∈ Dα , by invoking the domain-

sampling algorithm k times, on input α. Specifically, it selects random-tapes, r j ’s,
for D and sets x j = D(α, r j ), for j = 1, ..., k.
2. Using the evaluation algorithm, the receiver sets yi = f α (xi ).
3. For each j = i, the receiver sets y j = x j .

640

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

4. The receiver sends (y1 , y2 , ..., yk ) to the sender.

(Thus, the receiver knows f α−1 (yi ) = xi , but cannot predict b( f α−1 (y j )) for any
j = i.)
Step S2: Upon receiving (y1 , y2 , ..., yk ), using the inverting-with-trapdoor algorithm
and the trapdoor t, the sender computes z j = f α−1 (y j ), for every j ∈ {1, ..., k}. It
sends (σ1 ⊕ b(z 1 ), σ2 ⊕ b(z 2 ), ..., σk ⊕ b(z k )) to the receiver.
Step R2: Upon receiving (c1 , c2 , ..., ck ), the receiver locally outputs ci ⊕ b(xi ).
We first observe that this protocol correctly computes OTk1 : This is the case since the
receiver’s local output (i.e., ci ⊕ b(xi )) satisfies

ci ⊕ b(xi ) = (σi ⊕ b(z i )) ⊕ b(xi )

= σi ⊕ b( f α−1 (yi )) ⊕ b(xi )
= σi ⊕ b( f α−1 ( f α (xi ))) ⊕ b(xi )
= σi

We show next that the protocol indeed privately computes OTk1 . Intuitively, the sender
gets no information from the execution because, for any possible value of i, the senders
sees the same distribution; specifically, a sequence of k uniformly and independently
distributed elements of Dα . (Indeed, the key observation is that applying f α to a uni-
formly distributed element of Dα yields a uniformly distributed element of Dα .) In-
tuitively, the receiver gains no computational knowledge from the execution since, for
j = i, the only data it has regarding σ j is the triplet (α, r j , σ j ⊕ b( f α−1 (x j ))), where
x j = D(α, r j ), from which it is infeasible to predict σ j better than by a random guess.
Specifically, we rely on the “enhanced one-way” hypothesis by which, given α and r j ,
it is infeasible to find f α−1 (x j ) (or guess b( f α−1 (x j )) better than at random). A formal
argument is indeed due and given next.

Proposition 7.3.6: Suppose that { fi : Di → Di }i∈I constitutes a collection of enhanced

trapdoor permutations (as in Definition C.1.1) and that b constitutes a hard-core predi-
cate for it. Then, Construction 7.3.5 constitutes a protocol for privately computing OTk1
(in the semi-honest model).

We comment that the intractability assumption used in Proposition 7.3.6 will propagate
to all subsequent results in the current and next section (i.e., Sections 7.3 and 7.4).
In fact, the implementation of OTk1 seems to be the bottleneck of the intractability
assumptions used in these sections.

Proof Sketch: Note that since we are dealing with a deterministic functionality, we
may use the special (simpler) form of Definition 7.2.1 (which only refers to each
party’s view). Thus, we will present a simulator for the view of each party. Recall that
these simulators are given the local input (which also includes the security parameter)
and the local output of the corresponding party. The following schematic depiction of

641
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

the information flow in Construction 7.3.5 may be useful toward the constructions of
these simulators:

Sender (S) Receiver (R)

input (σ1 , ..., σk ) i
Step S1 (α, t) ← G(1n )
−→ α −→
Step R1 generates y j ’s
←− (y1 , ..., yk ) ←− (knows xi = f α−1 (yi ))
Step S2 c j = σ j ⊕ b( f α−1 (y j ))
−→ (c1 , ..., ck ) −→
R2 (output) λ ci ⊕ b(xi )

We start by presenting a simulator for the sender’s view. On input (((σ1 , ..., σk ), 1n ), λ),
this simulator randomly selects α (as in Step S1) and generates uniformly and inde-
pendently y1 , ..., yk ∈ Dα . That is, let r denote the sequence of coins used to generate
α, and assume without loss of generality that the inverting-with-trapdoor algorithm
is deterministic (which is typically the case anyhow). Then the simulator outputs
(((σ1 , ..., σk ), 1n ), r, (y1 , ..., yk )), where the first element represents the party’s input,
the second its random choices, and the third the (single) message that the party
has received. Clearly, this output distribution is identical to the view of the sender
in the real execution. (This holds because f α is a permutation, and thus applying
it to a uniformly distributed element of Dα yields a uniformly distributed element
of Dα .)
We now turn to the receiver. On input ((i, 1n ), σi ), the simulator (of the receiver’s
view) proceeds as follows:

1. Emulating Step S1, the simulator uniformly selects an index-trapdoor pair, (α, t), by
running the generation algorithm on input 1n .
2. As in Step R1, it uniformly and independently selects r1 , ..., rk for the domain sampler
D, and sets x j = D(α, r j ) for j = 1, ..., k. Next, it sets yi = f α (xi ) and y j = x j , for
each j = i.
3. It sets ci = σi ⊕ b(xi ), and uniformly selects c j ∈ {0, 1}, for each j = i.
4. Finally, it outputs ((i, 1n ), (r1 , ..., rk ), (α, (c1 , ..., ck ))), where the first element repre-
sents the party’s input, the second its random choices, and the third element represents
the two messages that the party has received.

Note that, except for the sequence of c j ’s, this output is distributed identically to the
corresponding prefix of the receiver’s view in the real execution. Furthermore, the said
equality holds even if we include the bit ci (which equals σi ⊕ b( f α−1 (yi )) = σi ⊕ b(xi )
in the real execution as well as in the simulation). Thus, the two distributions dif-
fer only in the values of the other c j ’s: For j = i, in the simulation c j is uni-
form and independent of anything else, whereas in the real execution c j equals

642

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

b( f α−1 (y j )) = b( f α−1 (x j )) (and hence depends on r j , which determines x j ). However,

it is impossible to distinguish the two cases, because x j is uniformly distributed and
the distinguisher is only given α and r j (but not the trapdoor to f α ). Here is where we
use the hypothesis that b is a hard-core of an enhanced collection of trapdoor permu-
tations (as in Definition C.1.1), rather than merely a standard collection of trapdoor
permutations.

Other Variants of Oblivious Transfer. A variety of different variants of the Oblivious

Transfer functionality were considered in the literature, but most treatments refer to the
(more challenging) problem of implementing these variants securely in the malicious
model (rather than in the semi-honest model). We briefly mention two of these other
variants:

1. Extensions of 1-out-of-k Oblivious Transfer to k secrets that are bit strings rather
than single bits.
2. Oblivious Transfer of a single secret (denoted σ ) that is to be delivered with prob-
ability 1/2. That is, the randomized functionality that maps (σ, λ) to (λ, σ ) with
probability 1/2 and to (λ, λ) otherwise.

Privacy reductions among these variants can be easily constructed (see Exercise 6).

7.3.3. Privately Computing c1 + c2 = (a1 + a2 ) · (b1 + b2 )

We now turn to the functionality defined in Eq. (7.17) – (7.18). Recall that this func-
tionality is a randomized mapping ((a1 , b1 ), (a2 , b2 )) → (c1 , c2 ) satisfying c1 + c2 =
(a1 + a2 ) · (b1 + b2 ), where the arithmetic is in GF(2). We reduce the private compu-
tation of this (finite) functionality to (the private computation of) OT41 .

Construction 7.3.7 (privately reducing the functionality of Eq. (7.17) – (7.18) to OT41 ):

Inputs: Party i holds (ai , bi ) ∈ {0, 1} × {0, 1}, for i = 1, 2.

Step 1: The first party uniformly selects c1 ∈ {0, 1}.
Step 2 – Reduction: The aim of this step is to privately compute the (residual)
deterministic functionality ((a1 , b1 , c1 ), (a2 , b2 )) → (λ, f a2 ,b2 (a1 , b1 , c1 )), where
def
f a,b (x, y, z) = z + (x + a) · (y + b). The parties privately reduce the computation
of this functionality to OT41 . Specifically, Party 1 plays the sender and Party 2 plays
the receiver. Using its input (a1 , b1 ) and coin c1 , Party 1 sets the sender’s input (in
the OT41 ) to equal the 4-tuple

( f 0,0 (a1 , b1 , c1 ) , f 0,1 (a1 , b1 , c1 ) , f 1,0 (a1 , b1 , c1 ) , f 1,1 (a1 , b1 , c1 )) . (7.21)

Using its input (a2 , b2 ), Party 2 sets the receiver’s input (in the OT41 ) to equal 1 +
2a2 + b2 ∈ {1, 2, 3, 4}.

643
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Thus, the receiver’s output will be the (1 + 2a2 + b2 )th element in Eq. (7.21), which
in turn equals f a2 ,b2 (a1 , b1 , c1 ). That is:

Input of Party 2 Receiver’s input in OT41 Receiver’s output in OT41

(i.e., (a2 , b2 )) (i.e., 1 + 2a2 + b2 ) (i.e., fa2 ,b2 (a1 , b1 , c1 ))
(0, 0) 1 c1 + a1 b1
(0, 1) 2 c1 + a1 · (b1 + 1)
(1, 0) 3 c1 + (a1 + 1) · b1
(1, 1) 4 c1 + (a1 + 1) · (b1 + 1)

Recall that f a2 ,b2 (a1 , b1 , c1 ) = c1 + (a1 + a2 ) · (b1 + b2 ).

Outputs: Party 1 outputs c1 , whereas Party 2 outputs the result obtained from the OT41
invocation.
We first observe that the reduction is valid; that is, when Party i enters with input (ai , bi ),
the output of Party 2 equals f a2 ,b2 (a1 , b1 , c1 ) = c1 + (a1 + a2 ) · (b1 + b2 ), where c1
is the output of Party 1. That is, the output pair is uniformly distributed among the
pairs (c1 , c2 ) for which c1 + c2 = (a1 + a2 ) · (b1 + b2 ) holds. Thus, each of the local
outputs (i.e, of either Party 1 or Party 2) is uniformly distributed, although the two local
outputs are dependent of one another (as in Eq. (7.18)). It is also easy to see that the
reduction is private. That is,

Proposition 7.3.8: Construction 7.3.7 privately reduces the computation of

Eq. (7.17) – (7.18) to OT41 .

Proof Sketch: Simulators for the oracle-aided protocol of Construction 7.3.7 are easily
constructed. Specifically, the simulator of the view of Party 1 has input ((a1 , b1 ), c1 )
(i.e., the input and output of Party 1), which is identical to the view of Party 1 in
the corresponding execution (where here c1 serves as coins to Party 1). Thus, the
simulation is trivial (i.e., by the identity transformation). The same also holds for
the simulator of the view of Party 2: It gets input ((a2 , b2 ), c1 + (a1 + a2 ) · (b1 + b2 ))
(i.e., the input and output of Party 2), which is identical to the view of Party 2 in the
corresponding execution (where here c1 + (a1 + a2 ) · (b1 + b2 ) serves as the oracle
response to Party 2). Thus, again, the simulation is trivial. We conclude that the view of
each party can be perfectly simulated (rather than just be simulated in a computationally
indistinguishable manner). The same holds when we also account for the parties’ outputs
(as required in the general form of Definition 7.2.1), and the proposition follows.28

On the Generic Nature of Construction 7.3.7. The idea underlying Step 2 of Con-
struction 7.3.7 can be applied in order to reduce the computation of any determinis-
tic functionality of the form (x, y) → (λ, f y (x)) to 1-out-of-2|y| Oblivious Transfer.
Indeed, this reduction is applicable only when y is short (i.e., the number of pos-
sible y’s is at most polynomial in the security parameter). Specifically, consider the

28 An alternative proof is presented in Exercise 9.

644

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

functions f y : {0, 1}k → {0, 1}, for y ∈ {0, 1} (when in Construction 7.3.7  = 2 (and
k = 3)). Then, privately computing (x, y) → (λ, f y (x)) is reduced to 1-out-of-2 Obliv-
ious Transfer by letting the first party play the sender with input set to the 2 -tuple
( f 0 (x), ..., f 1 (x)) and the second party play the receiver with input set to the index of
y among the -bit long strings.

7.3.4. The Circuit Evaluation Protocol

We now show that the computation of any deterministic functionality, which is repre-
sented by an arithmetic circuit over GF(2), is privately reducible to the functionality of
Eq. (7.17) – (7.18). Recall that the latter functionality corresponds to a private compu-
tation of multiplication of inputs that are shared by the two parties. We thus refer to this
functionality as the multiplication-gate emulation.
Our reduction follows the overview presented in the beginning of this section (i.e.,
Section 7.3). In particular, the sharing of a bit-value v between the two parties
means a uniformly distributed pair of bits (v1 , v2 ) such that v = v1 + v2 , where the first
party holds v1 and the second holds v2 . Our aim is to propagate, via private computation,
shares of the input-wires of the circuit to shares of all wires of the circuit, so that finally
we obtain shares of the output-wires of the circuit.
Arithmetic circuits – the basics: Recall that an arithmetic circuit over GF(2) is a
directed acyclic graph with internal vertices corresponding to gates, where inter-
nal vertices are vertices having both incoming and outgoing edges. Without loss
of generality, we will consider two types of gates, called addition and multiplica-
tion. We will assume that each internal vertex has two incoming edges, called its
input-wires, and several outgoing edges called its output-wires. Boolean values are
propagated through such gates in the natural manner (i.e., each outgoing wire holds
the sum or multiple of the values of the incoming wires of the gate). Vertices with no
incoming edges are called sources, and vertices with no outgoing edges are called
sinks. Without loss of generality, each source has a single outgoing edge, which
is called an input-wire of the circuit, and each sink has a single incoming edge,
which is called an output-wire of the circuit. When placing Boolean values on the
input-wires of the circuit, the propagation of values through the gates determines
values to all output-wires. The function from input values to output values defined
this way is called the function computed by the circuit.
A tedious comment: For the sake of simplicity, we do not provide the circuit with
constant values (i.e., 0 and 1). The constant 0 can be easily produced by adding
any GF(2) value to itself, but omitting the constant 1 weakens the power of such
circuits (because this constant is essential to the computation of non-monotone
functions). However, the evaluation of any circuit that uses the constant 1 can be
privately reduced to the evaluation of a corresponding circuit that does not use the
constant 1.29
29 Given a circuit C that uses the constant 1, we derive a circuit C
that lacks constant inputs by introducing an
auxiliary variable that is to be set to 1 (and replacing any occurrence of the constant 1 by an occurrence of
the new auxiliary variable). Thus, C(x) = C
(x, 1) (or rather C(x 1 , x2 ) = C
(1x1 , σ x 2 ), for any σ ∈ {0, 1}).
Clearly, the private evaluation of C (on the input pair (x1 , x 2 )) is reducible to the private evaluation of C
(e.g.,
by a single oracle call that asks for the evaluation of C
, say, on the input (1x1 , 0x2 )).

645
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

We will consider an enumeration of all wires in the circuit. The input-wires of the
circuit, n per each party, will be numbered 1, 2...., 2n so that, for j = 1, ..., n, the j-th
input of party i corresponds to the (i − 1) · n + j-th wire. The wires will be numbered
so that the output-wires of each gate have a larger numbering than its input wires.
The output-wires of the circuit are clearly the last ones. For the sake of simplicity we
assume that each party obtains n output bits, and that the output bits of the second party
correspond to the last n wires of the circuit.

Construction 7.3.9 (reducing the evaluation of any circuit to the emulation of a

multiplication gate): For simplicity, we assume that the circuit is either fixed or can be
determined in poly(n)-time as a function of n, which denotes the length of the input to
each party.30

Inputs: Party i holds the bit string xi1 · · · xin ∈ {0, 1}n , for i = 1, 2.
Step 1 – Sharing the Inputs: Each party (splits and) shares each of its input bits with
the other party. That is, for every i = 1, 2 and j = 1, ..., n, Party i uniformly se-
j
lects a bit ri and sends it to the other party as the other party’s share of the input
wire (i − 1) · n + j. Party i sets its own share of the (i − 1) · n + j th input wire
j j
to xi + ri .
Step 2 – Circuit Emulation: Proceeding by the order of wires, the parties use their
shares of the two input-wires to a gate in order to privately compute shares for the
output-wire(s) of the gate. Suppose that the parties hold shares to the two input-wires
of a gate; that is, Party 1 holds the shares a1 , b1 and Party 2 holds the shares a2 , b2 ,
where a1 , a2 are the shares of the first wire and b1 , b2 are the shares of the second
wire. We consider two cases.31
Emulation of an addition gate: Party 1 just sets its share of the output-wire of the
gate to be a1 + b1 , and Party 2 sets its share of the output-wire to be a2 + b2 .
Emulation of a multiplication gate: Shares of the output-wire of the gate are obtained
by invoking the oracle for the functionality of Eq. (7.17) – (7.18), where Party 1
supplies the input (query part) (a1 , b1 ), and Party 2 supplies (a2 , b2 ). When the
oracle responds, each party sets its share of the output-wire of the gate to equal
its part of the oracle answer. Recall that, by Eq. (7.18), the two parts of the oracle
answer sum up to (a1 + b1 ) · (a2 + b2 ).

Step 3 – Recovering the Output Bits: Once the shares of the circuit-output wires are
computed, each party sends its share of each such wire to the party with which the
wire is associated. That is, the shares of the last n wires are sent by Party 1 to Party 2,

30 Alternatively, we may let the circuit be part of the input to both parties, which essentially means that the protocol
is computing the “universal circuit-evaluation” function.
31 In the text, we implicitly assume that each gate has a single output wire, but this assumption is immaterial and

the treatment extends easily to the case that the gates have several output wires. In the case of a multiplication
gate, both the natural possibilities (which follow) are fine. The first (more natural) possibility is to invoke the
oracle once per each multiplication gate and have each party use the same share for all output wires. The second
possibility is to invoke the oracle once per each output-wire (of a multiplication gate).

646

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

whereas the shares of the preceding n wires are sent by Party 2 to Party 1. Each
party recovers the corresponding output bits by adding up the two shares, that is,
the share it had obtained in Step 2 and the share it has obtained in the current step.
Outputs: Each party locally outputs the bits recovered in Step 3.
For starters, let us verify that the output is indeed correct. This can be shown by induction
on the wires of the circuits. The induction claim is that the shares of each wire sum up
to the correct value of the wire. The base case of the induction are the input-wires of the
j j
circuits. Specifically, the (i − 1) · n + j-th wire has value xi , and its shares are ri and
j j j
ri + xi (indeed, summing up to xi ). For the induction step we consider the emulation
of a gate. Suppose that the values of the input-wires (to the gate) are a and b, and that
their shares a1 , a2 and b1 , b2 indeed satisfy a1 + a2 = a and b1 + b2 = b. In the case
of an addition gate, the shares of the output-wire were set to be a1 + b1 and a2 + b2 ,
indeed satisfying
(a1 + b1 ) + (a2 + b2 ) = (a1 + a2 ) + (b1 + b2 ) = a + b
In the case of a multiplication gate, the shares of the output-wire were set to be c1 and
c2 such that c1 + c2 = (a1 + a2 ) · (b1 + b2 ). Thus, c1 + c2 = a · b as required.

Privacy of the Reduction. We now turn to show that Construction 7.3.9 indeed
privately reduces the computation of a circuit to the multiplication-gate emulation.
That is,

Proposition 7.3.10 (privately reducing circuit evaluation to multiplication-gate emu-

lation): Construction 7.3.9 privately reduces the evaluation of arithmetic circuits over
GF(2) to the functionality of Eq. (7.17) – (7.18).

Proof Sketch: Note that since we are dealing with a deterministic functionality, we
may use the special (simpler) form of Definition 7.2.1 and only refer to simulating
the view of each party. Recall that these simulators should produce the view of the
party in an oracle-aided execution (i.e., an execution of Construction 7.3.9, which is an
oracle-aided protocol). Without loss of generality, we present a simulator for the view
of Party 1. This simulator gets the party’s input x11 , ..., x1n , as well as its output, denoted
y 1 , ..., y n . It operates as follows:
j
1. The simulator uniformly selects r11 , ..., r1n and r21 , ..., r2n , as in Step 1. (The r1 ’s will
be used as the coins of Party 1, which are part of the view of the execution, whereas
j
the r2 ’s will be used as the message Party 1 receives at Step 1.) For each j ≤ n, the
j j
simulator sets x 1 + r1 as the party’s share of the value of the j-th wire. Similarly,
j
for j ≤ n, the party’s share of the n + j-th wire is set to r2 .
This completes the computation of the party’s shares of all the 2n circuit-input wires.
2. The party’s shares for all other wires are computed, iteratively gate by gate, as
follows:
r The party’s share of the output-wire of an addition gate is set to be the sum of the
party’s shares of the input-wires of the gate.
647
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

r The party’s share of the output-wire of a multiplication gate is selected uniformly

in {0, 1}.
(The shares computed for output-wires of multiplication gates will be used as the
answers obtained, by Party 1, from the oracle.)
3. For each wire corresponding to an output, denoted y j , that is available to Party 1, the
simulator sets the value z j to equal the sum of y j and the party’s share of that wire.
4. The simulator outputs
((x11 , ..., x1n ), (y 1 , ..., y n ), (r11 , ..., r1n ), V 1 , V 2 , V 3 )
where V 1 = (r21 , ..., r2n ) corresponds to the view of Party 1 in Step 1 of the protocol,
the string V 2 equals the concatenation of the bits selected for the output-wires of
multiplication gates (corresponding to the party’s view of the oracle answers in Step 2
of a real execution), and V 3 = (z 1 , ..., z n ) corresponds to the party’s view in Step 3
(i.e., the messages it would have obtained from Party 2 in Step 3 of the execution).
We claim that the output of the simulation is distributed identically to the view of Party 1
in the execution of the oracle-aided protocol. The claim clearly holds with respect to
the first four parts of the view; that is, the claim holds with respect to the party’s input
(i.e., x11 , ..., x1n ), its output (i.e., y 1 , ..., y n ), its internal coin-tosses (i.e., r11 , ..., r1n ), and
the message obtained from Party 2 in Step 1 (i.e., (r21 , ..., r2n ) = V 1 ). Also, by defi-
nition of the functionality of Eq. (7.17) – (7.18), the oracle answers to each party are
uniformly distributed independently of (the parts of) the party’s queries. Thus, this part
of the view of Party 1 is uniformly distributed, identically to V 2 . It follows that all
shares held by Party 1 are set by the simulator to have exactly the same distribution
as they have in a real execution. This holds, in particular, for the shares of the output
wires held by Party 1. Finally, we observe that both in the real execution and in the
simulation, adding the latter shares (i.e., the shares of the output wires held by Party 1)
to the messages sent by Party 2 in Step 3 (resp., to V 3 ) yields the corresponding bits
of the local output of Party 1. Thus, conditioned on the view so far, V 3 is distributed
identically to the messages sent by Party 2 in Step 3. We conclude that the simu-
lation is perfect (not only computationally indistinguishable), and so the proposition
follows.

Conclusion. Combining Propositions 7.3.4, 7.3.10, and 7.3.8 with the transitivity of
privacy reductions (see Exercise 5), we obtain:

Theorem 7.3.11: Any functionality is privately reducible to OT41 .

Combining Theorem 7.3.11 and Proposition 7.3.6 with the Composition Theorem (The-
orem 7.3.3), we obtain:32

32 Alternatively, one may avoid relying on the transitivity of privacy reductions by successively applying the Com-
position Theorem to derive private protocols first for the multiplication functionality, then for any deterministic
functionality, and finally for any functionality. That is, in the first application we use Propositions 7.3.8 and 7.3.6,

648

www.Ebook777.com
 7.3* PRIVATELY COMPUTING (TWO-PARTY) FUNCTIONALITIES

Theorem 7.3.12: Suppose that there exist collections of enhanced trapdoor per-
mutations. Then any functionality can be privately computable (in the semi-honest
model).

For the sake of future usage (in Section 7.4), we point out a property of the protocols
underlying the proof of Theorem 7.3.12.

Definition 7.3.13 (canonical semi-honest protocols): A protocol
for privately com-

puting the functionality f is called canonical if it proceeds by executing the following
two stages:

Stage 1: The parties privately compute the functionality (x, y) → ((r1 , r2 ), (s1 , s2 )),
where the ri ’s and si ’s are uniformly distributed among all possibilities that satisfy
(r1 ⊕ s1 , r2 ⊕ s2 ) = f (x, y).
Stage 2: Party 2 sends s1 to Party 1, which responds with r2 . Each party computes its
own output; that is, Party i outputs ri ⊕ si .
Indeed, the protocols underlying the proof of Theorem 7.3.12 are canonical. Hence,

Theorem 7.3.14: Suppose that there exist collections of enhanced trapdoor permuta-
tions. Then any functionality can be privately computable by a canonical protocol.

We present two alternative proofs of Theorem 7.3.14: The first proof depends on the
structure of the protocols used in establishing Theorem 7.3.11, whereas the second
proof is generic and uses an additional reduction.

First Proof of Theorem 7.3.14: Recall that the oracle-aided protocol claimed in The-
orem 7.3.11 is obtained by composing the reduction in Proposition 7.3.4 with Con-
structions 7.3.9 and 7.3.7. The high-level structure of the resulting protocol is induced
by the circuit-evaluation protocol (of Construction 7.3.9), which is clearly canonical
(with Step 3 fitting Stage 2 in Definition 7.3.13). Indeed, it is important that in Step 3
exactly two messages are sent and that Party 1 sends the last message. The fact that the
said oracle-aided protocol is canonical is also preserved when replacing the OT41 oracle
by an adequate sub-protocol.

Second Proof of Theorem 7.3.14: Using Theorem 7.3.12, we can first derive a protocol
for privately computing the functionality of Stage 1 (in Definition 7.3.13). Augment-
ing this protocol by the trivial Stage 2, we derive a canonical protocol for privately
computing the original functionality (i.e., f itself).

in the second we use Proposition 7.3.10 and the protocol resulting from the first application, and in the last
application we use Proposition 7.3.4 and the protocol resulting from the second application.

649
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

7.4.* Forcing (Two-Party) Semi-Honest Behavior

Our aim is to use Theorem 7.3.12 (or rather Theorem 7.3.14) in order to establish the
main result of this chapter; that is,
Theorem 7.4.1 (main result for the two-party case): Suppose that there exist collections
of enhanced trapdoor permutations. Then any two-party functionality can be securely
computable (in the malicious model).
Theorem 7.4.1 will be established by compiling any protocol for the semi-honest model
into an “equivalent” protocol for the malicious model. The current section is devoted
to the construction of the said compiler, which was already outlined in Section 7.1.3.1.
Loosely speaking, the compiler works by replacing the original instructions by macros
that force each party to either effectively behave in a semi-honest manner (hence, the
title of the current section) or be detected as cheating (in which case, the protocol aborts).

Teaching Tip. Some readers may prefer to see a concrete protocol (and its security
analysis) before getting to the general protocol compiler (and coping with the abstrac-
tions used in its exposition). We advise such readers to read Section 7.4.3.1 before
reading Sections 7.4.1 and 7.4.2.

7.4.1. The Protocol Compiler: Motivation and Overview

We are given a protocol for the semi-honest model. In this protocol, each party has a
local input and uses a uniformly distributed local random-tape. Such a protocol may be
used to privately compute some functionality (either a deterministic or a probabilistic
one), but the compiler does not refer to this functionality. The compiler is supposed to
produce an “equivalent protocol” for the malicious model. That is, any input–output
behavior that a malicious adversary can induce by attacking the resulting protocol
can also be induced by a semi-honest adversary that attacks the original protocol. To
motivate the protocol compiler, let us start by considering what a malicious party may
do (beyond whatever a semi-honest party can do).

1. A malicious party may enter the actual execution of the protocol with an input differ-
ent from the one it is given (i.e., “substitute its input”). As discussed in Section 7.2.3,
this is unavoidable. What we need to guarantee is that this substitution is done obliv-
iously of the input of the other party, that is, that the substitution only depends on
the original input.
Jumping ahead, we mention that the input-commitment phase of the compiled proto-
col is aimed at achieving this goal. The tools used here are commitment schemes (see
Section 4.4.1) and strong zero-knowledge proofs-of-knowledge (see Section 4.7.6).
Sequential executions of these proofs-of-knowledge guarantee the effective indepen-
dence of the committed values.
2. A malicious party may enter the actual execution of the protocol with a random-
tape that is not uniformly distributed. What we need to do is force the party
650

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

to use a random-tape (for the emulated semi-honest protocol) that is uniformly

distributed.
The coin-generation phase of the compiled protocol is aimed at achieving this goal.
The tool used here is an augmented coin-tossing into the well protocol, which in turn
uses tools as in Item 1.
3. A malicious party may try to send messages different from the ones specified by
the original (semi-honest model) protocol. So we need to force the party to send
messages as specified by its (already committed) local input and random-tape.
The protocol-emulation phase of the compiled protocol is aimed at achieving this
goal. The tool used here is zero-knowledge proof systems (for NP-statements). In fact,
forcing parties to act consistently with some known information is the archetypical
application of zero-knowledge proofs.
In accordance with this discussion, the protocols produced by the compiler consist of
three phases.

Input-Commitment Phase: Each of the parties commits to its input by using a

secure implementation of the input-commitment functionality (to be defined in
Section 7.4.3.6). The latter functionality guarantees that the committing party actu-
ally knows the value to which it has committed, and that the secrecy of the committed
value is preserved. It follows that each party commits to a value that is essentially
independent of the value committed to by the other party. Furthermore, the input-
commitment functionality provides the committer with the corresponding decom-
mitment information (to be used in the protocol-emulation phase).

Coin-Generation Phase: The parties generate random-tapes for the emulation of the
original protocol. Each party obtains the value of the random-tape to be held by it,
whereas the other party obtains a commitment to this value. The party holding the
value also obtains the corresponding decommitment information. All this is obtained
by using a secure implementation of the (augmented) coin-tossing functionality (to
be defined in Section 7.4.3.5). It follows that each party obtains a random-tape that
is essentially random and independent of anything else.

Protocol Emulation Phase: The parties use a secure implementation of the

authenticated-computation functionality (to be defined in Section 7.4.3.4) in order to
emulate each step of the original protocol. Specifically, each message transmission
in the original protocol is replaced by an invocation of the said sub-protocol (imple-
menting this functionality), where the roles played by the parties and the inputs fed
to the sub-protocol are as follows. The party that is supposed to send the message in
the original protocol invokes the sub-protocol with an input that consists of its initial
input (as committed in the first stage), its random-tape (as generated in the second
stage), the decommitment information provided to it in the two corresponding stages,
and the sequence of all incoming messages (of the original protocol as emulated so
far). The input provided by the other party (i.e., the designated receiver) consists of
the commitments it holds for the sender’s input and random-tape (received in the first
and second stage), as well as the sequence of all messages that it has previously sent
651
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

to the sender. The functionality guarantees that either the corresponding (next-step)
message is delivered or the designated receiver detects cheating.
In order to allow a modular presentation of the compiled protocols, we start by defining
an adequate notion of reducibility (where here the oracle-aided protocol needs to be
secure in the malicious model rather than in the semi-honest one). We next turn to
constructing secure protocols for several basic functionalities, and use the latter to
construct secure protocols for the three main functionalities mentioned here. Finally,
we present and analyze the actual compiler.

7.4.2. Security Reductions and a Composition Theorem

Analogously to Section 7.3.1, we now define what we mean by saying that one func-
tionality securely reduces to another functionality. We use the same definition of an
oracle-aided protocol (i.e., Definition 7.3.1), but require such a protocol to be secure
in the malicious model (rather than secure in the semi-honest model, as required in
Definition 7.3.2). Recall that the basic syntax of an oracle-aided protocol allows se-
quential (but not parallel) oracle calls. For simplicity of our exposition, we require that
the length of each oracle-query can be determined from the length of the initial input
to the oracle-aided protocol.

Definition 7.4.2 (security reductions):

r As in Definition 7.3.2, an oracle-aided protocol is said to be using the oracle-
functionality f if the oracle answers are according to f . However, in accordance
with the definition of the ideal model (for the invoked functionality), the oracle does
not answer both parties concurrently, but rather answers first the real-model party
that requested this specific oracle call (in the oracle-aided protocol). When receiving
its part of the oracle answer, this party (i.e., the real-model party that requested the
oracle call) instructs the oracle whether or not to respond to the other party.
We consider only protocols in which the length of each oracle-query is a polynomial-
time computable function of the length of the initial input to the protocol. Further-
more, as in Definition 7.3.2, the length of each query must be polynomially related
to the length of the initial input.
We consider executions of such a protocol by a pair of parties, with strategies repre-
sented by probabilistic polynomial-time algorithms A1 and A2 , such that one of the
parties follows the oracle-aided protocol. Such a pair is called admissible. Analo-
gously to Definition 7.2.5, the joint execution of an oracle-aided protocol
with
oracle f under A = (A1 , A2 ) in the real model (on input pair (x, y) and auxiliary
f
input z), denoted real
, A(z) (x, y), is defined as the output pair resulting from the
interaction between A1 (x, z) and A2 (y, z), where oracle calls are answered using f .
We stress that here the real model corresponds to an execution of an oracle-aided
protocol.
r An oracle-aided protocol
using the oracle-functionality f is said to securely
compute g if a condition analogous to the one in Definition 7.2.6 holds. That is, the

652

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

effect of any admissible real-model strategies as in the previous item can be simulated
by admissible strategies for the ideal model, where the ideal model for computing g
is exactly as in Definition 7.2.4.
More specifically, the oracle-aided protocol
(using oracle f ) is said to securely
compute g (in the malicious model) if for every probabilistic polynomial-time
pair A = (A1 , A2 ) that is admissible for the real model of the oracle-aided com-
putation, there exists a probabilistic polynomial-time pair B = (B1 , B2 ) that is
admissible for the ideal model (of Definition 7.2.4) such that

c f
{idealg, B(z) (x, y)}x, y,z ≡ {real
, A(z) (x, y)}x, y,z

where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).
r An oracle-aided protocol is said to securely reduce g to f if it securely computes
g when using the oracle-functionality f . In such a case, we say that g is securely
reducible to f ,
We are now ready to state a composition theorem for the malicious model.

Theorem 7.4.3 (Composition Theorem for the malicious model): Suppose that g is
securely reducible to f and that there exists a protocol for securely computing f . Then
there exists a protocol for securely computing g.

Recall that the syntax of oracle-aided protocols disallows concurrent oracle calls, and
thus Theorem 7.4.3 is actually a sequential composition theorem. As in the semi-
honest case, the Composition Theorem can be generalized to yield transitivity of secure
reductions; that is, if g is securely reducible to f and f is securely reducible to e, then
g is securely reducible to e (see Exercise 13).
As hinted in Section 7.3.1, the proof of Theorem 7.4.3 is significantly more complex
than the proof of Theorem 7.3.3. This does not refer to the construction of the resulting
protocol, but rather to establishing its security.

Proof Sketch: Analogously to the proof of Theorem 7.3.3, we are given an oracle-aided
protocol, denoted
g| f , that securely reduces g to f , and an ordinary protocol
f that
securely computes f . Again, we construct a protocol
for computing g in the natural
manner; that is, starting with
g| f , we replace each invocation of the oracle (i.e., of f )
by an execution of the protocol
f .
Clearly,
computes g, and we need to show that
securely computes g. Specifically,
we should present a transformation of real-model adversaries for
into ideal-model
adversaries for g. We have at our disposal two transformations of real-model adver-
saries (for
g| f and for
f ) into corresponding ideal-model adversaries (for g and f ,
respectively). So the first thing we should do is derive, from the real-model adversaries
of
, real-model adversaries for
g| f and for
f .
We assume, without loss of generality, that all real-model adversaries output their
view of the execution. (Recall that any other output can be efficiently computed from
the view, and that any adversary can be easily modified to output its view.)
653
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Let A = (A1 , A2 ) be an admissible pair of real-model strategies of
. We first derive

from it a pair of strategies A = (A
1 , A
2 ) that represents the behavior of A during (each
of) the invocations of
f . (We stress that we derive a single pair of real-model strategies
that represents the behavior of the adversary during all the invocations of
f .)33 Since
the honest Ai just behaves according to
, it follows that the induced Ai
just behaves
according to
f , which means that Ai
is honest. Thus, we focus on the other (i.e.,
dishonest) Ai . In this case, the derived Ai
is a real-model adversary of
f that gets as
auxiliary input the history of the execution of
up to the current invocation of
f .
Formally, Ai
takes two inputs, one representing (as usual) the history of the current
execution of
f , and the other (i.e., an auxiliary one) being the history of the execution
of
up to the current invocation of
f . When Ai
completes (or aborts) the current
execution of
f , it outputs its view of that execution. Loosely speaking, we derive


the corresponding ideal-model adversary for f , denoted B = (B1
, B2
), by employing
the guaranteed transformation. A few technical difficulties arise and are resolved as
follows:

r Party i (i.e., A
) is not necessarily the party that plays the i-th party in
f (i.e.,
i
Party 1 is not necessarily the party in
g| f that requests this particular oracle call
to f ). Furthermore, the identity of the party (in
f ) played by Ai
is not fixed,
but is rather determined by the history of the execution of
(which is given to
Ai
as auxiliary input). In contrast, our definitions refer to adversaries that play a
predetermined party. This technical discrepancy can be overcome by considering
two versions of Ai
, denoted Ai,1


and Ai,2 , such that Ai,
j in used (instead of Ai
) in
case Party i is the party that plays the j-th party in
f . Indeed, Ai,
j is always used
to plays the j-th party in
f .
r A minor problem is that Ai may have its own auxiliary input, in which case the
resulting Ai
will have two auxiliary inputs (i.e., the first identical to the one of
Ai , and the second representing a partial execution transcript of
). Clearly, these
two auxiliary inputs can be combined into a single auxiliary input. (This fact holds
generically, but more so in this specific setting in which it is anyhow natural to
incorporate the inputs to an adversary in its view of the execution transcript.)
r The last problem is that it is not clear what “initial input” should be given to the
adversary Ai
toward its current execution of
f (i.e., the input that is supposed to be
used for computing f ). However, this problem (which is more confusing than real)
has little impact on our argument, because what matters is the actual actions of Ai

during the current execution of
f , and these are determined based on its (actual)
auxiliary input (which represent the history of the execution of
). Still, the initial
inputs for the executions of
f have to be defined so that they can be passed to
the ideal-model adversary that we derive from Ai
. We may almost set these initial
inputs arbitrarily, except that (by our conventions regarding functionalities) we must

33 The simpler alternative of deriving a different pair of (real-model) strategies for each invocation of
f would
have sufficed for handling oracle-aided protocols that make a constant number of oracle calls. The point is
that the corresponding ideal-model strategies (with respect to f ) need to be combined into a single real-model
strategy for
g| f .

654

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

set them to strings of correct length (i.e., equal to the length of the other party’s
f -input). Here we use the hypothesis that this length can be determined from the
length of the input to
itself.34


Thus, we have obtained an (admissible) ideal-adversary pair B = (B1
, B2
) correspond-
ing to f such that
c
{ideal f, B
(z
) (x
, y
)}x
, y
,z
≡ {real
f , A
(z
) (x
, y
)}x
, y
,z
(7.22)

We comment that when applying Eq. (7.22), we set the input of the honest party to
equal the value on which the sub-protocol (or functionality) was invoked, and set the
auxiliary input to equal the current execution transcript of the high-level protocol (as
seen by the adversary). (As explained earlier, the setting of the primary input to the
dishonest party is immaterial, because the latter determines its actions according to its
auxiliary input.)



Our next step is to derive from A = (A1 , A2 ) a pair of strategies A = (A

1 , A

2 ) that
represents the behavior of A during the
g| f -part of
. Again, the honest Ai induces
a corresponding Ai

that just behaves according to
g| f . Turning to the dishonest Ai ,
we derive Ai

by replacing the (real) actions of Ai
that take place in Ai by simulated
actions of the ideal-model Bi
. That is, the adversary Ai

runs machine Ai locally, while
interacting with the actual other party of
g| f , obtaining the messages that Ai would
have sent in a real execution of
, and feeding Ai with messages that it expects to
receive (i.e., messages that Ai would have received in a real execution of
). The
handling of Ai ’s messages depends on whether they belong to the
g| f -part or to one
of the invocations of
f . The key point is the handling of the latter messages.

Handling Messages of
g| f : These messages are forwarded to/from the other party
without change. That is, Ai

uses Ai in order to determine the next message to be
sent, and does so by feeding Ai with the history of the execution so far (which
contains
g| f -part messages that Ai

has received before, as well as the
f -
parts that it has generated so far by itself). In particular, if Ai aborts, then so
does Ai

.
Handling Messages of
f : Upon entering a new invocation of
f , the adversary Ai


sets h i to record the history of the execution of
so far. Now, rather than executing

f using Ai
(h i ) (as Ai would have done), the adversary Ai

invokes Bi
(h i ), where
Bi
is the ideal-model adversary for f (derived from Ai
, which in turn was derived
from Ai ). Recall that Bi
sends no messages and makes a single oracle-query (which it
views as sending a message to its imaginary trusted party). The real-model adversary
Ai

(for the oracle-aided protocol
g| f ) forwards this query to its own oracle (i.e.,

34 We comment that when using the alternative conventions discussed at the end of Section 7.2.1.1, we may waive
the requirement that the query length be determined by the input length. Instead, we postulate that all oracle
calls made by the oracle-aided program use the same security parameter as the one with which the program is
invoked. On the other hand, under the current conventions, when trying to extend the composition theorem to
partial functionalities (or when removing the “length determination” hypothesis), we run into trouble because
we need to determine some f -input that fits the unknown f -input of the other party. (This problem can be
resolved by introducing an adequate interface to oracle calls.)

655
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

f ), and feeds Bi
with the oracle answer. At some point Bi
terminates, and Ai

uses
its output to update the simulated history of the execution of
. In particular, oracle-
stopping events caused by Bi
(h i ) (in case Party i requested this specific oracle call)
and ⊥-answers of the oracle (in the other case) are handled in the straightforward
manner.
On stopping the oracle and ⊥-answers: Suppose first that Party i has re-
quested this specific oracle call. In this case, after receiving the oracle answer
(which it views as the answer of its trusted party), the ideal-model adversary Bi

may stop its trusted party. If this happens, then machine Ai

instructs its own
oracle (i.e., f ) not to respond to the other party. Next, suppose that Party i is
the party responding to this specific oracle call (rather than requesting it). In this
case, it may happen that the oracle is stopped by the other party (i.e., the oracle
is not allowed to answer Party i). When notified of this event (i.e., receiving a
⊥-answer from its oracle), machine Ai

feeds ⊥ as answer to Bi
.

This completes the handling of the current invocation of
f .

When Ai halts with some output, Ai

halts with the same output. Note that A = (A

1 , A

2 )
is admissible as a real-model adversary for the oracle-aided protocol
g| f (which



computes g with oracle to f ). Thus, we can derive from A a corresponding ideal-



model adversary for g, denoted B = (B1

, B2

), by employing the second guaranteed
transformation, such that
c f
{idealg, B

(z) (x, y)}x, y,z ≡ {real

(x, y)}x, y,z (7.23)

g| f , A (z)

Thus, given a real-model adversary A for
, we have derived an ideal-model adversary

def


B = B for g. It is left to show that indeed the following holds:
c
{idealg, B(z) (x, y)}x, y,z ≡ {real
, A(z) (x, y)}x, y,z (7.24)
Note that the left-hand side of Eq. (7.24) equals the left-hand side of Eq. (7.23), so
it suffices to show that their corresponding right-hand sides are computationally in-
f
distinguishable. But real
, A(z) (x, y) differs from real g| f

(x, y) only in that the

, A (z)

f - invocations in the former are replaced in the latter by ideal calls to f . However,
by Eq. (7.22), each
f invocation is computationally indistinguishable from an ideal
call to f , where computational indistinguishability holds also with respect to auxiliary
inputs (which are used here to represent the execution transcript upto the point of the
current invocation). Using a hybrid argument (corresponding to a gradual substitution
f
of
f -invocations by ideal calls to f ), one can show that {real g| f

(x, y)}x, y,z

, A (z)
and {real
, A(z) (x, y)}x, y,z are computationally indistinguishable.35 This establishes
Eq. (7.24), and the theorem follows.

35 Here we use the hypothesis that the query lengths are polynomially related to the length of the input. The issue is
that in Eq. (7.22), computational indistinguishability is with respect to the length of the queries (to f ), whereas
we need computational indistinguishability with respect to the length of the initial inputs. We also highlight the



key role of the auxiliary inputs to A and B in this argument (cf. the analysis of the sequential composition of
zero-knowledge [i.e., proof of Lemma 4.3.11]).

656

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

Security Reduction of General Functionalities to Deterministic Ones. The follow-

ing reduction will not be used in our compiler, because the compiler refers to protocols
(rather to functionalities), and we have already obtained protocols for privately com-
puting general functionalities (by privately reducing them to deterministic ones). Still,
we consider it of interest to state that the reduction presented in Proposition 7.3.4 is, in
fact, secure in the malicious model.

Proposition 7.4.4 (securely reducing a randomized Functionality to a Deterministic

One): Let g be a randomized functionality, f be as defined in Eq. (7.19), and
be the
oracle-aided protocol for g using the oracle f as presented in Proposition 7.3.4. Then

securely computes g.

Proof Sketch: Suppose, without loss of generality, that Party 1 is malicious, and denote
by (x1
, r1
) the query it makes to f . Denoting by xi the initial input of Party i (in
), it
follows that the oracle answer is f ((x1
, r1
), (x2 , r2 )), where r2 is uniformly distributed
(because Party 2 is honest). Recalling that f ((x1
, r1
), (x2 , r2 )) = g(r1
⊕ r2 , (x1
, x2 )), it
follows that the oracle answer is distributed identically to g(x1
, x 2 ). Furthermore, by
the definition of
, all that Party 1 gets is f 1 ((x 1
, r1
), (x2 , U|r1
| )) ≡ g1 (x1
, x2 ). This is
easily simulated by a corresponding ideal-model adversary that sets x1
according to the
real-model adversary, and sends x1
to the trusted third party (which answers according
to g).

Remark 7.4.5 (reductions to a set of functionalities): We extend the notion of se-

curity reductions to account for protocols that use several oracles rather than one.
Specifically, g is securely reducible to a set of functionalities F = { f 1 , ..., f t } if there
exists an oracle-aided protocol that securely computes g when given oracles f 1 , ..., f t .
Theorem 7.4.3 also extends to assert that if g is securely reducible to F, and each
functionality in F can be securely computed, then so can g. We comment that the
entire remark is a matter of semantics, because one can “pack” the set F in one func-
def
tionality f (e.g., f ((i, x), (i, y)) = f i (x, y)).

7.4.3. The Compiler: Functionalities in Use

As stated in Section 7.4.1, the protocols produced by our compiler make extensive use
of protocols that securely compute three functionalities that are the core of the three
corresponding phases of the compiled protocols. In the current section, we explicitly
define these functionalities and present protocols for securely computing them.
We start by considering three natural functionalities that are related to the functional-
ities used by the compiler. Specifically, we first consider the coin-tossing functionality
(see Section 7.4.3.1), a restricted notion of the authenticated-computation functional-
ity (Section 7.4.3.2), and an “unauthenticated-computation functionality” (called image
transmission in Section 7.4.3.3). Next, using these three functionalities, we present se-
cure protocols for a general notion of authenticated-computation functionality (see
Section 7.4.3.4), for an augmented notion of coin-tossing (Section 7.4.3.5), and for the
657
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Commitment
ORDER OF PRESENTATION schemes ZK proofs ZK POKs TOOLS

1
COIN
TOSSING 2
restricted
AUTH. C. 3
IMAGE
TRANS.

4
5 AUTH.
COMMIT
6 AUG. Comput.
COIN
INPUT

THE COMPILED PROTOCOL

Figure 7.2: The functionalities used in the compiled protocol.

input-commitment functionality (Section 7.4.3.6). The latter three functionalities will

be used directly in the compiled protocols (see Figure 7.2, where solid arrows indicate
direct and essential use). We comment that although the material in Section 7.4.3.2
is not used directly in the rest of this work, it is instructive to the rest of the current
section.
We comment that it is easy to present protocols for privately computing all the
abovementioned functionalities (in the semi-honest model; see Exercise 11). Our aim,
however, is to present (for later use in the compiler) protocols for securely computing
these functionalities in the malicious model.

Basic Tools and Conventions Regarding Them. Let us recall some facts and notations
regarding three tools that we will use:

r Commitment schemes (as defined in Definition 4.4.1). For the sake of simplicity, we
will use a non-interactive commitment scheme (as in Construction 4.4.2). We assume,
for simplicity, that on security parameter n, the commitment scheme utilizes exactly
n random bits. We denote by Cr (b) the commitment to the bit b using (security
parameter n and) randomness r ∈ {0, 1}n , and by C(b) the value of Cr (b) for a
uniformly distributed r ∈ {0, 1}n (where n is understood from the context).
r Zero-knowledge proofs of NP-assertions. We rely on the fact (cf. Theorem 4.4.11)
that there exist such proof systems in which the prover strategy can be implemented
in probabilistic polynomial-time, when given an NP-witness as auxiliary input. We
stress that by this we mean (zero-knowledge) proof systems with negligible sound-
ness error. Furthermore, we rely on the fact that these proof systems have perfect
completeness (i.e., the verifier accepts a valid statement with probability 1).
658

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

r Zero-knowledge proofs-of-knowledge of NP-witnesses. We will use the definition of

a strong proof of knowledge (see Definition 4.7.13). We again rely on the analogous
fact regarding the complexity of adequate prover strategies: That is, strong proofs-
of-knowledge that are zero-knowledge exist for any NP-relation, and furthermore,
the prover strategy can be implemented in probabilistic polynomial-time, when given
an NP-witness as auxiliary input (see Construction 4.7.14).
All these tools are known to exist assuming the existence of one-way 1-1 functions. In
fact, the 1-1 requirement can be avoided at the cost of using an interactive commitment
scheme.

On the Adversaries Being Considered. For the sake of simplicity, in all the proofs of
security presented in this section, we only refer to malicious (real-model) adversaries
with no auxiliary input. Furthermore, we will assume that these malicious (real-model)
adversaries are deterministic. As discussed in Section 7.2.3.1 (see text following Def-
inition 7.2.5), the treatment of randomized adversaries (with auxiliary inputs) can be
reduced to the treatment of deterministic adversaries with auxiliary inputs, and so the
issue here is actually the fact that we ignore auxiliary inputs. However, in all cases,
the extension of our treatment to malicious adversaries with auxiliary input is straight-
forward. Specifically, in all cases, we construct ideal-model adversaries by using the
real-model adversaries as subroutines. This black-box usage easily supports the ex-
tension to adversaries with auxiliary inputs, because all that is needed is to pass the
auxiliary-input (given to the ideal-model adversary) to the real-model adversary (which
is invoked as a subroutine).

Comments Regarding the Following Exposition. All protocols are presented by spec-
ifying the behavior of honest parties, while keeping in mind that dishonest parties may
deviate from the specified behavior. Thus, we may instruct one party to send a specific
message that satisfies some property and next instruct the other party to check that the
message received indeed satisfies this property. When transforming real-model adver-
saries to ideal-model adversaries, we sometimes allow the latter to halt before invoking
the trusted party. As discussed in Section 7.2.3.1 (see text preceding Definition 7.2.4),
this can be viewed as invoking the trusted party with a special abort symbol, where in
this case, the latter responds to all parties with a special abort symbol.

7.4.3.1. Coin-Tossing
We start our assembly of functionalities that are useful for the compiler by presenting
and implementing a very natural functionality, which is of independent interest. Specif-
ically, we refer to the coin-tossing functionality (1n , 1n ) → (b, b), where b is uniformly
distributed in {0, 1}. This functionality allows a pair of distrustful parties to agree on a
common random value.36

36 Actually, in order to conform with the convention that the functionality has to be defined for any input pair, we
may consider the formulation (x, y) → (b, b).

659
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Definition 7.4.6 (coin-tossing into the well, basic version): A coin-tossing-into-the-

well protocol is a two-party protocol for securely computing (in the malicious model) the
randomized functionality (1n , 1n ) → (b, b), where b is uniformly distributed in {0, 1}.
That is, in spite of malicious behavior by any one party, a non-aborting execution of a
coin-tossing-into-the-well protocol ends with both parties holding the same uniformly
distributed bit, b. Recall that our definition of security allows (b, ⊥) to appear as output
in case Party 1 aborts. (It would have been impossible to securely implement the coin-
tossing functionality if the definition had not allowed this slackness; see Section 7.7.1.1.)
The coin-tossing functionality will not be used directly in the compiled protocols, but
it will be used to implement an augmented notion of coin-tossing (see Section 7.4.3.5),
which in turn will be used directly in these protocols.

Construction 7.4.7 (a coin-tossing-into-the-well protocol): For every r , let Cr :

{0, 1} → {0, 1}∗ .

Inputs: Both parties get security parameter 1n .

def
Step C1: Party 1 uniformly selects σ ∈ {0, 1} and s ∈ {0, 1}n , and sends c = Cs (σ ) to
Party 2.
To simplify the exposition, we adopt the convention by which failure of Party 1 to
send a message (i.e., aborting) is interpreted as an arbitrary bit string, say C0n (0).
Step C2: Party 2 uniformly selects σ
∈ {0, 1}, and sends σ
to Party 1.
Similarly, any possible response of Party 2, including abort, will be interpreted by
Party 1 as a bit.37
Step C3: Party 1 outputs the value σ ⊕ σ
, and sends (σ, s) to Party 2.
Step C4: Party 2 checks whether or not c = Cs (σ ). It outputs σ ⊕ σ
if c = Cs (σ ) and
halts with output ⊥ otherwise.
In contrast to Steps C1–C2, here any illegal answer is interpreted as abort.
def
Outputs: Party 1 always outputs b = σ ⊕ σ
, whereas Party 2 either outputs b
or ⊥.
Intuitively, Steps C1–C2 may be viewed as “tossing a coin into the well.” At this point,
the value of the coin is determined (essentially as a random value), but only one party
knows (“can see”) this value. Clearly, if both parties are honest, then they both output
the same uniformly chosen bit, recovered in Steps C3 and C4, respectively.

Proposition 7.4.8: Suppose that C is a bit commitment scheme. Then, Construc-

tion 7.4.7 constitutes a coin-tossing-into-the-well protocol.

Proof Sketch: We need to transform any admissible pair, (A1 , A2 ), for the real model
into a corresponding pair, (B1 , B2 ), for the ideal model. We treat separately each of the

37 These two conventions prevent the parties from aborting the execution before Step C3.

660

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

two cases corresponding to the identity of the honest party. Recall that we may assume,
for simplicity, that the adversary is deterministic (see discussion toward the end of the
preamble of Section 7.4.3). Also, for simplicity, we omit the input 1n in some places.
The following schematic depiction of the information flow in Construction 7.4.7 may
be useful toward the following analysis:

Party 1 Party 2
C1 selects (σ, s)
c ← Cs (σ ) −→ c −→
C2 selects σ
∈ {0, 1}
←− σ
←−


C3 b ←σ ⊕σ
−→ (σ, s) −→
output b b or ⊥
(depending on whether c = Cs (σ ))

We start with the case where the first party is honest. In this case, B1 is determined
(by the protocol), and we transform the real-model adversary A2 into an ideal-model
adversary B2 . Machine B2 will run machine A2 locally, obtaining the single message that
A2 would have sent in a real execution of the protocol (i.e., σ
∈ {0, 1}) and feeding A2
with the messages that it expects to receive. Recall that A2 expects to see the messages
Cs (σ ) and (σ, s) (and that B2 gets input 1n ).

1. B2 sends 1n to the trusted party and obtains an answer (bit), denoted b, which is
uniformly distributed. (Recall that b is also handed to Party 1.)
2. B2 tries to generate an execution view (of A2 ) ending with output b. This is done by
repeating the following steps at most n times:
def
(a) B2 uniformly select σ ∈ {0, 1} and s ∈ {0, 1}n , and feeds A2 with c = Cs (σ ).
Recall that A2 always responds with a bit, denoted σ
, which may depend on c
(i.e., σ
← A2 (c)).
(b) If σ ⊕ σ
= b, then B2 feeds A2 with the execution view (c, (σ, s)), and outputs
whatever A2 does. Otherwise, it continues to the next iteration.
In case all n iterations were completed unsuccessfully (i.e., without output), B2
outputs a special failure symbol.
We need to show that for the coin-tossing functionality, denoted f , and for Construc-
tion 7.4.7, denoted
, it holds that
c
{ideal f, B (1n , 1n )}n∈N ≡ {real
, A (1n , 1n )}n∈N

In fact, we will show that the two ensembles are statistically indistinguishable. We start
by showing that the probability that B2 outputs failure is exponentially small. This
is shown by proving that for every b ∈ {0, 1}, each iteration of Step 2 succeeds with
probability approximately 1/2. Such an iteration succeeds if and only if σ ⊕ σ
= b,
661
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

that is, if A2 (Cs (σ )) = b ⊕ σ , where (σ, s) ∈ {0, 1} × {0, 1}n is uniformly chosen.
We have
Prσ,s [A2 (Cs (σ )) = b ⊕ σ ]
1 1
= · Pr[A2 (C(0)) = b] + · Pr[A2 (C(1)) = b ⊕ 1]
2 2
1 1
= + · (Pr[A2 (C(0)) = b] − Pr[A2 (C(1)) = b])
2 2
Using the hypothesis that C is a commitment scheme, the second term is a negligible
function in n, and so our claim regarding the probability that B2 outputs failure
follows. Letting µ denote an appropriate negligible function, we state the following for
future reference:
1
Prσ,s [A2 (Cs (σ )) = b ⊕ σ ] = ± µ(n) (7.25)
2
Next, we show that conditioned on B2 not outputting failure, the distribution
ideal f, B (1n , 1n ) is statistically indistinguishable from the distribution real
, A (1n , 1n ).
Both distributions have the form (b , A2 (Cs (σ ), (σ, s))), with b = σ ⊕ A2 (Cs (σ )), and
thus both are determined by the (σ, s)-pairs. In real
, A (1n , 1n ), all (σ, s)-pairs are
equally likely (i.e., each appears with probability 2−(n+1) ); whereas (as proven next) in
ideal f, B (1n , 1n ), each pair (σ, s) appears with probability
1 1
· (7.26)
2 |Sσ ⊕A2 (Cs (σ )) |
def
where Sb = {(x, y) ∈ {0, 1} × {0, 1}n : x ⊕ A2 (C y (x)) = b} is the set of pairs that pass
the condition in Step 2b (with respect to the value b obtained in Step 1). To justify
Eq. (7.26), observe that the pair (σ, s) appears as output if and only if it is selected
in Step 2a and the trusted party answers with σ ⊕ A2 (Cs (σ )), where the latter event
occurs with probability 1/2. Furthermore, the successful pairs, selected in Step 2a and
passing the condition in Step 2b, are uniformly distributed in Sσ ⊕A2 (Cs (σ )) , which justifies
Eq. (7.26). We next show that |Sb | ≈ 2n , for every b ∈ {0, 1}. By Eq. (7.25), for every
fixed b ∈ {0, 1} and uniformly distributed (σ, s) ∈ {0, 1} × {0, 1}n , the event (σ, s) ∈ Sb
(i.e., σ ⊕ A2 (Cs (σ )) = b) occurs with probability that is negligibly close to 1/2, and
so |Sb | = (1 ± µ(n)) · 12 · 2n+1 , where µ is a negligible function. Thus, for every pair
(σ, s), it holds that |Sσ ⊕A2 (Cs (σ )) | ∈ {|S0 |, |S1 |} resides in the interval (1 ± µ(n)) · 2n . It
follows that the value of Eq. (7.26) is (1 ± µ(n)) · 2−(n+1) , and so real
, A (1n , 1n ) and
ideal f, B (1n , 1n ) are statistically indistinguishable.
We now turn to the case where the second party is honest. In this case, B2 is de-
termined, and we transform A1 into B1 (for the ideal model). On input 1n , machine
B1 runs machine A1 locally, obtaining the messages that A1 would have sent in a real
execution of the protocol and feeding A1 with the single message (i.e., σ
∈ {0, 1}) that
it expects to receive.
1. B1 invokes A1 (on input 1n ). Recall that by our conventions, A1 always sends a mes-
sage in Step C1. Let us denote this message (which is supposedly a commitment
662

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

using C) by c. Recall that c may be in the range of C(σ ) for at most one
σ ∈ {0, 1}.

2. Machine B1 tries to obtain the answers of A1 (in Step C3) to both possible messages
that could be sent in Step C2:
(a) B1 feeds A1 with the (Step C2) message 0 and records the answer, which is either
abort or (σ0 , s0 ). The case in which c = Cs0 (σ0 ) is treated as if A1 has aborted.
(b) Rewinding A1 to the beginning of Step C2, machine B1 feeds A1 with the message
1 and records the answer, which is either abort or (σ1 , s1 ). (Again, the case in
which c = Cs1 (σ1 ) is treated as abort.)
If A1 aborts in both cases, then machine B1 aborts with output A1 (1n , σ
), for a
uniformly chosen σ
∈ {0, 1} (and does so without invoking the trusted party, which
means that the honest Party 2 receives ⊥ from the latter).38 (In the following, we refer
to this case as to Case 0.) Otherwise, B1 proceed as follows, distinguishing two cases:
Case 1: A1 answers properly (in the previous experiment) for a single 0-1 value,
def
denoted σ
. In this case, we define σ = σσ
.
Case 2: A1 answers properly for both values. In this case, the values σ0 and σ1
(defined in Step 1) must be identical, because C s0 (σ0 ) = c = Cs1 (σ1 ), whereas the
def
ranges of C(0) and C(1) are disjoint. In this case, we define σ = σ0 (= σ1 ).

3. Machine B1 sends 1n to the trusted party, which responds with a uniformly selected
value b ∈ {0, 1}. Recall that the trusted party has not responded to Party 2 yet, and
that B1 still has the option of stopping the trusted party before it responds to Party 2.

4. In Case 1, machine B1 stops the trusted party if b = σ ⊕ σ
(where σ
is as defined

in Case 1), and otherwise allows it to send b to Party 2 (in which case b = σ ⊕ σ

holds). In Case 2, machine B1 sets σ
= b ⊕ σ and allows the trusted party to send
b to Party 2. Next, in both cases, B1 feeds σ
to A1 , which responds with the Step C3
message (σ, sσ
). Note that if the trusted party sent b to Party 2, then indeed σ ⊕ σ
=
b holds (in both Case 1 and Case 2).

5. Finally, B1 feeds A1 with the execution view, (1n , σ
), and outputs whatever A1 does.
We now show that ideal f, B (1n , 1n ) and real
, A (1n , 1n ) are actually identically dis-
tributed. Consider first the case where A1 (and so B1 ) never aborts (i.e., Case 2). In this
case, we have

ideal f, B (1n , 1n ) = (A1 (1n , σ ⊕ b) , b)

real
, A (1n , 1n ) = (A1 (1n , σ
) , σ ⊕ σ
)

where σ
and b are uniformly distributed in {0, 1}, and σ is determined by c = A1 (1n )
(i.e., σ = C −1 (c)). Observe that σ
is distributed uniformly independently of σ , and so
38 We comment that whenever B1 is determined to abort, it need not invoke the trusted party at all, because it (i.e.,
B1 ) can simulate the trusted party’s answer by itself. The only reason to invoke the trusted party is to provide
Party 2 with an answer that is related to the output of B1 .

663
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

σ ⊕ σ
is uniformly distributed over {0, 1}. We conclude that ( A1 (1n , σ ⊕ b) , b) and

(A1 (1n , σ ⊕ (σ ⊕ σ
)) , σ ⊕ σ
) are identically distributed.
Next, consider the case that B1 always aborts (due to improper A1 behavior in
Step C3). In this case (i.e., the previous Case 0), B1 aborts before invoking the trusted
party, and so both ensembles are identical (i.e., both equal ( A1 (1n , σ
), ⊥) for a random
σ
). Since A1 is deterministic (see beginning of the proof), the only case left to consider
is where A1 responds properly (in Step C3) to a single value, denoted σ
. In this case
(i.e., Case 1), the real execution of
is completed only if Party 2 sends σ
as its Step C2
message (which happens with probability 1/2), and is aborted otherwise. Similarly, in
the ideal model, the execution is completed (without B1 aborting) if the trusted party
answers with b = σ ⊕ σ
(which happens with probability 1/2).39 In both models,
the joint non-aborted execution equals (A1 (1n , σ
) , σ ⊕ σ
), whereas the joint aborted
execution equals (A1 (1n , σ
⊕ 1) , ⊥).

7.4.3.2. Authenticated Computation (Partial Version)

We continue our assembly of functionalities that are useful for the compiler by pre-
senting and implementing another natural functionality, which is of independent in-
terest. Specifically, we refer to the archetypical application of zero-knowledge proofs
(cf. Section 4.4.3), which is to solve the following problem. For two predetermined
(polynomial-time computable) functions, f and h, a party holding a secret α should
send the correct value of f (α) to the other party, which holds h(α), while not revealing
anything else to the other party. That is, we are talking about securely computing the
functionality (α, h(α)) → (λ, f (α)), where typically h is 1-1 (and so the value of its
image uniquely determines its preimage).
We stress that the functionality described here has a partial domain; that is, it is not
defined over all pairs of inputs (of equal length), but rather only for pairs of the form
(α, h(α)). This restriction (i.e., definability over a partial domain) coincides with the
standard archetypical application of zero-knowledge proofs and is easier to implement.
However, this restriction does not suffice for a modular exposition of the compiled
protocols (because composition of partial functionalities is more complex than the
composition result captured by Theorem 7.4.3). Indeed, in Section 7.4.3.4 we waive
the restriction (to the partial domain) and consider an extension of the authenticated
computation functionality to arbitrary pairs of (equal-length) strings.

Definition 7.4.9 (authenticated computation, partial version): Let f : {0, 1}∗ ×

{0, 1}∗ → {0, 1}∗ and h : {0, 1}∗ → {0, 1}∗ be polynomial-time computable. The h-
authenticated f -computation functionality is defined by

(α, h(α)) → (λ, f (α)) (7.27)

We assume, for simplicity, that h is length preserving. Otherwise, the definition may
be modified to consider the functionality ((α, 1|h(α)| ) , (h(α), 1|α| )) → (λ, f (α)). To

39 Recall that, in this case, σ and σ
are determined by the Step C1 message.

664

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

facilitate the implementation, we assume that the function h is one-to-one, as is the case
in typical applications. This allows us to use (ordinary) zero-knowledge proofs, rather
than strong (zero-knowledge) proofs-of-knowledge. The issue is further discussed in
Section 7.4.3.3.
The functionality of Eq. (7.27) is implemented by having Party 1 send f (α) to
Party 2, and then prove in zero-knowledge the correctness of the value sent (with
respect to the common input h(α)). Note that this statement is of the NP type and
that Party 1 has the corresponding NP-witness. Actually, the following protocol is the
archetypical application of zero-knowledge proof systems.

Construction 7.4.10 (authenticated computation protocol, partial version): Let L be

the set of pairs (u, v) satisfying Eq. (7.28) to follow, and (P, V ) be an interactive proof
for L. Furthermore, suppose that P can be implemented in probabilistic polynomial-
time when given an adequate auxiliary-input (i.e., an NP-witness for membership of
the common input in L).
Inputs: Party 1 gets input α ∈ {0, 1}∗ , and Party 2 gets input u = h(α).
def
Step C1: Party 1 sends v = f (α) to Party 2.
Step C2: The parties invoke the proof system (P, V ) such that Party 1 plays the prover
and Party 2 plays the verifier. The common input to the proof system is (u, v), the
prover gets auxiliary input α, and its objective is to prove that
∃x s.t. (u = h(x)) ∧ (v = f (x)) (7.28)
(Each party locally determines the common input (u, v) according to its own view of
the execution so far.)40 In case the verifier rejects the proof, Party 2 halts with output
⊥ (otherwise the output will be v).
(Any possible response – including abort – of Party 2 during the execution of this
step will be interpreted by Party 1 as a canonical legitimate message.)
Outputs: In case Party 2 has not halted with output ⊥ (indicating improper behavior of
Party 1), Party 2 sets its local output to v. (Party 1 has no output [or, alternatively,
always outputs λ].)

Observe that the specified strategies are indeed implementable in polynomial-time. In

particular, in Step C2, Party 1 supplies the prover subroutine with the NP-witness α such
that Eq. (7.28) is satisfied with x = α. Also, using the perfect completeness condition
of the proof system, it follows that if both parties are honest, then neither aborts and
the output is as required.

Proposition 7.4.11: Suppose that the function h is one-to-one and that (P, V ) is
a zero-knowledge interactive proof (with negligible soundness error) for L. Then,
Construction 7.4.10 securely computes (in the malicious model) the h-authenticated
f -computation functionality of Eq. (7.27).

40 In particular, Party 1 sets (u, v) = (h(α), f (α)), whereas Party 2 sets u according to its own input and v according
to the message received in Step C1.

665
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

We stress that Proposition 7.4.11 refers to the security of a protocol for computing a
partial functionality, as discussed in Remark 7.2.7. In the case of Eq. (7.27), this means
that the ideal-model adversary is not allowed to “modify its input” (i.e., it must pass its
initial input to the trusted party), because its initial input is the unique value that fits
the other party’s input.
Proof Sketch: Again, we need to transform any admissible pair, (A1 , A2 ), for the real
model into a corresponding pair, (B1 , B2 ), for the ideal model. We treat separately each
of the two cases, corresponding to the identity of the honest party.
We start with the case where the first party is honest. In this case, B1 is determined,
and we transform (the real-model adversary) A2 into (an ideal-model adversary) B2 ,
which uses A2 as a subroutine. Recall that B2 gets input u = h(α), where α is the input
of the honest Party 1.
1. B2 sends u to the trusted party and obtains the value v, which equals f (α) for α
handed by (the honest) Party 1 to the trusted party. Thus, indeed, B2 does not modify
its input and (u, v) ∈ L. (Recall that Party 1 always obtains λ from the trusted party.)
2. B2 invokes the simulator guaranteed for the zero-knowledge proof system (P, V ), on
input (u, v), using (the residual) A2 as a possible malicious verifier.41 Note that we are
simulating the actions of the prescribed prover P, which in the real protocol is played
by the honest Party 1. Denote the obtained simulation transcript by S = S(u, v),
where (indeed) A2 is implicit in the notation.
3. Finally, B2 feeds A2 with the alleged execution view (v, S), and outputs whatever A2
does.
We need to show that for the functionality, denoted F, of Eq. (7.27) and for Construc-
tion 7.4.10, denoted
, it holds that
c
{ideal F, B (α, h(α))}α∈{0,1}∗ ≡ {real
, A (α, h(α))}α∈{0,1}∗ (7.29)
Let R(α) denote the verifier’s view of the real interaction with P on common input
(h(α), f (α)) and prover’s auxiliary input α, where the verifier is played by A2 . Then,
real
, A (α, h(α)) = (λ , A2 (h(α), f (α), R(α)))
ideal F, B (α, h(α)) = (λ , A2 (h(α), f (α), S(h(α), f (α))))
However, by the standard formulation of zero-knowledge, it follows that {R(α)}α∈{0,1}∗
and {S(h(α), f (α))}α∈{0,1}∗ are computationally indistinguishable (also when given α
as auxiliary input), and so Eq. (7.29) follows.
We now turn to the case where the second party is honest. In this case, B2 is deter-
mined, and we transform (real-model) A1 into (ideal-model) B1 , which uses A1 as a
subroutine. Recall that B1 gets input α ∈ {0, 1}n .
1. B1 invokes A1 on input α. As (implicit) in the protocol, any action of A1 in Step C1
(including abort) is interpreted as sending a string. Let us denote by v the message
sent by A1 (i.e., v ← A1 (α)).

41 The case in which A2 executes Step C2 with respect to a different common input is just a special case of a
malicious behavior.

666

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

2. Intuitively, machine B1 checks whether or not v = f (α), where α is as in Step 1

(i.e., the input to B1 ). Actually, B1 checks whether or not an honest verifier would
have been convinced by (the residual) A1 that v = f (α) holds, which is equivalent
to being convinced that (h(α), v) ∈ L. Specifically, B1 emulates the execution of
Step C2 (i.e., the execution of the proof system (P, V ) on common input (h(α), v)),
while using the strategy A1 to determine the moves of the (possibly cheating) prover
(and playing the honest verifier in a straightforward manner).42
Recall that this proof system has negligible soundness error, and so if (h(α), v) does
not satisfy Eq. (7.28), this fact is detected with probability 1 − µ(n), where µ is
some negligible function. If the verifier (played by B1 itself) rejects, then machine
B1 aborts (without invoking the trusted party).43 Otherwise, we proceed assuming
that (h(α), v) satisfies Eq. (7.28). Note that since h is 1-1 and Eq. (7.28) is satisfied,
it must be the case that v = f (h −1 (h(α))) = f (α).44
3. Assuming that machine B1 has not aborted, it sends α to the trusted party and allows
the latter to respond to Party 2. (The trusted party’s response will be f (α) = v. Again,
note that, indeed, B1 does not modify its input.)
4. Finally, B1 feeds A1 with the execution view, which consists of the prover’s view of
the emulation of Step C2 (produced in Step 2), and outputs whatever A1 does.
We now show that
c
{ideal F, B (α, h(α))}α∈{0,1}∗ ≡ {real
, A (α, h(α))}α∈{0,1}∗ (7.30)

Actually, we will show that these two ensembles are statistically indistinguishable,
where the statistical difference is due to the case where the real adversary A1 succeeds
in convincing the verifier (played by the honest Party 2) that (u, v) satisfies Eq. (7.28),
and yet this claim is false. By the soundness of the proof system, this event happens only
with negligible probability. On the other hand, in case (u, v) satisfies Eq. (7.28), we show
that ideal F, B (α, h(α)) and real
, A (α, h(α)) are identically distributed. Details follow.
One key observation is that the emulation of the proof system (with prover strategy
A1 (α)) performed in Step 2 by B1 is distributed identically to the real execution of the
proof system that takes place in Step C2 of
.
def def
Fixing any α, recall that v = A1 (α) need not equal f (α), and that u = h(α) uniquely
determines α (because h is 1-1). We denote by p the probability that A1 (α) (playing a
possibly cheating prover) convinces the verifier (played in Step C2 by Party 2) that (u, v)
satisfies Eq. (7.28). (Since A1 is deterministic, v = A1 (α) is fixed and the probability
is only taken over the moves of Party 2.) We consider two cases corresponding to the

42 In particular, if A1 aborts the execution of Step C2, then the honest verifier will not be convinced.
43 Alternatively, machine B1 may invoke the trusted party but prevent it from answering Party 2. The difference
is immaterial, because Party 1 gets nothing from the trusted party. What matters is that (in either case) Party 2
will get an abort symbol (i.e., ⊥).
44 We comment that even if h were not 1-1 but a strong proof-of-knowledge (rather than an ordinary proof system)

was used in Step C2, then one could have inferred that Party 1 knows an α
so that h(α
) = u and v = f (α
),
whereas α
does not necessarily equal α. Sending α
to the trusted party in the next (emulation) step, we would
have been fine, as it would have (also) meant that the trusted party’s response to Party 2 is v.

667
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

relation between p and the soundness error-bound function µ associated with the proof
system (P, V ).45
1. Suppose p > µ(n). In this case, by the soundness condition, it must be the
case that A1 (α) = v = f (α), because in this case (u, v) satisfies Eq. (7.28) and
so v = f (h −1 (u)) = f (h −1 (h(α))) = f (α). Thus, in both the real and the ideal
model, with probability p, the joint execution view is non-aborting and equals
(A1 (α, T ), A1 (α)) = (A1 (α, T ), f (α)), where T represents the prover’s view of the
execution of Step C2 (on common input (h(α), f (α)), where the prover is played by
A1 (α), and the verifier is honest). On the other hand, in both models, with proba-
bility 1 − p, the joint execution is aborting and equals (A1 (α, T ), ⊥), where T is as
before (except that here it is a rejecting execution transcript). Thus, in this case, the
distributions in Eq. (7.30) are identical.
We call the reader’s attention to the reliance of our analysis on the fact that the
emulation of the proof system (with prover A1 (α)) that is performed in Step 2 by B1
is distributed identically to the real execution of the proof system that takes place in
Step C2 of
.

2. Suppose that p ≤ µ(n). Again, in both models, aborting executions are identical
and occur with probability 1 − p. However, in this case, we have no handle on
the non-aborting executions in the real model (because it is no longer guaranteed
that A1 (α) = f (h −1 (u)) holds in the real non-aborting execution, whereas in the
ideal model it still holds that in non-aborting executions, Party 2 outputs f (α) =
f (h −1 (u))). But we do not care, because (in this case) these non-aborting executions
occur with negligible probability (i.e., p ≤ µ(n)). Thus, in this case, the distribution
ensembles in Eq. (7.30) are statistically indistinguishable.
The proposition follows.

We comment that this treatment can be extended to the case that h is a randomized
process, rather than a function (as long as the image of h uniquely determines its pre-
image). Details are omitted in view of the fact that a much more general treatment will
be provided in Section 7.4.3.4.

7.4.3.3. Image Transmission

We now consider the following functionality, called image transmission (or unauthen-
ticated computation):
(α, 1|α| ) → (λ, f (α)) (7.31)
46
where (as in Section 7.4.3.2) the function f is polynomial-time computable. In con-
trast to Section 7.4.3.2, the value f (α) is not verifiable (with respect to a value h(α)

45 We stress that an explicit error-bound can be associated with all standard zero-knowledge proof systems, and
that here we use a system for which µ is negligible. Furthermore, we may use a proof system with error-bound
def
µ(n) = 2−n .
46 Actually, in order to conform with the convention that the functionality has to be defined for any input pair, we
may consider the formulation (α, β) → (λ, f (α)).

668

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

that is known to the second party and uniquely determines α). In other words, the value
output by Party 2 is only required to be an image of f (corresponding to a pre-image of
a given length). Thus, at first glance, one may think that securely computing Eq. (7.31)
should be easier than securely computing Eq. (7.27), especially in case f is onto (in
which case any string is an f -image). This impression is wrong, because securely com-
puting Eq. (7.31) means emulating an ideal model in which Party 1 knows the string
it sends to the trusted party. That is, in a secure protocol for Eq. (7.31), whenever
Party 2 outputs some image (of f ), Party 1 must know a corresponding pre-image
(under f ).47 Still, proving knowledge of a pre-image (and doing so in zero-knowledge)
is what a zero-knowledge proof-of-knowledge is all about. Actually, in order to avoid
expected probabilistic polynomial-time adversaries, we use zero-knowledge strong-
proof-of-knowledge (as defined and constructed in Section 4.7.6). We will show that
Construction 7.4.10 can be easily adapted in order to yield a secure implementation
of Eq. (7.31). Specifically, all that is needed is to use (in Step C2) a zero-knowledge
strong-proof-of-knowledge (rather than an ordinary zero-knowledge proof), and set h
to be a constant function.

Proposition 7.4.12: Suppose that (P, V ) is a zero-knowledge strong-proof-of-

def
knowledge for the relation R = {(v, w) : v = f (w)}, and let h be a constant function.
Then, Construction 7.4.10 securely computes (in the malicious model) the functionality
of Eq. (7.31).

Proof Sketch: Recall that P is postulated to be implemented in probabilistic

polynomial-time when given an adequate auxiliary-input (i.e., a pre-image under f
of the common input). For clarity, we reproduce the modified protocol, omitting all
mention of the (constant) function h.
Inputs: Party 1 gets input α ∈ {0, 1}∗ , and Party 2 gets input 1|α| .
def
Step C1: Party 1 sends v = f (α) to Party 2.
Step C2: Analogously to Construction 7.4.10, the parties invoke the zero-knowledge
strong-proof-of-knowledge (for R) such that Party 1 plays the prover and Party 2
plays the verifier. The common input to the proof system is v, the prover gets α as
auxiliary input, and its objective is to prove that it knows a w such that (v, w) ∈ R
(i.e., v = f (w)). In case the verifier rejects the proof, Party 2 halts with output ⊥
(otherwise the output will be v).
Outputs: In case Party 2 did not output ⊥, it halts with output v. (Party 1 has no output.)

The analysis of this protocol, denoted
, follows the ideas underlying the proof of
Proposition 7.4.11. The only significant modification is in the construction of ideal-
model adversaries for Party 1.

47 We comment that the same also holds with respect to Eq. (7.27). But there, the knowledge of a pre-image (of
the output v under f ) is guaranteed by the fact that security implies that the pre-image of v under f must be
consistent with h(α), whereas the only such pre-image is α itself, which in turn is the initial input of Party 1
and thus known to it.

669
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Let us first justify why the treatment of the case in which Party 1 is honest is
exactly as in the proof of Proposition 7.4.11. In this case, we can use exactly the
same transformation of the real-model adversary A2 into an ideal-model adversary B2 ,
because what this transformation does is essentially invoke the simulator associated
with (the residual prover) A 2 on input the string v = f (α) that it obtains from the
trusted party. Furthermore, the adequateness of this transformation is established by
only referring to the adequateness of the (zero-knowledge) simulator, which holds also
here.
We now turn to the case where the second party is honest. In this case, B2 is deter-
mined, and we transform (real-model) A1 into (ideal-model) B1 , which uses A1 as a
subroutine. Recall that B1 gets input α ∈ {0, 1}n :

1. B1 invokes A1 on input α. As (implicit) in the protocol, any action of A1 in Step C1

(including abort) is interpreted as sending a string. Let us denote by v the message
sent by A1 (i.e., v ← A1 (α)).
2. Machine B1 tries to obtain a pre-image of v under f . Toward this end, B1 uses the
(strong) knowledge-extractor associated with the proof system of Step C2. Specif-
ically, providing the strong knowledge-extractor with oracle access to (the residual
prover) A1 (α), machine B1 tries to extract (from A1 ) a string w such that f (w) = v.
def def
In case the extractor succeeds, B1 sets α
= w. Otherwise, B1 sets α
= ⊥.
3. Machine B1 now emulates an execution of Step C2. Specifically, it lets A 1 (α) play
the prover and emulates by itself the (honest) verifier interacting with A1 (α) (i.e., B1
behaves like A2 ).
r In case the emulated verifier rejects, machine B1 aborts (without invoking the
trusted party), and outputs whatever A1 does (when fed with this emulated proof
transcript).
r Otherwise (i.e., the emulated verifier accepts), if α
= ⊥, then B1 sends α
to the
trusted party and allows it to respond to Party 2. (The response will be f (α
),
which by Step 2 must equal v.) In case α
= ⊥, this sub-step will fail, and B1
aborts as in the case that the emulated verifier rejects.
4. Finally, B1 feeds A1 with the execution view, which consists of the prover’s view of
the emulation of Step C2 (produced in Step 3), and outputs whatever A1 does.
Denoting the functionality of Eq. (7.31) by F, we now show that
c
{ideal F, B (α, 1|α| )}α∈{0,1}∗ ≡ {real
, A (α, 1|α| )}α∈{0,1}∗ (7.32)

Actually, we will show that these two ensembles are statistically indistinguishable, where
the statistical difference is due to the case where the real-model adversary A1 succeeds
in convincing the knowledge-verifier (played by the honest A2 ) that it knows a pre-
image of v under f , and yet the knowledge-extractor failed to find such a preimage. By
definition of strong knowledge-verifiers, such an event may occur only with negligible
probability. Loosely speaking, ignoring the rare case in which extraction fails although
the knowledge-verifier (played by A2 ) is convinced, it can be shown that the distributions
ideal f, B ((σ, r ), 1n ) and real
, A ((σ, r ), 1n ) are identical. Details follow.
670

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

def
Fixing any α, recall that v = A1 (α) need not be an image of f (let alone that it may
not equal f (α)). We denote by p the probability that A1 (α), playing a possibly cheating
prover, convinces the knowledge-verifier (played in Step C2 by Party 2) that it knows a
pre-image of v under f . We consider two cases corresponding to the relation between
p and the error-bound function µ referred to in Definition 4.7.13:

1. Suppose that p > µ(n). In this case, by Definition 4.7.13, with probability at least
1 − µ(n), machine B1 has successfully extracted a pre-image α
(of v = A1 (α) under
f ). In the real model, with probability p, the joint execution ends up non-aborting.
By the aforementioned extraction property, in the ideal model, a joint execution is
non-aborting with probability p ± µ(n) (actually, the probability is at least p − µ(n)
and at most p). Thus, in both models, with probability p ± µ(n), a joint execution is
non-aborting and equals ( A1 (α, T ), A1 (α)) = (A1 (α, T ), f (α
)), where T represents
the prover’s view of an execution of Step C2 (on common input f (α
) = v = A1 (α),
where the role of the prover is played by the residual strategy A1 (α) and the verifier is
honest). On the other hand, in both models, with probability 1 − p ± µ(n), the joint
execution is aborting and equals (A1 (α, T ), ⊥), where T is as before (except that
here it is a rejecting execution transcript). Thus, the statistical difference between the
two models is due only to the difference in the probability of producing an aborting
execution in the two models, which in turn is negligible.
We call the reader’s attention to the reliance of our analysis on the fact that the
emulation of the proof system (with prover A1 (α)) performed in Step 2 by B1 is
distributed identically to the real execution of the proof system that takes place in
Step C2 of
.

2. Suppose that p ≤ µ(n). Again, in the real model the non-aborting probability is p,
which in this case is negligible. Thus, we ignore these executions and focus on the
aborting executions, which occur with probability at least 1 − p ≥ 1 − µ(n) in both
models. Recalling that aborting executions are identically distributed in both models,
we conclude that the statistical difference between the two models is at most µ(n).
Thus, in both case, the distribution ensembles in Eq. (7.32) are statistically indistin-
guishable. The proposition follows.

7.4.3.4. Authenticated Computation, Revisited

We now generalize the image-transmission functionality to treat the case where Party 2
has some partial information of the input of Party 1. In the extreme case, the informa-
tion available to Party 2 uniquely determines the input of Party 1 (although obtaining
the latter from the former may be infeasible). Thus, in a sense, we revisit the authenti-
cated computation functionality, which was considered in Section 7.4.3.2. The impor-
tant aspect of the current treatment is that we consider a functionality that is defined
on all pairs of (equal-length) strings, rather than a partial functionality (as treated in
Section 7.4.3.2).
671
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Definition 7.4.13 (authenticated computation, revisited): Let f : {0, 1}∗ × {0, 1}∗ →
{0, 1}∗ and h : {0, 1}∗ → {0, 1}∗ be polynomial-time computable. The h-authen-
ticated f -computation functionality is redefined by
(λ , f (α)) if β = h(α)
(α, β) → (7.33)
(λ , (h(α), f (α))) otherwise

In the intended applications of the h-authenticated f -computation functionality, Party 2

is supposed to input β = h(α), and so the first case in Eq. (7.33) holds, provided
that both parties are honest. Indeed, if Party 2 is honest, then either it gets the cor-
rect value of f (α) (i.e., which fits h(α) known to it) or it gets an indication that
Party 1 is cheating. The specific form of the second case was designed to facil-
itate the implementation, while not causing any harm.48 What matters is that the
outputs in the two cases are different, and so Party 2 can tell whether or not it re-
ceived the correct value of f (α). We stress that in the intended applications, Party 2
knows h(α) and is supposed to obtain f (α), and so it causes no harm to provide
Party 2 with both of them (even in case Party 2 misbehaves and enters an input other
than h(α)).
We assume again, for simplicity, that h is length preserving, which again can be
“enforced” by considering α
= (α, 1|h(α)| ) and h
(α
) = (h(α), 1|α| ). However, we make
no further assumptions concerning the function h, and thus Eq. (7.31) is essentially a
special case (obtained by setting h(α) = 1|α| ).
The functionality of Eq. (7.33) is implemented by having Party 1 use the image-
transmission functionality to send the pair (h(α), f (α)) to Party 2, which compares
the first element to its own input and acts accordingly. That is, we use the following
(oracle-aided) protocol:

Construction 7.4.14 (authenticated computation protocol, general version):

Inputs: Party 1 gets input α ∈ {0, 1}∗ , and Party 2 gets input β ∈ {0, 1}|α| .
def
Step C1: Party 1 uses the image-transmission functionality to send the pair (u, v) =
(h(α), f (α)) to Party 2. That is, the parties invoke the functionality of Eq. (7.31) with
def
respect to the function g(α) = (h(α), f (α)), where Party 1 enters the input α and
Party 2 is to obtain g(α).
Step C2: Assuming that Step C1 was not aborted by Party 1 and that Party 2 receives
the pair (u, v) in Step C1, Party 2 outputs v if u = β and (u, v) otherwise.
Outputs: If not aborted (with output ⊥), Party 2 sets its local output as directed in
Step C2. (Party 1 has no output.)
We stress that in the oracle invocation (of Step C1), Party i plays the i-th party (with
respect to the oracle call). Recall that (unlike Party 2), Party 1 may abort and in particular

48 In contrast, even privately computing the more natural functionality (α, β) → (λ , v), where v = f (α) if β =
h(α) and v = λ otherwise, is significantly harder than (securely or privately) implementing Eq. (7.33); see
Exercise 12. The difference is that Eq. (7.33) allows for revealing h(α) to Party 2 (specifically in case h(α) = β),
whereas the more natural functionality does not allow this.

672

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

do so during Step C1. Since Step C1 consists of an oracle invocation, aborting during
Step C1 means instructing the oracle not to answer Party 2.

Proposition 7.4.15: Construction 7.4.14 securely reduces the h-authenticated f -

computation functionality of Eq. (7.33) to the image-transmission functionality of
Eq. (7.31).

Proof Sketch: We need to transform any admissible pair, (A 1 , A2 ), for the real oracle-
aided model into a corresponding pair, (B1 , B2 ), for the ideal model. We start by assum-
ing that the first party is honest and by transforming the real-model adversary A 2 (for
the oracle-aided execution) into a corresponding ideal-model adversary B2 . On input
β, the latter proceeds as follows:

def
1. Machine B2 sends β to the trusted party and obtains the answer, which equals v =
def
f (α) if β = h(α) and (u, v) = (h(α), f (α)) otherwise, where α is the (unknown
def
to B2 ) input of Party 1.49 In the first case, B2 sets u = β, and so in both cases
(u, v) = (h(α), f (α)).
2. Machine B2 emulates the protocol, by feeding A2 with β and the pair (u, v),
which A2 expects to get in Step C1, and outputting whatever the latter outputs (in
Step C2).
Note that both the ideal execution under (B1 , B2 ) and the real execution (in the oracle-
aided model) under ( A1 , A2 ) yield the output pair (λ , A2 (β, (h(α), f (α)))). Thus, the
ideal and real ensembles are identical.
We now turn to the case where the second party is honest and transform the real-
model adversary A1 into a corresponding ideal-model adversary B1 . On input α,
the latter proceeds as follows:

1. Machine B1 emulates Step C1 of the protocol, by obtaining from A1 the input α
←

A1 (α) (that A1 wishes to transmit via Eq. (7.31)) and feeding A1 with the expected
answer λ.
2. If A1 instructs the oracle not to answer Party 2, then B1 halts without invoking the
trusted party. Otherwise, B1 sends α
to the trusted party and lets it answer Party 2.
In both cases, B1 halts with output equal to the corresponding output of A1 .
Note that if h(α
) = β, where β is the (unknown to B1 ) input of Party 2, then the
trusted party answers Party 2 with f (α
), and otherwise it answers Party 2 with
(h(α
), f (α
)).
Note that both the ideal execution under (B1 , B2 ) and the real execution (in the oracle-
aided model) under ( A1 , A2 ) yield the output pair ( A1 (α, λ, ⊥) , ⊥) if A1 (α, λ) = ⊥
and (A1 (α, λ) , F(A1 (α), β) otherwise, where F(α
, β) is as in Eq. (7.33); that is,

49 Recall that, in either case, the trusted party will send Party 1 the answer λ. Also note that the emulation will
remain valid regardless of which |β|-bit long string B2 sends to the trusted party (because, for any such choice,
B2 will [explicitly] receive f (α), as well as [explicitly or implicitly] receive h(α)).

673
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

F(α
, β) = f (α
) if h(α
) = β and F(α
, β) = (h(α
), f (α
)) otherwise. Thus, also here

the ideal and real ensembles are identical.

7.4.3.5. Augmented Coin-Tossing

In this section, we generalize the coin-tossing functionality (of Section 7.4.3.1) in two
ways. Firstly, we consider the generation of a random (n)-bit long string, rather than a
single bit. Secondly, we provide the second party with a function of the coin-outcomes
obtained by the first party, rather than providing it with the outcomes themselves. That
is, for any positive polynomial  : N → N and a polynomial-time computable function
g, we consider the randomized functionality
(1n , 1n ) → (r, g(r )), where r is uniformly distributed in {0, 1}(n) . (7.34)
def def
Indeed, Definition 7.4.6 is a special case (obtained by setting (n) = 1 and g(r ) = r ).
The augmented coin-tossing functionality (mentioned in Section 7.4.1) will be derived
as a special case (see Proposition 7.4.19). But first we show that Eq. (7.34) can be
securely reduced to the set of functionalities presented earlier (see discussion of this
notion of a reduction in Remark 7.4.5). That is, we present an oracle-aided protocol that
uses two of the latter functionalities (i.e., basic coin-tossing and general authenticated
computation), as well as a commitment scheme C. The protocol can be viewed as
a “robust” version of Construction 7.4.7 (i.e., simple operations, such as sending a
commitment to a value and tossing a coin, are replaced by corresponding functionalities
that prevent various abuses).

Construction 7.4.16 (an oracle-aided protocol for Eq. (7.34)): For r1 , ..., r ∈ {0, 1}n
and σ1 , ..., σ ∈ {0, 1}, we let C r1 ,...,r (σ1 , ..., σ ) = (Cr1 (σ1 ), ..., Cr (σ )).
def
Inputs: Both parties get security parameter 1n , and set  = (n).
Step C1: Party 1 uniformly selects σ1 , ..., σ ∈ {0, 1} and s1 , ..., s ∈ {0, 1}n , and lets
r
= σ1 · · · σ and s = s1 · · · s .
def
Step C2: Party 1 uses the image-transmission functionality to send c = C s (r
) to
Party 2. Actually, since image-transmission functionality is a special case of the
general authenticated-computation functionality, we use the latter. That is, Party 1
enters Eq. (7.33) with input (r
, s), Party 2 enters with input 1+·n , and Party 2 is
def
supposed to obtain f (C2) (r
, s) = C s (r
).
Recall that, by definition, a party cannot abort the execution of an oracle call that
was not initiated (requested) by it, and so Party 2 cannot abort Steps C2–C4. For
simplicity, we assume that Party 1 does not abort Steps C2 and C3, but it may abort
Step C4.
Step C3: The parties invoke the basic coin-tossing functionality  times to generate a
common random string r

∈ {0, 1} . That is, in the i-th invocation of the functionality
of Definition 7.4.6, the parties obtain the i-th bit of r

.
def
Step C4: Party 1 sets r = r
⊕ r

, and uses the authenticated-computation functional-
ity to send g(r ) to Party 2. Specifically, Party 1 enters Eq. (7.33) with input (r
, s, r

),
674

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

def
Party 2 enters with input (c, r

), where (c, r

) is supposed to equal h (C4) (r
, s, r

) =
def
(C s (r
), r

), and Party 2 is supposed to obtain f (C4) (r
, s, r

) = g(r
⊕ r

). In case
Party 1 aborts or Party 2 obtains an answer of a different format, which happens if
the inputs to the functionality do not match, Party 2 halts with output ⊥ (indicating
that Party 1 misbehaved).
We comment that r = r
⊕ r

is uniquely determined by c and r

.
Outputs: Party 1 outputs r , and Party 2 outputs the value determined in Step C4, which
is either g(r ) or ⊥.
We stress that, in all oracle calls, Party 1 is the party initiating (requesting) the call. We
comment that more efficient alternatives to Construction 7.4.16 do exist; it is just that
we find Construction 7.4.16 easiest to analyze.

Proposition 7.4.17: Let F be the set of functionalities defined in Definition 7.4.6 and
Eq. (7.33), respectively. Then Construction 7.4.16 constitutes a security reduction from
the generalized coin-tossing functionality of Eq. (7.34) to F.

Proof Sketch: We start by assuming that the first party is honest and by transforming
the real-model adversary A2 (for the oracle-aided execution) into a corresponding ideal-
model adversary B2 . On input 1n , the latter proceeds as follows:
1. Machine B2 emulates the local actions of the honest Party 1 in Step C1 of the protocol,
by uniformly selecting r
∈ {0, 1} and s ∈ {0, 1}·n .
def
2. Machine B2 emulates Step C2 of the protocol, by feeding A2 with c = C s (r
). (Recall
that by our convention, A2 never aborts.)
3. Machine B2 emulates Step C3 of the protocol, by uniformly selecting r

∈ {0, 1}
and feeding A2 with it.
4. Machine B2 invokes the trusted party with input 1n and obtains the answer g(r ),
for a uniformly distributed r ∈ {0, 1} that is handed to Party 1.50 Next, machine
B2 obtains the input (or query) of A2 to the functionality of Step C4. If this input
(i.e., A2 (λ, C s (r
), r

), where λ represents the Step 1 emulation of Step C1) does not
equal the pair of values (C s (r
), r

) fed to A2 in Steps 2–3, then B2 halts with output
A2 (λ, c, r

, ((c, r

), g(r ))). Otherwise, B2 halts with output A2 (λ, c, r

, g(r )).
Note that in both cases, the output of B2 corresponds to the output of A2 when
fed with the corresponding emulation of Steps C1–C4. In particular, B2 emulates
Step C4 by feeding A2 either with g(r ) or with (h (C4) (r
, s, r

), g(r )), where the
decision depends on whether or not A2 (λ, C s (r
), r

) = (C s (r
), r

). (Recall that
(C s (r
), r

) = h (C4) (r
, s, r

).) Indeed, B2 is cheating (in the emulation of Step C4),
because A2 expects to get either f (C4) (r
, s, r

) = g(r
⊕ r

) or (h (C4) (r
, s, r

), g(r
⊕
r

)), but (as we shall see) this cheating is undetectable.
Let us first assume that the input entered by A2 to the functionality of Step C4
does fit its view of Steps C2 and C3, an event that occurs with equal probability

50 Indeed, this part of the current step could also take place at an earlier stage.

675
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

in both models (because the emulation of Steps C2–C3 is perfect). In this case,
the ideal-model execution under (B1 , B2 ) yields the pair (r , A2 (λ, C(r
), r

, g(r ))),
where r
, r

, r are uniformly and independently distributed. On the other hand, the
real-model execution (in the oracle-aided model) under (A1 , A2 ) yields the pair
(r
⊕ r

, A2 (λ, C(r
), r

, g(r
⊕ r

))), where r
and r

are as before, which (for
r = r
⊕ r

) is distributed identically to (r , A2 (λ, C(r ⊕ r

), r

, g(r ))). However, due
to the hiding property of C, the two ensembles are computationally indistinguish-
able. In case the input entered by A2 to the functionality of Step C4 does not fit
its view of Steps C2 and C3, the ideal-model execution under (B1 , B2 ) yields the
pair (r , A2 (λ, C(r
), r

, ((C(r
), r

), g(r )))), whereas the real-model execution under
(A1 , A2 ) yields the pair (r
⊕ r

, A2 (λ, C(r
), r

, ((C(r
), r

), g(r
⊕ r

)))), which is
distributed identically to (r , A2 (λ, C(r ⊕ r

), r

, ((C(r ⊕ r

), r

), g(r )))). Again, the
two ensembles are computationally indistinguishable.
We now turn to the case where the second party is honest and transform the real-
model adversary A1 into a corresponding ideal-model adversary B1 . On input 1n , the
latter proceeds as follows:

1. Machine B1 emulates Step C1 of the protocol, by obtaining (r
, s) ← A1 (1n ), which

is the query that A1 will use in Step C2.
2. Machine B1 emulates Step C2 by doing nothing.
Note that the real-model adversary A1 would have made the oracle query (r
, s) and
would have obtained λ as an answer.
3. Machine B1 invokes the trusted party (on input 1n ) and obtains a uniformly distributed
r ∈ {0, 1} . We stress that at this time, B1 does not instruct the trusted party whether
def
or not to answer Party 2. Machine B1 emulates Step C3, by feeding r

= r ⊕ r
to
A1 .
4. Machine B1 starts its emulation of Step C4, by checking whether or not the query
that A1 wishes to make (i.e., A1 (1n , λ, r

)) fits the tuple (r
, s, r

) in the sense
def
that it yields the same value (C s (r
), r

). That is, let (q
, q, q

) = A1 (1n , λ, r

). If
(C q (q
), q

) = (C s (r
), r

), then B1 instructs the trusted party to answer Party 2,
or else B1 instructs the trusted party to stop (without answering Party 2).51 Finally,
B1 outputs whatever A1 does (i.e., A1 (1n , λ, r

, λ), where the four inputs of A 1
correspond to its view in each of the four steps).

Note that the output of Party 1 in both the real model (under the Ai ’s) and the ideal model
(under the Bi ’s) equals A1 (1n , λ, r

, λ), where r

is uniformly distributed (in both mod-
els). The issue is the correlation of this output to the output of Party 2, which is relevant
only if Party 2 does have an output. Recall that Party 2 obtains an output (in both models)
only if the corresponding Party 1 does not abort (or stops the trusted party). Furthermore,
in both models, an output is obtained if and only if (C q (q
), q

) = (C s (r
), r

) holds,
def def
where (r
, s) = A1 (1n ), and (q
, q, q

) = A1 (1n , λ, r

). In particular, (C q (q
), q

) =

51 In particular, if (in contrary to our simplifying assumption) A1 aborts before Step C4, then the sequence
(q
, q, q

) equals ⊥ and does not fit (C s (r
), r

).

676

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

(C s (r
), r

) implies that (q
, q

) = (r
, r

) and that the inputs entered in Step C4 do

match (i.e., h (C4) (q
, q, q

) = (C s (r
), r

)). This means that in the real model, the output
of Party 2 is f (C4) (q
, q, q

) = f (C4) (r
, q, r

) = g(r
⊕ r

), whereas in the ideal model,
it equals g(r ) = g(r
⊕ r

). We conclude that the ideal model perfectly emulates the
real model, and the proposition follows.

An Important Special Case. An important special case of Eq. (7.34) is when g(r, s) =
C s (r ), where |s| = n · |r |. This special case will be called the augmented coin-tossing
functionality.

Definition 7.4.18 (coin-tossing into the well, augmented): An augmented coin-

tossing-into-the-well protocol is a two-party protocol for securely computing the
following randomized functionality with respect to some fixed commitment scheme,
C, and a positive polynomial :

(1n , 1n ) → ((r, s), C s (r )) (7.35)

where (r, s) is uniformly distributed in {0, 1}(n) × {0, 1}(n)·n .

An augmented coin-tossing protocol is exactly what is needed for the implementation
of the coin-generation phase of the compiler. In particular, the string s, included in the
output of Party 1, allows it to (later) prove in zero-knowledge statements regarding the
actual value, r , committed (to Party 2). This fact will be used in the protocol emulation
phase of the compiler.

Proposition 7.4.19: Let F be as in Proposition 7.4.17, and suppose that C is a commit-

ment scheme. Then Construction 7.4.16, when applied to g = C, constitutes a secure
reduction of the augmented coin-tossing functionality Eq. (7.35) to the set of function-
alities F.

7.4.3.6. Input Commitment

The last component needed for the compiler is a functionality that captures what is
required in the input-commitment phase of the compiler. Specifically, we want to force
Party 1 to make a random commitment to an input of its choice, while knowing the com-
mitted value and the corresponding decommitment. Knowledge of the latter will allow
the party to (later) prove in zero-knowledge statements regarding the actual committed
value, and this fact will be used in the protocol-emulation phase of the compiler.
Let C be a commitment scheme, and let C be defined as in Section 7.4.3.5. We
consider the input commitment functionality

(x, 1|x| ) → (r, C r (x)), where r is uniformly distributed in {0, 1}|x|

2
(7.36)

Certainly, the naive protocol of just letting Party 1 send Party 2 a commitment to x does
not constitute a secure implementation of Eq. (7.36): This naive suggestion does not
guarantee that the output is in the range of the commitment scheme, let alone that it is a
random commitment for which Party 1 knows a corresponding decommitment. Thus,

677
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

the naive protocol must be augmented by mechanisms that address all these concerns.
We show that Eq. (7.36) can be securely reduced to the set of functionalities presented
in previous subsections.

Construction 7.4.20 (an oracle-aided protocol for Eq. (7.36)):

Inputs: Party 1 has input x ∈ {0, 1}n , whereas Party 2 gets input 1n .
Step C1: Party 1 selects uniformly r
∈ {0, 1}n .
2

def
Step C2: Party 1 uses the image-transmission functionality to send c
= C r
(x) to
Party 2. Again, we actually use the authenticated-computation functionality, where
Party 1 enters Eq. (7.33) with input (x, r
), Party 2 inputs 1n+n , and Party 2 is sup-
2

def
posed to obtain f (C2) (x, r
) = C r
(x). Thus, Steps C1–C2 yield an initial commitment
to the input.
As in Construction 7.4.16, we recall that Party 2 cannot abort Steps C2–C4, and
assume that Party 1 does not abort Steps C2 and C3.
Step C3: Generating coins for the final commitment. The parties use the augmented
def
coin-tossing functionality to obtain the outputs (r, r

) and c

= C r

(r ), respectively,
n2

n3
where r ∈ {0, 1} and r ∈ {0, 1} are uniformly and independently distributed.
That is, Party 1 gets (r, r

), while Party 2 gets c

.
Step C4: Sending the final commitment. Party 1 uses the authenticated-computation
functionality to send C r (x) to Party 2, where (x, r ) is uniquely determined by (c
, c

).
Specifically, Party 1 enters Eq. (7.33) with input (x, r, r
, r

), Party 2 enters with input
def
(c
, c

), where (c
, c

) is supposed to equal h (C4) (x, r, r
, r

) = (C r
(x), C r

(r )), and
def
Party 2 is supposed to obtain f (C4) (x, r, r
, r

) = C r (x).
In case Party 1 aborts or Party 2 obtains an answer of a different format, which
happens if the inputs to the functionality do not match, Party 2 halts with output ⊥
(indicating that Party 1 misbehaved).
Outputs: Party 1 outputs r , and Party 2 outputs the value determined in Step C4, which
is either C r (x) or ⊥.
Again, more efficient alternatives to Construction 7.4.20 do exist, but we prefer to
analyze the one here.

Proposition 7.4.21: Construction 7.4.20 constitutes a security reduction from

Eq. (7.36) to the set of two functionalities defined in Eq. (7.35) and Eq. (7.33),
respectively.

Proof Sketch: We start by assuming that the first party is honest and by transforming
the real-model adversary A2 (for the oracle-aided execution) into a corresponding ideal-
model adversary B2 . On input 1n , the latter proceeds as follows:

1. Machine B2 emulates (the actions of the honest Party 1 in) Step C1 of the protocol,
by uniformly selecting r
∈ {0, 1}n .
2

678

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

def
2. Machine B2 emulates Step C2 of the protocol, by feeding A2 with c
= C r
(0n ).
(Clearly, B2 is cheating, because A2 is supposed to be fed with C(x), where x is the
(unknown to B2 ) input of Party 1. However, A2 cannot detect this cheating.)
2
3. Machine B2 emulates Step C3 of the protocol, by uniformly selecting s ∈ {0, 1}n
def
and r

∈ {0, 1}n , and feeding A2 with c

= C r

(s).
3

4. Machine B2 invokes the trusted party with input 1n and obtains the answer C r (x),
2
for a uniformly distributed r ∈ {0, 1}n that is handed to Party 1.52 Next, machine B2
obtains the input (or query) of A2 to the functionality of Step C4. If this input (i.e.,
A2 (λ, c
, c

)) does not equal the pair of values (c
, c

) = (C r
(0n ), C r

(s)) fed to A2
in Steps 2–3, then B2 halts with output A2 (λ, c
, c

, ((c
, c

), C r (x))). Otherwise, B2
halts with output A2 (λ, c
, c

, C r (x)).

Note that in both cases, the output of B2 corresponds to the output of A2 when
fed with the corresponding emulation of Steps C1–C4. In particular, B2 emulates
Step C4 by feeding A2 with either C r (x) or with ((C(0n ), C(s)), C r (x)), where the
decision depends on whether or not A2 (λ, C r
(0n ), C r

(s)) = (C r
(0n ), C r

(s)). (Re-
call that (C r
(0n ), C r

(s)) = h (C4) (0n , s, r
, r

).) Indeed, on top of cheating in the
emulation of Step C2, machine B2 cheats in the emulation of Step C4, firstly be-
cause the decision is supposed to depend on whether or not A2 (λ, C r
(x), C r

(r )) =
(C r
(x), C r

(r )), where (C r
(x), C r

(r )) = h (C4) (x, r, r
, r

), and secondly because
A2 expects to get either C r (x) = f (C4) (x, r, r
, r

) or ((C(x), C(r )), C r (x)) ≡
(h (C4) (x, r, r
, r

), f (C4) (x, r, r
, r

)). However, as we shall see, this cheating is
undetectable.
Let us first assume that the input entered by A2 to the functionality of Step C4 does
fit its view of Steps C2 and C3. In this case, the ideal-model execution under (B1 , B2 )
yields the pair (r , A2 (λ, C(0n ), C(s), C r (x))), where r and s are uniformly and inde-
pendently distributed. On the other hand, the corresponding real-model execution (in
the oracle-aided model) under (A1 , A2 ) yields the pair (r , A2 (λ, C(x), C(r ), C r (x))),
where r is as before. However, due to the hiding property of C, the two ensembles
are computationally indistinguishable.53 In case the input entered by A2 to the func-
tionality of Step C4 does not fit its view of Steps C2 and C3, the ideal-model exe-
cution under (B1 , B2 ) yields the pair (r , A2 (λ, C(0n ), C(s), ((C(0n ), C(s)), C r (x)))),
whereas the corresponding real-model execution under (A1 , A2 ) yields the pair
(r , A2 (λ, C(x), C(r ), ((C(x), C(r )), C r (x)))). Again, the two ensembles are com-
putationally indistinguishable. Since the two cases occur with almost the same
probability in both models (because the decision depends on A2 (λ, c
, c

), where
(c
, c

) is either (C(0n ), C(s)) or (C(x), C(r ))), the outputs in the two models are
indistinguishable.

52 Indeed, this part of the current step could also take place at an earlier stage.
53 In fact, the said ensembles are computationally indistinguishable even when r and s are fixed, rather than being
random. That is, the ensembles {(C(0|x| ), C(s), C r (x))}x,r,s and {(C(x), C(r ), C r (x))}x,r,s are computationally
indistinguishable, where (as usual) the distribution’s index (x, r, s) is also given to the potential distinguisher.
This follows from the computational indistinguishability of {(C(0|x| ), C(s))}x,r,s and {(C(x), C(r ))} x,r,s , which
in turn follows from the hiding property of C.

679
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

We now turn to the case where the second party is honest and transform the real-
model adversary A1 into a corresponding ideal-model adversary B1 . On input x, the
latter proceeds as follows:

1. Machine B1 emulates Step C1 of the protocol, by obtaining r
from A1 (x). Actually,

B1 obtains (x
, r
) ← A1 (x), which is the query that A1 will use in Step C2.
2. Machine B1 emulates Step C2 by doing nothing.
Note that the real-model adversary A1 would have made the oracle query (x
, r
) and
would have obtained λ as an answer.
3. Machine B1 invokes the trusted party on input x
, and obtains a uniformly distributed
2
r ∈ {0, 1}n . We stress that at this time, B1 does not instruct the trusted party whether
or not to answer Party 2. Machine B1 emulates Step C3, by uniformly selecting
r

∈ {0, 1}n and feeding (r, r

) to A1 .
3

4. Machine B1 starts its emulation of Step C4, by checking whether or not the query that
A1 wishes to make (i.e., A1 (x, λ, (r, r

))) fits the tuple (x
, r, r
, r

) in the sense that it
def
yields the same value (C r
(x
), C r

(r )). That is, let (q1 , q2 , s1 , s2 ) = A1 (x, λ, (r, r

)).
If (C s1 (q1 ), C s2 (q2 )) = (C r
(x
), C r

(r )), then B1 instructs the trusted party to answer
Party 2; otherwise B1 instructs the trusted party to stop (without answering Party 2).
Finally, B1 outputs whatever A1 does (i.e., A1 (x, λ, (r, r

), λ), where the four inputs
of A1 correspond to its view in each of the four steps).

Note that the output of Party 1 in both the real model (under the Ai ’s) and the ideal model
(under the Bi ’s) equals A1 (x, λ, (r, r

), λ), where r ∈ {0, 1}n and r

∈ {0, 1}n are uni-
2 3

formly and independently distributed (in both models). The issue is the correlation of
this output to the output of Party 2, which is relevant only if Party 2 does have an output.
Recall that Party 2 obtains an output (in both models) only if the corresponding Party 1
does not abort (or stops the trusted party). Furthermore, in both models, an output is ob-
tained if and only if (C s1 (q1 ), C s2 (q2 )) = (C r
(x
), C r

(r )), where (x
, r
) = A1 (x) and
(q1 , q2 , s1 , s2 ) = A1 (x, λ, (r, r

)). In particular, (C s1 (q1 ), C s2 (q2 )) = (C r
(x
), C r

(r ))
implies that (q1 , q2 ) = (x
, r ) and that the inputs entered in Step C4 do match (i.e.,
h (C4) (q1 , q2 , s1 , s2 ) = (C r
(x
), C r

(r ))), which means that in the real model, the output
of Party 2 is f (C4) (q1 , q2 , s1 , s2 ) = f (C4) (x
, r, s1 , s2 ) = C r (x
) (exactly as in the ideal
model). We conclude that the ideal model perfectly emulates the real model, and the
proposition follows.

7.4.3.7. Summary
Combining Proposition 7.4.8 (resp., Proposition 7.4.12) with suitable results regarding
the underlying primitives, we conclude that coin-tossing (resp., image transmission
as in Eq. (7.31)) can be securely implemented based on any 1-1 one-way function.
Combining Proposition 7.4.15 (resp., Proposition 7.4.19) [resp., Proposition 7.4.21]
with the previous results, by using the Composition Theorem (i.e., Theorem 7.4.3 or
Remark 7.4.5), we obtain secure implementations of the authenticated-computation
functionality (resp., augmented coin-tossing) [resp., input-commitment functionality].
The 1-1 restriction can be waived by using a slightly more cumbersome construction that
680

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

utilizes the commitment scheme of Construction 4.4.4 (instead of the simple scheme
of Construction 4.4.2). We thus state the following for future reference:

Proposition 7.4.22: Assuming the existence of (non-uniformly strong) one-way func-

tions, the following three functionalities can be securely computed:

1. The input-commitment functionality as defined in Eq. (7.36).

2. The augmented coin-tossing functionality as defined in Eq. (7.35).
3. The authenticated-computation functionality as defined in Eq. (7.33).

7.4.4. The Compiler Itself

We are now ready to present the compiler. Recall that we are given a protocol,
, for the
semi-honest model, and we want to generate an equivalent protocol

for the malicious
model. The meaning of the term “equivalent” will be clarified in Section 7.4.4.1. We
start by compiling
into an oracle-aided protocol

that uses the three functionalities
referred to in Proposition 7.4.22.
We assume, without loss of generality, that on any input of length n, each party to

tosses (n) = poly(n) coins. Recall that C is a (non-interactive) (string) commit-
ment scheme, derived from the bit commitment scheme C, and that C r (v) denotes the
commitment to value v using the random-tape r .

Construction 7.4.23 (the compiled protocol, oracle-aided version): Given a protocol,

, for the semi-honest model, we consider the following oracle-aided protocol,

, for
the malicious model:

Inputs: Party 1 gets input x ∈ {0, 1}n and Party 2 gets input y ∈ {0, 1}n .
Input-Commitment Phase: Each of the two parties commits to its input by using the
input-commitment functionality of Eq. (7.36). Recall that Eq. (7.36) maps the input
2
pair (u, 1n ) to the output pair (s, C s (u)), where s is uniformly distributed in {0, 1}n .
Thus, each of the parties obtains decommitment information that will allow it to
perform its role in the protocol-emulation phase.

Specifically, we are talking about two invocations of Eq. (7.36). In the first invocation,
Party 1 wishing to commit to x, plays the role of the first party in Eq. (7.36), and
2
obtains a uniformly distributed ρ 1 ∈ {0, 1}n , whereas Party 2 (which plays the role
def
of the second party in Eq. (7.36)) obtains γ 1 = C ρ 1 (x). Likewise, in the second
invocation, Party 2, wishing to commit to y, plays the role of the first party in
2
Eq. (7.36), and obtains a uniformly distributed ρ 2 ∈ {0, 1}n , whereas Party 1 (which
def
plays the role of the second party in Eq. (7.36)) obtains γ 2 = C ρ 2 (y).
Coin-Generation Phase: Each of the parties generates a random-tape for the emula-
tion of
, by invoking the augmented coin-tossing functionality of Eq. (7.35). Recall
that this functionality maps the input pair (1n , 1n ) to the output pair ((r, s), C s (r )),
where (r, s) is uniformly distributed in {0, 1}(n) × {0, 1}n·(n) . Thus, each party
681
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

obtains the random-tape to be held by it, whereas the other party obtains a commit-
ment to this value. The party holding the random-tape also obtains the randomization
used in the corresponding commitment, which it will use in performing its role in the
protocol-emulation phase.
Specifically, we are talking about two invocations of Eq. (7.35). In the first (resp.,
second) invocation, Party 1 (resp., Party 2) plays the role of the first party in
Eq. (7.35), and obtains a uniformly distributed (r 1 , ω1 ) ∈ {0, 1}(n) × {0, 1}n·(n)
(resp., (r 2 , ω2 ) ∈ {0, 1}(n) × {0, 1}n·(n) ), whereas Party 2 (resp., Party 1) which plays
def def
the other role, obtains δ 1 = C ω1 (r 1 ) (resp., δ 2 = C ω2 (r 2 )).
Protocol-Emulation Phase: The parties use the authenticated-computation function-
ality of Eq. (7.33) in order to emulate each step of protocol
. Recall that, for
predetermined functions h and f , this functionality maps the input pair (α, β) to
the output pair (λ, f (α)) if β = h(α) and to (λ , (h(α), f (α))) otherwise, where the
second case is treated as abort.
The party that is supposed to send a message plays the role of the first (i.e., initiating)
party in Eq. (7.33), and the party that is supposed to receive the message plays the
role of the second party. Suppose that the current message in
is to be sent by
def def
Party j, and let u = x if j = 1 and u = y otherwise. Then the functions h, f and
the inputs α, β, for the functionality of Eq. (7.33), are set as follows:
r The string α is set to equal (α1 , α2 , α3 ), where α1 = (u, ρ j ) is the query and
answer of Party j in the oracle call that it initiated in the input-commitment
phase, α2 = (r j , ω j ) is the answer that Party j obtained in the oracle call that
it initiated in the coin-generation phase, and α3 is the sequence of messages that
Party j obtained so far in the emulation of
. The string β equals (γ j , δ j , α3 ),
where γ j and δ j are the answers that the other party obtained in the same oracle
calls in the first two phases (and α3 is as before).
In particular, u is the input to which Party j committed in the input-commitment
phase, and r j is the random-tape generated for it in the coin-generation phase.
Together with α3 , they determine the message that is to be sent by Party j in
.
The auxiliary strings ρ j and ω j will be used to authenticate u and r j , as reflected
in the following definition of h.
r The function h is defined such that h((v1 , s1 ), (v2 , s2 ), v3 ) equals
(C s1 (v1 ), C s2 (v2 ), v3 ). Indeed, it holds that h(α1 , α2 , α3 ) = (C ρ j (u), C ω j (r j ),
α3 ) = β.
r The function f equals the computation that determines the message to be sent
in
. Note that this message is computable in polynomial-time from the party’s
input (denoted u and being part of α1 ), its random-tape (denoted r j and being
part of α2 ), and the messages it has received so far (i.e., α3 ). Indeed, it holds that
f (α1 , α2 , α3 ) is the message that Party j should send in
.

Recall that the party that plays the receiver in the current oracle call obtains either
f (α) or (h(α), f (α)). It treats the second case as if the other party has aborted,
which is also possible per se.
682

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

Aborting: In case any of the functionalities invoked in any of these phases terminates in
an abort state, the party (or parties) obtaining this indication aborts the execution,
and sets its output to ⊥. Otherwise, outputs are as follows.
Outputs: At the end of the emulation phase, each party holds the corresponding output
of the party in protocol
. The party just locally outputs this value.
Clearly, in case both parties are honest, the input–output relation of

is identical
to that of
. (We will show that essentially the same also holds in general.) We note
that the transformation of
to

can be implemented in polynomial-time. Finally,
replacing the oracle calls by the sub-protocols provided in Proposition 7.4.22 yields a
standard protocol for the malicious model.

7.4.4.1. The Effect of the Compiler

As will be shown, given a protocol as constructed in the proof of Theorem 7.3.12, the
compiler produces a protocol that securely computes the same functionality. Thus, for
any functionality f , the compiler transforms a specific protocol for privately computing
f (in the semi-honest model) into a protocol for securely computing f (in the malicious
model). This suffices to establish our main result (i.e., Theorem 7.4.1), yet it does not
say what the compiler does when given an arbitrary protocol (i.e., one not constructed
as in the proof of Theorem 7.3.12). In order to analyze the action of the compiler,
in general, we introduce the following model that is a hybrid of the semi-honest and
the malicious models.54 We call this new model, which is of independent interest, the
augmented semi-honest model.

Definition 7.4.24 (the augmented semi-honest model): Let
be a two-party protocol.

An augmented semi-honest behavior (with respect to
) is a (feasible) strategy that
satisfies the following conditions:

Entering the execution: Depending on its initial input, denoted u, the party may abort
before taking any step in the execution of
. Otherwise, again depending on u, it
enters the execution with any input u
∈ {0, 1}|u| of its choice. From this point on, u

is fixed.
Proper selection of a random-tape: The party selects the random-tape to be used in

uniformly among all strings of the length specified by
. That is, the selection of the
random-tape is exactly as specified by
.
Proper message transmission or abort: In each step of
, depending on its view of the
execution so far, the party may either abort or send a message as instructed by
.
We stress that the message is computed as
instructs based on input u
, the selected
random-tape, and all messages received so far.

54 Indeed, Theorem 7.4.1 will follow as a special case of the general analysis of the compiler (as provided later).
See further discussion following the statement of Proposition 7.4.25.

683
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Output: At the end of the interaction, the party produces an output depending on its
entire view of the interaction. We stress that the view consists of the initial input u,
the selected random-tape, and all messages received so far.
A pair of probabilistic polynomial-time strategies, C = (C 1 , C2 ), is admissible with
respect to
in the augmented semi-honest model if one strategy implements
and
the other implements an augmented semi-honest behavior with respect to
.

The augmented semi-honest model extends the ordinary semi-honest model in allowing
adversaries to modify their initial input and to abort the execution at an arbitrary time.
The augmented semi-honest model is arguably more appealing than the semi-honest
model because in many settings, input modification and aborting can also be performed
at a high level, without modifying the prescribed program. In contrast, implementing
an effective malicious adversary may require some insight into the original protocol,
and it typically requires substitution of the program’s code.
Intuitively, the compiler transforms any protocol
into an (oracle-aided) protocol


, such that executions of

in the malicious model correspond to executions of

in the augmented semi-honest model. That is:

Proposition 7.4.25 (general analysis of the two-party compiler): Let

be the (oracle-
aided) protocol produced by Construction 7.4.23 when given the protocol
, and let G
denote the set of the three oracle functionalities that are used by protocol

. Then, for
every pair of probabilistic polynomial-time strategies A = (A1 , A2 ) that are admissible
(with respect to

) for the (real) malicious model (of Definition 7.4.2),55 there exists a
pair of probabilistic polynomial-time strategies B = (B1 , B2 ) that are admissible with
respect to
for the augmented semi-honest model (of Definition 7.4.24), such that
c
{real
, B(z) (x, y)}x, y,z ≡ {real

G

, A(z) (x, y)} x, y,z

where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).

Proposition 7.4.25 will be applied to protocols as constructed in the proof of Theo-

rem 7.3.12. Actually, we will apply Proposition 7.4.25 to Theorem 7.3.14 (which pro-
vides canonical protocols for privately computing any functionality). As we shall see
(in Section 7.4.4.2), for these specific protocols, the augmented semi-honest model (of
Definition 7.4.24) can be emulated by the ideal malicious model (of Definition 7.2.4).
Thus, we obtain secure (oracle-aided) protocols (with oracle to G) for any functional-
ity, because (schematically speaking) for every functionality f , there exist
and


such that ideal f,malicious (x, y) equals real
,aug−semi−honest (x, y), which in turn equals
real
G

,malicious (x, y). (Ordinary secure protocols are obtained by using secure imple-

mentations of the oracles in G (which are provided by Proposition 7.4.22).) Thus,

Theorem 7.4.1 is proven by combining the properties of the compiler, as stated in

55 Recall the definition of real-model adversaries for an oracle-aided protocol (i.e., Definition 7.4.2) extends the
definition of real-model adversaries for ordinary protocols (i.e., Definition 7.2.5).

684

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

Proposition 7.4.25, with the properties of specific protocols to be compiled by it. We

start by establishing Proposition 7.4.25.

Proof Sketch: Given a pair of strategies, (A1 , A2 ), which is admissible with respect
to

for the real malicious model, we present a corresponding pair, (B1 , B2 ), that
is admissible with respect to
for the augmented semi-honest model. In the current
proof, the treatment of the two cases for the identity of the honest party is symmetric.
Hence, we use a generic symbol for the said identity. (Alternatively, without loss of
generality, one may assume that Party 1 is honest.)
We denote by hon the identity of the honest party and by mal the identity of the
malicious party (i.e., {hon, mal} = {1, 2}). Thus, Bhon is determined by
, and we
transform (the malicious adversary) Amal into (an augmented semi-honest adversary)
Bmal , which uses Amal as a subroutine. In particular, machine Bmal will emulate all
the oracles that are used in

(which is an oracle-aided protocol compiled out of the
ordinary protocol
). On input u ∈ {0, 1}n , machine Bmal behaves as follows:

Entering the execution: Machine Bmal invokes Amal on input u, and decides whether
to enter the protocol, and if so, with what input. Toward this end, machine Bmal
emulates the input-committing phase of

, using Amal (as subroutine). Machine Bmal
obtains from Amal the oracle-query that it makes to the input-committing functionality
(initiated by it), and uses this query to determine the replaced input u
(to be used
in the rest of the execution). It also provides Amal with the oracle answers that A mal
expects to get. Details follow.
Recall that the the input-committing phase consists of two invocations of the input-
committing functionality, one by Partyhon and the other by Partymal . In each invoca-
tion, one party supplies an input and the other party gets a commitment to it (while
the first party gets the corresponding commitment coins).

r In the invocation of the input-committing functionality in which Party commits

hon
to its input, machine Bmal generates a dummy commitment (supposedly to the
input of Partyhon ) and feeds it to Amal , which expects to get a commitment (as
2
answer from the oracle). Specifically, Bmal uniformly selects ρ hon ∈ {0, 1}n , and
def
computes the commitment γ hon = C ρ hon (0n ), where 0n is an arbitrary (dummy)
value (which replaces the unknown input of Partyhon ). Machine Bmal feeds Amal
with γ hon (as if γ hon were the oracle answer).
r In the invocation of the input-committing functionality in which Party com-
mal
mits to its input, machine Bmal tries to obtain the committed value (provided by
Partymal ) and feeds Amal with decommitment information (which it expects to
get). Specifically, Bmal obtains the query, denoted u
, that Amal makes to the input-
2
committing functionality, and feeds it with a uniformly selected ρ mal ∈ {0, 1}n .
We stress that Bmal will use this u
as its modified input in its (augmented semi-
honest) execution of
.
In case Amal has aborted this oracle call, machine Bmal aborts (i.e., does not enter
the execution of
).
685
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

In case Bmal did not abort, it enters protocol
with input u
. Note that this entire
step is implemented in polynomial-time, and the resulting u
depends only on u (the
initial input of Bmal ).
Selection of random-tape: Bmal selects its random-tape uniformly in {0, 1}(n) (as spec-
ified by
), and emulates the execution of the coin-generation phase of

ending
with this outcome, so as to place Amal in the appropriate state toward the protocol-
emulation phase. To achieve the latter goal, machine Bmal supplies Amal with the
oracle answers that it expects to see. Again, we distinguish between the two oracle
calls (to the augmented coin-tossing functionality) made during the coin-generation
phase of

:
r In the invocation of the augmented coin-tossing functionality in which Party ob-
hon
tains the outcome of the coin-toss, machine Bmal generates a dummy commitment
(supposedly to the random-tape of Partyhon ) and feeds it to Amal , which expects
to get a commitment (as answer from the oracle). Specifically, Bmal uniformly se-
def
lects ωhon ∈ {0, 1}n·(n) , and computes the commitment δ hon = C ωhon (0(n) ), where
0(n) is an arbitrary (dummy) value (which replaces the unknown random-tape of
Partyhon ). Machine Bmal feeds Amal with δ hon (as if δ hon were the oracle answer).
r In the invocation of the augmented coin-tossing functionality in which Party
mal
obtains the outcome of the coin-toss, machine Bmal first selects uniformly r mal ∈
{0, 1}(n) and ωmal ∈ {0, 1}n·(n) , and feeds Amal with the pair (r mal , ωmal ). Machine
Bmal will use r mal as its random-tape in its (augmented semi-honest) execution of

. If Amal aborts this oracle call, then Bmal aborts.
In case Bmal did not abort, it will use r mal as its random-tape in the subsequent steps
of protocol
. Note that this entire step is implemented in polynomial-time, and that
r mal is selected uniformly in {0, 1}(n) independent of anything else.
Subsequent steps – message transmission: Machine Bmal now enters the actual execu-
tion of
. It proceeds in this real execution along with emulating the correspond-
ing oracle answers of the authenticated-computation functionality. In a message-
transmission step by Partyhon (in
), machine Bmal obtains from Partyhon (in the
real execution of
) a message, and emulates the answer given to Partymal by the
authenticated-computation functionality. In a message-transmission step by Partymal
in
, machine Bmal computes the message to be sent to Partyhon (in
) as instructed
by
, based on the input u
determined earlier, the random-tape r mal selected earlier,
and the messages obtained so far from Partyhon (in
). It then checks if Amal makes
the correct oracle-query, in which case it sends Partyhon the message just computed,
and otherwise it aborts. Details follow:
r In a message-transmission step by Party (in
), machine Bmal first obtains
hon
from Partyhon (in the real execution of
) a message, denoted msg. Next, ma-
chine Bmal obtains from Amal the query that Amal makes to the authenticated-
computation functionality. Let us denote this query by β = (q1 , q2 , q3 ). If
(q1 , q2 ) = (γ hon , δ hon ) and q3 equals the sequence of messages sent so far (by
Bmal to Partyhon ), then Bmal feeds Amal with the received message msg. Other-
wise, Bmal feeds Amal with ((γ hon , δ hon , α3 ), msg), where α3 is the sequence of
686

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

messages sent so far (by Bmal to Partyhon ). (The latter case means that Amal is
cheating, but Partyhon does not detect this fact (because it obtains no answer from
the authenticated-computation functionality).)
r In a message-transmission step by Party (in
), machine Bmal first computes
mal
the message, denoted msg, that it should send (according to
) on input u
(as de-
termined earlier), random-tape r mal (as recorded earlier), and the messages re-
ceived so far (from Partyhon in execution of
). Next, machine Bmal obtains
from Amal the query that Amal makes to the authenticated-computation function-
ality. Let us denote this query by ((u

, ρ

), (r

, ω

), α3

). If C ρ

(u

) = C ρ mal (u
),
C ω

(r

) = C ωmal (r mal ), and α3

equals the sequence of messages received so far
(from Partyhon ), then Bmal sends the message msg to Partyhon . Otherwise, Bmal
aborts
. (The latter case means that Amal is cheating in

, and Partyhon detects
this fact and treats it as if Partymal has aborted in

.)

Output: Machine Bmal just outputs, whatever machine Amal outputs, given the execution
history (in

) emulated earlier.
Clearly, machine Bmal (as described) implements an augmented semi-honest behavior
with respect to
. It is left to show that
c
{real

, A(z) (x, y)} x, y,z ≡ {real
, B(z) (x, y)}x, y,z
G
(7.37)

There is only one difference between the two ensembles referred to in Eq. (7.37): In
the first distribution (i.e., real

G

, A(z) (x, y)), the commitments obtained by A mal in the

input-commitment and coin-generation phases are to the true input and true random-
tape of Partyhon . On the other hand, in the second distribution (i.e., real
, B(z) (x, y)),
the emulated machine Amal is given commitments to dummy values (and the ac-
tions of Bmal are determined accordingly). We stress that, other than this differ-
ence, Bmal perfectly emulates Amal . However, the difference is “undetectable” (i.e.,
computationally indistinguishable) due to the hiding property of the commitment
scheme.
Composing the oracle-aided protocols produced by the compiler with secure imple-
mentations of these oracles (as provided by Proposition 7.4.22), and using the Compo-
sition Theorem and Proposition 7.4.25, we obtain:

Corollary 7.4.26 (compilation of two-party protocols): Assuming the existence of

(non-uniformly strong) one-way functions, any two-party protocol
can be efficiently
transformed into a two-party protocol

such that the following holds. For every pair
of probabilistic polynomial-time strategies A = (A1 , A2 ) that are admissible (with re-
spect to

) for the (real) malicious model (of Definition 7.2.5), there exists a pair of
probabilistic polynomial-time strategies B = (B1 , B2 ) that are admissible with respect
to
for the augmented semi-honest model (of Definition 7.4.24), such that
c
{real
, B(z) (x, y)}x, y,z ≡ {real

, A(z) (x, y)}x, y,z

where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).
687
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

7.4.4.2. Canonical Protocols and the Augmented Semi-Honest Model

Recall that a protocol for privately computing some functionality is guaranteed to be
secure with respect to semi-honest behavior. Thus, a real semi-honest execution of this
protocol can be emulated by an ideal semi-honest computation of the functionality. The
question is what happens to such a protocol when it is run under the augmented semi-
honest model. We now show that for canonical protocols (e.g., the protocols constructed
in the proof of Theorem 7.3.12), a real augmented semi-honest execution of such a
protocol can be emulated by an ideal malicious computation of the functionality. That
is, these protocols have the salient property of allowing emulatation of the (wider)
class of real augmented semi-honest executions by the (wider) class of ideal malicious
computations. Combined with Corollary 7.4.26, this fact means that if one applies
the compiler to a canonical protocol
that privately computes f , then the resulting
protocol

securely computes f (because malicious executions of

can be emulated
by augmented semi-honest executions of
, which in turn can be emulated by the ideal
malicious model for f ).
Recall that the augmented semi-honest model allows two things that go beyond the
semi-honest model: (1) oblivious substitution of inputs, and (2) abort. The first type of
behavior has a correspondence in the malicious ideal model, and so poses no problem.
To account for the second type of behavior, we need to match an aborting execution
in the augmented semi-honest model with an aborting execution in the ideal malicious
model. Here is where the extra property of the specific protocols, constructed in the
proof of Theorem 7.3.12, comes to help. Specifically, we refer to the fact that these
protocols are canonical, which means that the output of each party is determined only
after it receives the very last message (and no knowledge of the output is obtained
before). Thus, aborting before this stage is essentially equivalent to not entering the
execution at all, whereas aborting at the last stage is accounted for by the malicious
ideal model.

Proposition 7.4.27 (on canonical protocols): Let
be a canonical protocol that pri-
vately computes the functionality f . Then, for every probabilistic polynomial-time pair
B = (B1 , B2 ) that is admissible for the (real) augmented semi-honest model (of Def-
inition 7.4.24), there exists a probabilistic polynomial-time pair C = (C 1 , C 2 ) that is
admissible for the ideal malicious model (of Definition 7.2.4) such that
c
{real
, B(z) (x, y)}x, y,z ≡ {ideal f,C(z) (x, y)}x, y,z
where x, y, z ∈ {0, 1}∗ such that |x| = |y| and |z| = poly(|x|).

We comment that the statement of Proposition 7.4.27 implicitly introduces a notion of

security in the augmented semi-honest model. Indeed, if the real-model adversary is
allowed augmented semi-honest behavior, then it is natural to allow a corresponding
behavior in the ideal model, which then coincides with the ideal malicious model.
Viewed in these terms, Proposition 7.4.27 asserts that canonical protocols are secure
in the augmented semi-honest model.

688

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

Proof Sketch: Recall that canonical protocols (cf. Definition 7.3.13) proceed in two
stages, where the first stage yields no information at all (to any semi-honest party) and
the second phase consists of the exchange of a single pair of messages (i.e., each party
sends a single message). We use the fact that canonical protocols admit a two-stage
simulation procedure (for the view of a semi-honest party). Such two-stage simulators
act as follows:

Input to simulator: A pair (u, v), where u is the initial input of the semi-honest party
and v the corresponding local output.
Simulation Stage 1: Based (only) on u, the simulator generates a transcript corre-
sponding to the view of the semi-honest party in the first stage of the canonical
protocol
.
Recall that this is a truncated execution of
, where the execution is truncated just
before the very last message is received by the semi-honest party. We stress that this
truncated view, denoted T , is produced without using v.
Simulation Stage 2: Based on T and v, the simulator produces a string corresponding
to the last message received by the semi-honest party. The simulator then outputs the
concatenation of T and this (last) message.

The reader may easily verify that any canonical protocol has two-stage simulators.
Loosely speaking, a simulator as required in Stage 1 is implicit in the definition of
a canonical protocol (cf. Definition 7.3.13), and the simulation of Stage 2 is trivial
(because Stage 1 in a canonical protocol ends with the parties holding shares of the
desired outputs, and Stage 2 consists of each party sending the share required by the
other party).
Next, for any protocol having two-stage simulators, given a pair (B1 , B2 ) that is
admissible with respect to
for the augmented semi-honest model, we construct a
pair, (C1 , C2 ) that is admissible for the ideal malicious model. We distinguish two
cases, corresponding to the identity of the honest party. The difference between these
cases amounts to the possibility of (meaningfully) aborting the execution after receiving
the last message (and just before sending the last message). This possibility exists for
a dishonest Party 1 but not for a dishonest Party 2 (see Figure 7.3).
We start with the case where Party 1 is honest (and Party 2 is dishonest). In this
case, C1 is determined (by
), and we need to transform the augmented semi-honest
real adversary B2 into a malicious ideal-model adversary C 2 . The latter operates as
follows, using the two-stage simulator, denoted S2 , provided for the view of Party 2 in
semi-honest executions of
(which privately computes f ). Recall that C2 gets input
y ∈ {0, 1}n .

1. Machine C2 first determines the input y
to be sent to the trusted party, where y

is determined according to the behavior of B2 during the entire emulation of the
(canonical) protocol
. In addition, C2 emulates the messages sent and received by

689
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Party 1 Party 2

Stage 1

(r1,r2) (s1,s2)

Stage 2
s1
meaningful
abort
r2

r1 + s1 r2+ s2
Figure 7.3: Schematic depiction of a canonical protocol.

B2 during the first phase of
, and also determines the last message of B2 (i.e., its
single Stage 2-message). This is done as follows:

(a) First, C2 computes the substituted input with which (the augmented semi-honest
adversary) B2 enters
. That is, y
← B2 (y). In case B2 aborts, machine C 2 sets
y
= ⊥ (so as to conform with the [simplifying] convention that the ideal-model
adversary always sends input to the trusted party).
(b) Next, C 2 invokes the first stage of the simulator S2 in order to obtain the view of
the execution of the first stage of
as seen by a semi-honest party having input
y
. Denote this view by T , and note that T includes y
. Machine C2 extracts from
T the random-tape, denoted r , of Party 2. This random-tape will be fixed for the
use of B2 .
(c) Using T , machine C 2 emulates the execution of B2 on input y
and random-tape
r , up to the point where Party 2 is to receive the last message (in
). We stress
that this point is just after Party 2 has sent its last message. Thus, the last message
of Party 2 (in
) is determined at this step. To perform the emulation, C 2 feeds
B2 with input y
and random-tape r , and iteratively feeds B2 with the sequence
of (incoming) messages as appearing in the corresponding locations in T . We
stress that although T is only the transcript of Stage 1 in
, it determines all
messages of Party 2 in
(including its single Stage 2 message).
Note that the augmented semi-honest strategy B2 may abort in such an execution,
but in case it does not abort, the messages it sends fit the transcript T . Conse-
quently, the view of (the augmented semi-honest adversary) B2 in an execution
of the first stage of
is emulated by a prefix of T (which in turn represents the
simulated view of a semi-honest party on input y
).
690

www.Ebook777.com
 7.4* FORCING (TWO-PARTY) SEMI-HONEST BEHAVIOR

In case B2 has aborted the execution (even just before sending the last message,
which belongs to Stage 2), machine C 2 resets y
to ⊥.

We stress that y
is determined based only on y, and that C 2 never aborts.

2. Machine C2 invokes the trusted party with input y
and obtains a response, denoted v.
(Since the trusted party answers Party 1 first, Party 2 does not have the option of
stopping the trusted party before it answers Party 1. But this option is not needed
because Party 2 cannot meaningfully abort
after receiving the last message in it.
That is, if B2 has not aborted so far, then it cannot (meaningfully) abort now, because
it has already sent (or rather determined) its last message.)

3. Finally, C 2 determines its output as follows:

(a) C2 invokes the second stage of the simulator S2 in order to obtain the last message
sent to Party 2. That is, C 2 supplies the simulator with the first-stage transcript
T and the output v, and obtains the last message, denoted msg.
(b) C2 now emulates the last step of B2 (i.e., its output computation) by supplying
it with the message msg.
(Note that the last message of B2 was already determined in Step 1, and so the
execution of C 2 ends here.)
The output of C 2 is set to be the output of B2 , regardless if B2 has aborted or completed
the execution.
We need to show that
c
{real
, B(z) (x, y)}x, y,z ≡ {ideal f,C(z) (x, y)}x, y,z (7.38)
Abusing notation, we replace the final value of y
by B2 (y), and get:
{real
, B (x, y)}x, y ≡ {(output


1 (x, B2 (y)) , B2 (view2 (x, B2 (y))))} x, y
c
≡ {( f 1 (x, B2 (y)) , B2 (S2 (y, f 2 (x, B2 (y)))))}x, y
≡ {( f 1 (x, C2 (y)) , C2 (y, f 2 (x, C 2 (y))))}x, y
≡ {ideal f,C (x, y)}x, y
where S2 (y, v) denotes the result of the two-stage simulation. Eq. (7.38) follows. We
stress that the first stage of the simulator S2 is used to determine the value of y
= B2 (y),
but unfortunately this fact is not explicit in our notation. An analogous comment holds
with respect to the case treated next (where Party 1 is dishonest).
We now turn to the case where Party 2 is totally honest (and Party 1 possibly dishonest).
In this case, C2 is determined, and we need to transform the augmented semi-honest
real adversary B1 into a malicious ideal-model adversary C1 . The latter operates as
follows, using the simulator, denoted S1 , provided for the view of Party 1. Recall that
C1 gets input x ∈ {0, 1}n .
1. Machine C 1 first determines the input x
to be sent to the trusted party, where x

is determined according to the behavior of B1 during Stage 1 of the (canonical)
691
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

protocol
. In addition, C1 emulates the messages sent and received by B1 during

the first phase of
. This is done as in the previous transformation of B2 to C2 , except
that here, the last message of B1 (i.e., its Stage 2 message) is still undetermined at
this step (and can be determined only when given the last message of Party 2, which
in turn is obtained only at Step 3).
2. Machine C1 invokes the trusted party with input x
and obtains a response,
denoted v.
We stress that, unlike in the case where Party 2 is dishonest, Party 1 (i.e., C 1 ) still
has the option of stopping the trusted party before it answers Party 2.
3. Next, C1 invokes the second stage of the simulator S1 in order to obtain the last
message sent (by Party 2) to Party 1. It supplies the simulator with the transcript of
the first stage and the output v, and obtains the last message, denoted msg.
4. Machine C1 now emulates the last step of B1 by supplying it with the message msg.
In case B1 aborts, machine C1 prevents the trusted party from answering Party 2 and
aborts. Otherwise, machine C 1 allows the trusted party to answer Party 2. We stress
that C1 does not abort in any prior step.
The output of C1 is set to be the output of B1 , regardless if B1 has aborted or completed
the execution.
We again need to show that Eq. (7.38) holds. The argument is analogous to the one
applied for a dishonest Party 2. Suppose first, for simplicity, that machine B1 never
aborts. In such a case, by definition of S1 ,
{real
, B (x, y)}x, y ≡ {(B1 (view


1 (B1 (x), y)) , output2 (B1 (x), y))}x, y
c
≡ {(B1 (S1 (B1 (x), f 1 (B1 (x), y))) , f 2 (B1 (x), y))}x, y
≡ {(C1 (x, f 1 (C1 (x), y)) , f 2 (C1 (x), y))}x, y
≡ {ideal f,C (x, y)}x, y
Next, suppose that B1 always aborts after receiving the last message, and before sending
its last message to Party 2. In this case, we have
{real
, B (x, y)}x, y ≡ {(B1 (view

1 (B1 (x), y)) , ⊥)} x, y
c
≡ {(B1 (S1 (B1 (x), f 1 (B1 (x), y))) , ⊥)}x, y
≡ {(C 1 (x, f 1 (C1 (x), y), ⊥) , ⊥)}x, y
≡ {ideal f,C (x, y)}x, y
In the general case, machine B1 may abort in some executions and not abort in others.
Whenever B1 aborts, it does so before sending its last message (possibly just after
receiving the last message). However, both the decision of whether or not to abort and
the output at such a case are determined by B1 based on its view of the execution so
far. This view can be simulated by S1 (B1 (x), f 1 (B1 (x), y)), which in turn is invoked by
C 1 in two stages (in Steps 1 and 3, respectively). Thus, Eq. (7.38) holds in this case,
too.
692

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

7.4.4.3. Conclusion: Proof of Theorem 7.4.1

Essentially, Theorem 7.4.1 follows by combining the following three results: (1) The-
orem 7.3.14 providing canonical protocols for privately computing any functionality,
(2) the general analysis of the compiler (i.e., Corollary 7.4.26), and (3) the special
properties of canonical protocols (i.e., Proposition 7.4.27). Specifically, let f be an ar-
bitrary functionality,
be a canonical protocol for privately computing f (guaranteed
by Theorem 7.3.14), and

be the protocol compiled from
by Construction 7.4.23
(using secure implementations of the functionalities in G). Now, let A be admissible
for the real malicious model (with respect to

), let B be the real-model adversary
(which is admissible with respect to
in the augmented semi-honest model) derived
by Corollary 7.4.26, and C be the ideal-model adversary (which is admissible for the
malicious model) derived by Proposition 7.4.27. Then
c
{real

, A(z) (x, y)}x, y,z ≡ {real
, B(z) (x, y)}x, y,z
c
≡ {ideal f,C(z) (x, y)}x, y,z

Theorem 7.4.1 follows.

7.5.* Extension to the Multi-Party Case

In this section, we extend the treatment of general secure protocols from the two-
party case to the multi-party case. Again, our ultimate goal is to design protocols that
withstand any feasible adversarial behavior, and again we proceed in two steps. We first
consider a benign type of adversary, called semi-honest, and construct protocols that
are secure with respect to such an adversary. The definition of this type of adversary
is very much the same as in the two-party case. Next, we turn to the case of general
adversary behavior, but here (unlike in the two-party case) we consider two different
models. The first model of malicious behavior mimics the treatment of adversaries in
the two-party case; it allows the adversary to control even a majority of the parties, but
it does not view the (unavoidable) early abort phenomena as a violation of security.
In the second model of malicious behavior, we assume that the adversary can control
only a strict minority of the parties. In this model, which would have been vacuous in
the two-party case, the early abort phenomena can be effectively prevented. We show
how to transform protocols secure in the semi-honest model into protocols secure in
each of the two malicious-behavior models. As in the two-party case, this is done by
forcing parties (in each of the latter models) to behave in an effectively semi-honest
manner.
The constructions are obtained by suitable modifications of the constructions used
in the two-party case. In fact, the construction of multi-party protocols for the semi-
honest model is a minor modification of the construction used in the two-party case. The
same holds for the compilation of protocols for the semi-honest model into protocols
for the first malicious model. When compiling protocols for the semi-honest model
into protocols for the second malicious model, we will use a new primitive, called
693
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Verifiable Secret Sharing (VSS), in order to “effectively prevent” minority parties from
aborting the protocol prematurely. Actually, we shall compile protocols secure in the
first malicious model into protocols secure in the second malicious model.
Our treatment touches upon a variety of issues that were ignored (or are inapplicable)
in the two-party case. These issues include the communication model (i.e., the type of
communication channels), the consideration of an external adversary, and the way the
latter selects dishonest parties (or corrupts parties). In particular, in some models (i.e.,
postulating private channels and a majority of honest participants), it is possible to obtain
secure protocols without relying on any intractability assumptions: See Section 7.6.

Teaching Tip. We strongly recommend reading Sections 7.2–7.4 before reading the
current section. In many places in the current section, motivating discussions and
technical details are omitted, while relying on the fact that analogue elaboration has
appeared in the treatment of the two-party case (i.e., in Sections 7.2–7.4).

7.5.1. Definitions
A multi-party protocol problem is cast by specifying a random process that maps se-
quences of inputs (one input per each party) to corresponding sequences of outputs.
Let m denote the number of parties. It will be convenient to think of m as being fixed,
yet one can certainly think of it as an additional parameter. An m-ary functionality, de-
noted f : ({0, 1}∗ )m → ({0, 1}∗ )m , is thus a random process mapping sequences of the
form x = (x1 , ..., xm ) into sequences of random variables, f (x) = ( f 1 (x), ..., f m (x)).
The semantics is that for every i, the i-th party, initially holds an input xi , and wishes to
obtain the i-th element in f (x1 , ..., x m ), denoted fi (x1 , ..., xm ). For example, consider
deterministic functionalities for computing the maximum, average, or any other statis-
tics of the individual values held by the parties (and see more examples in Exercises 14
and 15). The discussions and simplifying conventions made in Section 7.2.1 apply in
the current context, too. Most importantly, we assume throughout this section that all
parties hold inputs of equal length; that is, |x i | = |x j |.

Conventions Regarding the Number of Parties. For simplicity of exposition, we

assume throughout our exposition that m is fixed. From time to time, we comment
on what is needed in order to derive definitions (and constructions) for the case that
m is a parameter. We comment that it is natural to discuss multi-party functionalities
that are “uniform,” in the sense that there exist (uniform) algorithms for computing
them for each value of m (and of course each m-sequence). One such functionality
is the “universal functionality” that is given a description of a circuit, as well as a
corresponding sequence of inputs. (For example, the circuit may be part of the input of
each party, and in case these circuits are not identical, the value of the functionality is
defined as a sequence of ⊥’s.) Indeed, a universal functionality is natural to consider
also in the two-party case, but here (in view of the extra parameter m) its appeal is
enhanced.
694

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

7.5.1.1. The Communication Model and External Adversaries

In the definitional treatment of the two-party case, we viewed one of the communicating
parties as an adversary and considered its effect on the protocol’s execution. This
approach can be extended to the multi-party case, except that here, we may consider
coalitions of dishonest parties and their effect on the execution. Alternatively, we may
consider an (external) adversary that controls a subset of the parties that participate
in the execution. A variety of issues that arise includes the size of this subset, the
way it is selected (by the adversary), and the possible effect of the adversary on the
communication channels.

The Number of Parties Controlled by the Adversary. In the two-party case, we have
focused on the case in which the adversary is identified with one of the participants
in the execution. Clearly, the case in which the adversary controls both participants is
of no interest, but the case in which the adversary controls none of the participants
may be of interest in case the adversary can wire-tap the communication line (as will
be discussed). In the multi-party case, we will consider adversaries that control any
number of participants.56 (Of course, when defining security following the “ideal-vs.-
real” paradigm, we should insist that the corresponding ideal adversary controls the
same set of participants.)

The Selection of Parties Controlled by the Adversary. The notion of an external

adversary naturally leads to the issue of how this adversary selects the set of parties that
it controls. The basic (and simpler) model postulates that this set is determined before
the execution starts (and is, of course, not known to the honest parties). This model
is called non-adaptive as opposed to the adaptive model in which the adversary may
select the set of parties that it controls adaptively, during the execution of the protocol
and depending on information it has gathered so far. In this section, we only consider the
non-adaptive model, and defer the treatment of the adaptive model to Section 7.7.1.2.
We comment that the difference between the non-adaptive model and the adaptive
model becomes crucial when the number of parties (i.e., m) is treated as a parameter,
rather than being fixed.

The Communication Channels. Throughout this section, we assume a model of syn-

chronous communication. It is natural to assume that the external adversary may tap all
communication channels (i.e., specifically, the channels between honest parties). In such
a case, even an adversary that controls none of the participants is of interest, because it
may potentially gain information about the execution by wire-tapping. However, for the
sake of simplicity, we sometimes prefer to present and use definitions that refer to the
“private-channel model” (see also Section 7.6); that is, we sometimes presuppose that
honest parties may communicate in secrecy (or, put differently, we sometimes assume
that adversaries do not tap communication lines between honest parties). We comment
that in the non-adaptive model, the issue of implementing the private-channel model

56 Indeed, the case in which the adversary controls all parties is of no interest, and is often ignored.

695
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

over the “standard model” (i.e., providing secret communication) is well understood,
and can be (easily) decoupled from the main treatment. Specifically, protocols secure in
the private-channel model can be compiled to withstand wire-tapping adversaries (by
using encryption schemes). Similarly, we assume that messages sent between honest
parties arrive intact, whereas one may want to consider adversaries that may inject mes-
sages on the communication line between honest parties. Again, this can be counteracted
by the use of well-understood paradigms, in this case, the use of signature schemes.

7.5.1.2. The Semi-Honest Model

This model is defined exactly as in the two-party case (see Section 7.2.2.1). Recall that
a semi-honest party is one who follows the protocol properly, with the exception that
it keeps a record of all its intermediate computations. Loosely speaking, a multi-party
protocol privately computes f if whatever a set (or a coalition) of semi-honest parties
can obtain after participating in the protocol could be essentially obtained from the
input and output of these very parties. Thus, the only difference between the current
definition and the one used in the two-party case is that we consider the gain of a
coalition (rather than of a single party) from participating in the protocol.

Definition 7.5.1 (privacy with respect to semi-honest behavior, without wire-tapping):

Let f : ({0, 1}∗ )m → ({0, 1}∗ )m be an m-ary functionality, where f i (x1 , ..., x m ) de-
def
notes the i-th element of f (x1 , .., xm ). For I = {i 1 , ..., i t } ⊆ [m] = {1, ..., m}, we let
f I (x1 , ..., xm ) denote the subsequence f i1 (x1 , ..., x m ), ..., f it (x1 , ..., x m ). Let
be an
m-party protocol for computing f .57 The view of the i-th party during an execution
of
on x = (x 1 , ..., x m ), denoted viewi
(x), is defined as in Definition 7.2.1, and for
def
I = {i 1 , ..., i t }, we let view



I (x) = (I, viewi 1 (x), ..., viewi t (x)).

r (deterministic case) In case f is a deterministic m-ary functionality, we say that

privately computes f if there exists a probabilistic polynomial-time algorithm,
denoted S, such that for every I ⊆ [m], it holds that
c
{S(I, (xi1 , ..., xit ), f I (x))}x∈({0,1}∗ )m ≡ {view

I (x)} x∈({0,1}∗ )m (7.39)
r (general case) We say that
privately computes f if there exists a probabilis-
tic polynomial-time algorithm, denoted S, such that for every I ⊆ [m], it holds
that

{(S(I, (xi1 , ..., xit ), f I (x)) , f (x))}x∈({0,1}∗ )m (7.40)

c
≡ {(view


I (x) , output (x))}x∈({0,1}∗ )m

where output
(x) denotes the output sequence of all parties during the execution
represented in view

I (x).

57 As in Section 7.2, by saying that
computes (rather than privately computes) f , we mean that the output
distribution of the protocol (when played by honest or semi-honest parties) on the input sequence (x 1 , ..., xm )
is distributed identically to f (x 1 , ..., xm ).

696

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Eq. (7.40) asserts that the view of the parties in I can be efficiently simulated based
solely on their inputs and outputs. Note that view
I (x) includes only the local views of
parties in I , and does not include the messages sent between pairs of honest parties.
Thus, Definition 7.5.1 refers to the case in which the semi-honest parties do not (or
cannot) wire-tap the channels between honest parties (and, hence, is labeled “without
wire-tapping”), which is equivalent to assuming the existence of “private channels.”
To deal with the case of wire-tapping, one just needs to augment view
I (x) with the
transcript of the messages sent between all the pairs of honest parties. In this case, it is
more natural to consider an external adversary that obtains the views of all parties in
I , as well as all messages sent over all channels.
Definition 7.5.1 can be easily adapted to deal with a varying parameter m, by taking
advantage of the current order of quantifiers (i.e., “there exists an algorithm S such that
for every I ”).58 We also note that the simulator can certainly handle the trivial cases in
which either I = [m] or I = ∅. (The case I = [m] is always trivial, whereas the case
I = ∅ is trivial only because here we consider the case of no wire-tapping.)
As in the two-party case, Definition 7.5.1 is equivalent to a definition that can
be derived by following the real-vs.-ideal paradigm (analogously to the treatment in
Section 7.2.2.2).

7.5.1.3. The Two Malicious Models

We now turn to consider arbitrary feasible deviation of parties from a specified multi-
party protocol. As mentioned earlier, one may consider two alternative models:

1. A model in which the number of parties that deviate from the protocol is arbitrary.
The treatment of this case extends the treatment given in the two-party case. In
particular, in this model, one cannot prevent malicious parties from aborting the
protocol prematurely, and the definition of security has to account for this fact (if it
is to have a chance of being met).
2. A model in which the number of parties that deviate from the protocol is strictly less
than half the total number of parties. The definitional treatment of this case is sim-
pler than the treatment given in the two-party case. In particular, one may – in
some sense – (effectively) prevent malicious parties from aborting the protocol
prematurely.59 Consequently, the definition of security is “freed” from the need to
account for early stopping, and thus is simpler.
We further assume (toward achieving a higher level of security) that malicious parties
may communicate (without being detected by the honest parties), and may thus coor-
dinate their malicious actions. Actually, it will be instructive to think of all malicious
parties as being controlled by one (external) adversary. Our presentation follows the

58 Note that for a fixed m, it may make as much sense to reverse the order of quantifiers (i.e., require that “for
every I there exists an algorithm S I ”).
59 As we shall see, the assumption that malicious parties are in a minority opens the door to effectively preventing

them from aborting the protocol immaturely. This will be achieved by letting the majority parties have (together!)
enough information so as to be able to emulate the minority parties in case the latter abort.

697
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

ideal-vs.-real emulation paradigm introduced and used in the previous sections. The
difference between the two malicious models is reflected in a difference in the corre-
sponding ideal models, which captures the different types of benign behaviors that a
secure protocol is aimed at achieving. Another difference is in the number of malicious
parties considered in each model.

The first malicious model. Following the discussion in Section 7.2.3, we conclude
that three things cannot be avoided in the first malicious model:
1. Malicious parties may refuse to participate in the protocol (when the protocol is first
invoked). Actually, as explained in Section 7.2.3, this behavior may be viewed as a
special case of input substitution (as discussed in the next item).
2. Malicious parties may substitute their local inputs (and enter the protocol with inputs
other than the ones provided to them from the outside).
3. Malicious parties may abort the protocol prematurely (e.g., before sending their last
message).
Accordingly, the ideal model is derived by a straightforward generalization of Defini-
tion 7.2.4. In light of this similarity, we allow ourselves to be quite terse. To simplify
the exposition, we assume that for every I , first the trusted party supplies the adversary
with the I -part of the output (i.e., the value of f I ), and only next is it possibly allowed
(at the adversary’s discretion) to answer the other parties. Actually, as in the two-party
case, the adversary has the ability to prevent the trusted party from answering all parties
only in the case where it controls Party 1.60

Definition 7.5.2 (the ideal model – first malicious model): Let f : ({0, 1}∗ )m →
def
({0, 1}∗ )m be an m-ary functionality. For I = {i1 , ..., i t } ⊆ [m] = {1, ..., m}, let I =
[m] \ I and (x 1 , ..., x m ) I = (x i1 , ..., xit ). A pair (I, B), where I ⊆ [m] and B is a prob-
abilistic polynomial-time algorithm, represents an adversary in the ideal model. The
joint execution of f under (I, B) in the ideal model (on input x = (x1 , ..., xm ) and
(1)
auxiliary input z), denoted ideal f, I, B(z) (x), is defined by uniformly selecting a random-
(1) def
tape r for the adversary, and letting ideal f, I, B(z) (x) = ϒ(x, I, z, r ), where ϒ(x, I, z, r )
is defined as follows:
r In case Party 1 is honest (i.e., 1 ∈ I ),

ϒ(x, I, z, r ) = ( f I (x
) , B(x I , I, z, r, f I (x
)),

def
(7.41)

where x
= (x1
, ..., xm
) such that x i
= B(x I , I, z, r )i for i ∈ I and xi
= xi other-
def

wise.
r In case Party 1 is not honest (i.e., 1 ∈ I ), ϒ(x, I, z, r ) equals

(⊥|I | , B(x I , I, z, r, f I (x
), ⊥)) if B(x I , I, z, r, f I (x
)) = ⊥ (7.42)

( f I (x
) , B(x, I, z, r, f I (x
))) otherwise (7.43)

60 As in the two-party case, this convention is rather arbitrary; see the discussion at the end of Section 7.2.3.1.

698

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

where, in both cases, x
= (x1
, ..., xm
) such that xi
= B(x I , I, z, r )i for i ∈ I and
def

xi
= xi otherwise.
In all cases, the trusted party is invoked with possibly substituted inputs, denoted
x
= (x 1
, ..., xm
), where xi
= xi only if i ∈ I . Eq. (7.42) represents the case where
the trusted party is stopped right after supplying the adversary with the I -part of the
output (i.e., f I (x
)). This case is allowed only when 1 ∈ I , and so Party 1 can always
be “blamed” when this happens.61 Equations (7.41) and (7.43) represent the cases
where the trusted party is invoked with possibly substituted inputs, but is allowed to
answer all parties. We stress that either all honest parties get their output or all are
notified that the trusted party was stopped by the adversary. As usual, the definition of
security is obtained by requiring that for every feasible adversary in the real model,
there exists a corresponding adversary in the ideal model that achieves the same effect.
Specifically, in the real model, the adversary may tap all communication lines and
determine (adaptively) all the outgoing messages of all dishonest parties.

Definition 7.5.3 (Security in the first malicious model): Let f be as in Definition 7.5.2,
and
be an m-party protocol for computing f .
r The joint execution of
under (I, A) in the real model (on input a sequence
x = (x1 , ..., xm ) and auxiliary input z), denoted real
, I, A(z) (x), is defined as the
output sequence resulting from the interaction between the m parties, where the
messages of parties in I are computed according to A(x I , I, z) and the messages of
def
parties in I¯ = [m] \ I are computed according to
.62 Specifically, the messages of
malicious parties (i.e., parties in I ) are determined by the adversary A based on the
initial inputs of the parties in I , the auxiliary input z, and all messages sent so far
by all parties (including messages received by the honest parties [i.e., parties in I¯]).
r Protocol
is said to securely compute f (in the first malicious model) if for ev-
ery probabilistic polynomial-time algorithm A (representing a real-model adversary
strategy), there exists a probabilistic polynomial-time algorithm B (representing an
ideal-model adversary strategy), such that for every I ⊆ [m]
(1) c
{ideal f, I, B(z) (x)}x,z ≡ {real
, I, A(z) (x)}x,z
When the context is clear, we sometimes refer to
as an implementation of f .
We stress that the ideal-model adversary (i.e., B) controls exactly the same set of parties
(i.e., I ) as the the real-model adversary (i.e., A). Definition 7.5.3 (as well as the following
Definition 7.5.4) refers to an adversary that may wire-tap all communication channels.
This is reflected in the definition of real
, I, A(z) (x), which allows A to determine its
actions based on all messages communicated so far. (Thus, for m = 2, Definition 7.5.3
is stronger than Definition 7.2.6, because [unlike the latter] the former also refers to the

61 In fact, in the protocols presented in this work, early abort is always due to malicious behavior of Party 1. By
Definition 7.5.3, this translates to malicious behavior of Party 1 in the ideal model.
62 To fit the format used in Definition 7.5.2, the outputs of the parties (in real

, I, A(z) (x)) are arranged such that
the outputs of the honest parties appear on the left-hand side.

699
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

case I = ∅, which is non-trivial because it refers to an adversary that may wire-tap the
communication channel.) In order to derive a definition for the private-channel model,
one should modify the definition of real
, I, A(z) (x), such that A’s actions may depend
only on the messages received by parties in I .

The Second Malicious Model. In the second model, where malicious parties are in
a strict minority, the early-abort, phenomena can be effectively prevented. Thus, in
this case, there is no need to “tolerate” early-abort, and consequently our definition of
security requires “proper termination” of executions. This is reflected in the definition
of the ideal model, which actually becomes simpler.63

Definition 7.5.4 (security in the second malicious model, assuming an honest

majority): Let f and
be as in Definition 7.5.3:
r The ideal-model adversary is defined as in Definition 7.5.2, except that the abort
case captured by Eq. (7.42) is not allowed. The corresponding joint computation in
(2)
the ideal model, under (I, B), is denoted by ideal f, I, B(z) (x).
r The real-model adversary is defined exactly as in Definition 7.5.3. However, we will
only consider the case where such adversary controls strictly less than m/2 parties.
r Protocol
is said to securely compute f (in the second malicious model) if for ev-
ery probabilistic polynomial-time algorithm A (representing a real-model adversary
strategy), there exists a probabilistic polynomial-time algorithm B (representing an
ideal-model adversary strategy), such that for every I ⊂ [m] such that |I | < m/2, it
holds that
(2) c
{ideal f, I, B(z) (x)}x,z ≡ {real
, I, A(z) (x)}x,z

When the context is clear, we sometimes refer to
as an implementation of f .

We stress that in Definition 7.5.4, we consider only adversaries that control a strict
minority of the parties.

Discussion. The two alternative malicious models give rise to two appealing and yet
fundamentally incomparable notions of security. Put in other words, there is a trade-
off between the willingness to put up with early-abort (i.e., not consider it a breach
of security) and requiring the protocol to be robust also against malicious coalitions
controlling a majority of all parties. The question of which notion of security is prefer-
able depends on the application (or on the setting). In some settings, one may prefer
to be protected from malicious majorities, while giving up the guarantee that parties
cannot abort the protocol prematurely (while being detected doing so). On the other
hand, in settings in which a strict majority of the parties can be trusted to follow the
protocol, one may obtain the benefit of effectively preventing parties to abort the proto-
col prematurely. We stress that all definitions are easily adapted to deal with a varying
parameter m.

63 In this case, the definition extends the one presented in Section 7.2.3.2.

700

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

7.5.2. Security in the Semi-Honest Model

Our construction of private multi-party protocols (i.e., secure versus semi-honest behav-
ior) for any given multi-argument functionality follows the presentation of the two-party
case. For simplicity, we think of the number of parties m as being fixed. The reader may
verify that the dependence of our constructions on m is at most polynomial.
Our protocol construction adapts the one used in the two-party case (see Section 7.3).
That is, we consider a GF(2) circuit for evaluating the (deterministic) m-ary function-
ality f , and start by letting each party share its input bits with all other parties, such
that the sum of all shares equals the input bit. Next, scanning the circuit from its input
wires to its output wires, we propagate shares through the circuit gates, by using a
suitable private computation. As in the two-party case, we focus on the propagation of
shares through multiplication gates. That is, for Party i holding bits ai and bi , we wish
toconduct a  private computation
m such that this party ends up with a random bit ci and
m m
( i=1 ai ) · ( i=1 bi ) = i=1 ci holds. More precisely, we are interested in privately
computing the following randomized m-ary functionality:

((a1 , b1 ), ..., (am , bm )) → (c1 , ..., cm ) uniformly in {0, 1}m (7.44)

m
m
m
subject to ci = ai · bi . (7.45)
i=1 i=1 i=1

Thus, all that we need to do on top of Section 7.3 is to provide a private m-party
computation of this functionality. This is done by privately reducing, for arbitrary m,
the computation of Eq. (7.44) – (7.45) to the computation of the same functionality for
the case m = 2, which in turn coincides with Eq. (7.17) – (7.18). But first we need to
define an appropriate notion of a reduction. Indeed, the new notion of a reduction is
merely a generalization of the notion presented in Section 7.3.1.

7.5.2.1. A Composition Theorem

We wish to generalize the notion of privacy reduction presented in Section 7.3.1 (in
the context of two-party [semi-honest] computation). Here, the reduction is an m-party
protocol that may invoke a k-ary functionality in its oracle calls, where k ≤ m. In case
k < m, an oracle call also needs to specify the set of parties who are to provide the
corresponding k inputs. Actually, the oracle call needs to specify the order of these
parties (i.e., which party should supply which input, etc.). (We note that the ordering of
parties also needs to be specified in case k = 2, and indeed this was done implicitly in
Section 7.3.1, where the convention was that the party who makes the oracle request is
the one supplying the first input. In case k > 2, such a convention does not determine
the correspondence between parties and roles, and thus in the following we use an
explicit mechanism for defining the correspondence.)

Definition 7.5.5 (m-party protocols with k-ary oracle access): As in the two-party
case, an oracle-aided protocol is an ordinary protocol augmented by a pair of oracle-
tapes per each party, and oracle-call steps defined as follows. Each of the m parties
701
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

may send a special oracle-request message to all other parties. The oracle-request
message contains a sequence of k distinct parties, called the request sequence, that
are to supply queries in the current oracle call. In response, each party specified in the
request sequence writes a string, called its query, on its own write-only oracle-tape, and
responds to the requesting party with an oracle-call message. At this point, the oracle
is invoked and the result is that a string, not necessarily the same, is written by the
oracle on the read-only oracle-tape of each of the k specified parties. This k-sequence
of strings is called the oracle answer.

One may assume, without loss of generality, that the party who invokes the oracle is
the one who plays the role of the first party in the reduction (i.e., the first element in
the request sequence is always the identity of the party that requests the current oracle
call).

Definition 7.5.6 (multi-party privacy reductions):

r An m-party oracle-aided protocol is said to be using the k-ary oracle-functionality
f if the oracle answers are according to f . That is, when the oracle is invoked with
request sequence (i 1 , ..., i k ), and the query sequence q1 , ..., qk is supplied by parties
i 1 , ..., i k , the answer sequence is distributed as f (q1 , ..., qk ). Specifically, party i j
in the m-party protocol (the one which supplied q j ), is the one which obtains the
answer part f j (q1 , ..., qk ). As in Definition 7.3.2, we require that the length of each
query be polynomially related to the length of the initial input.
r An m-party oracle-aided protocol using the k-ary oracle-functionality f is said to
privately compute g if there exists a polynomial-time algorithm, denoted S, satis-
fying Eq. (7.40), where the corresponding views are defined in the natural manner.
r An m-party oracle-aided protocol is said to privately reduce the m-ary func-
tionality g to the k-ary functionality f if it privately computes g when using
the oracle-functionality f . In such a case, we say that g is privately reducible
to f ,
We are now ready to generalize Theorem 7.3.3:

Theorem 7.5.7 (Composition Theorem for the multi-party semi-honest model): Sup-
pose that the m-ary functionality g is privately reducible to the k-ary functionality f ,
and that there exists a k-party protocol for privately computing f . Then there exists an
m-party protocol for privately computing g.

As in the two-party case, the Composition Theorem can be generalized to yield tran-
sitivity of privacy reductions; that is, if g is privately reducible to f and f is privately
reducible to e, then g is privately reducible to e.

Proof Sketch: The construction supporting the theorem is identical to the one used in
the proof of Theorem 7.3.3: Let
g| f be an oracle-aided protocol that privately reduces
g to f , and let
f be a protocol that privately computes f . Then, a protocol
for

702

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

computing g is derived by starting with
g| f , and replacing each invocation of the
oracle by an execution of
f . Clearly,
computes g. We need to show that it privately
computes g (as per Definition 7.5.1).
We consider an arbitrary (non-trivial) set I ⊆ [m] of semi-honest parties in the
execution of
, where the trivial cases (i.e., I = ∅ and I = [m]) are treated (differently)
in a straightforward manner. Note that for k < m (unlike the situation in the two-
party case), the set I may induce different sets of semi-honest parties in the different
executions of
f (replacing different invocations of the oracle). Still, our “uniform”
definition of simulation (i.e., uniform over all possible sets of semi-honest parties)
keeps us away from trouble. Specifically, let S g| f and S f be the simulators guaranteed
for
g| f and
f , respectively. We construct a simulation S, for
, in the natural
manner. On input (I, x I , f I (x)), we first run S g| f (I, x I , f I (x)), and obtain the view
of the semi-honest coalition I = ∅ in
g| f . This view includes the sequence of all
oracle-call requests made during the execution, which in turn consists of the sequence
of parties that supply query-parts in each such call. The view also contains the query-
parts supplied by the parties in I , as well as the corresponding answer-parts. For each
such oracle call, we denote by J the subset of I that supplied query-parts in this
call and invoke S f , providing it with the subset J , as well as with the corresponding
J -parts of the queries and answers. Thus, we fill up the view of I in the current
execution of
f . (Recall that S f can also handle the trivial cases in which either |J | = k
or |J | = 0.)
It is left to show that S indeed generates a distribution indistinguishable from the
view of semi-honest parties in actual executions of
. As in the proof of Theorem 7.3.3,
this is done by introducing a hybrid distribution, denoted H . This hybrid distribution
represents the view of the parties in I (and output of all parties) in an execution of
g| f
that is augmented by corresponding invocations of S f . In other words, H represents
the execution of
, with the exception that the invocations of
f are replaced by
simulated transcripts. Using the guarantees regarding S f (resp., S g| f ), we show that
the distributions corresponding to H and
(resp., H and S) are computationally
indistinguishable. The theorem follows.

  
7.5.2.2. Privately Computing i ci = ( i ai ) · ( i bi )
We now turn to the m-ary functionality defined in Eq. (7.44) – (7.45). Recall that the
arithmetic is that of GF(2), and so −1 = +1, and so forth. The key observation is that
   

m
m
m
 
ai · bi = ai bi + ai b j + a j bi (7.46)
i=1 i=1 i=1 1≤i< j≤m

m

= (1 − (m − 1)) · ai bi + (ai + a j ) · (bi + b j )
i=1 1≤i< j≤m

m

=m· ai bi + (ai + a j ) · (bi + b j ) (7.47)
i=1 1≤i< j≤m

703
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

where the last equality relies on the specifics of GF(2). Now, looking at Eq. (7.47),
we observe that each party, i, can compute (by itself) the term m · ai bi , whereas each
2-subset, {i, j}, can privately compute shares to the term (ai + a j ) · (bi + b j ) by in-
voking the two-party functionality of Eq. (7.17) – (7.18). This leads to the following
construction:

Construction 7.5.8 (privately reducing the m-party computation of Eq. (7.44) – (7.45)
to the two-party computation of Eq. (7.17) – (7.18)):

Inputs: Party i holds (ai , bi ) ∈ {0, 1} × {0, 1}, for i = 1, ..., m.

Step 1 – Reduction: Each pair of parties, (i, j), where i < j, invokes the 2-ary func-
tionality of Eq. (7.17)–(7.18). Party i provides the input pair, (ai , bi ), whereas Party j
{i, j}
provides (a j , b j ). Let us denote the oracle response to Party i by ci , and the re-
{i, j}
sponse to Party j by c j .
 {i, j}
Step 2: Party i sets ci = mai bi + j =i ci .
Indeed, mai bi = 0 if m is even and mai bi = ai bi otherwise.
Outputs: Party i outputs ci .
We first observe that this reduction is valid; that is, the output of all parties indeed sum
up to what they should. It is also easy to see that the reduction is private. That is,

Proposition 7.5.9: Construction 7.5.8 privately reduces the computation of the m-ary
functionality given by Eq. (7.44)–(7.45) to the computation of the 2-ary functionality
given by Eq. (7.17)–(7.18).

Proof Sketch: We construct a simulator, denoted S, for the view of the parties in the
oracle-aided protocol, denoted
, of Construction 7.3.7. Given a set of semi-honest
parties, I = {i 1 , ..., i t } (with t < m), and a sequence of inputs (ai1 , bi1 ), ...., (ait , bit )
and outputs ci1 , ..., ci t , the simulator proceeds as follows:
{i, j}
1. For each pair, (i, j) ∈ I × I where i < j, the simulator uniformly selects ci ∈
{i, j} {i, j}
{0, 1} and sets c j = ci + (ai + a j ) · (bi + b j ).
def
2. Let I¯ = [m] \ I , and let  be the largest element in I¯. (Such an  ∈ [m] exists since
|I | < m).
{i, j}
(a) For each i ∈ I and each j ∈ I¯ \ {}, the simulator uniformly selects ci ∈
{0, 1}.  {i, j}
(b) For each i ∈ I , the simulator sets ci{i,} = ci + mai bi + j ∈{i,} ci , where the
{i, j}
latter ci ’s are as generated in Steps 1 and 2a.

{i, j}
3. The simulator outputs all ci ’s generated here. That is, it outputs the sequence of
{i, j}
ci ’s corresponding to all i ∈ I and j ∈ [m] \ {i}.

We claim that the output of the simulator is distributed identically to the view of the
parties in I during the execution of the oracle-aided protocol. Furthermore, we claim
704

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

that for every such I , every x = ((a1 , b1 ), ..., (am , bm )), and every possible outcome
(c1 , ..., cm ) of the functionality f of Eq. (7.44)–(7.45), it holds that the conditional
distribution of S(I, x I , f I (x)) is distributed identically to the conditional distribution
of view
I (x).
To prove this claim, we first note that f (x) is uniformly distributed over the m-
def  
bit, long sequences that sum up to c = ( i ai ) · ( i bi ). The same holds also for
the outputs of the parties in protocol
, because the sequence of the outputs of Par-
ties 1, ..., m − 1 is uniformly distributed over {0, 1}m−1 (due to the contribution of
{i,m}
ci to the output of Party i), whereas the sum of all m outputs equals c. Turning to
the conditional distributions (i.e., conditioning on f (x) = (c1 , ..., cm ) = output
(x)),
{i, j}
we show that the sequence of ci ’s (for i ∈ I ) is distributed identically in both distri-
butions (i.e., in the execution view and in the simulation). Specifically, in both cases,
{i, j}
the sequence (ci )i∈I, j∈[m]\{i} is uniformly distributed among the sequences satisfy-
{i, j} {i, j}  {i, j}
ing ci + c j = (ai + a j ) · (bi + b j ) (for each i ∈ I and j = i) and j =i ci =
ci + mai bi (for each i ∈ I ).
{i, j}
Details: Consider the distribution of the sub-sequence (ci )i∈I, j∈[m]\{i,} , where
 ∈ I¯ is as in the preceding. In both cases, the conditioning (on f (x) = (c1 , ..., cm ) =
{i,}
output
(x)) does not affect this distribution, because the ci ’s are missing. Thus,
in both cases, this sub-sequence is uniformly distributed among the sequences sat-
{i, j} {i, j}
isfying ci + c j = (ai + a j ) · (bi + b j ) (for each i = j ∈ I ). Furthermore, in
{i,}  {i, j}
both cases, the ci ’s are set such that j =i ci = ci + mai bi holds.

The proposition follows.

7.5.2.3. The Multi-Party Circuit-Evaluation Protocol

For sake of completeness, we explicitly present the m-party analogue of the protocol
of Section 7.3.4. Specifically, we show that the computation of any deterministic func-
tionality, which is expressed by an arithmetic circuit over GF(2), is privately reducible
to the functionality of Eq. (7.44) – (7.45).
Our reduction follows the overview presented in the beginning of this section. In
particular, the sharing of a bit-value v between m parties means a uniformly selected
m
m-sequence of bits (v1 , ..., vm ) satisfying v = i=1 vi , where the i-th party holds vi .
Our aim is to propagate, via private computation, shares of the input wires of the circuit
to shares of all wires of the circuit, so that finally we obtain shares of the output-wires
of the circuit.
We will consider an enumeration of all wires in the circuit. The input-wires of the
circuit, n per each party, will be numbered 1, 2...., m · n such that, for j = 1, ..., n,
the j-th input of Party i corresponds to the (i − 1) · n + j th wire. The wires will be
numbered so that the output-wires of each gate have a larger numbering than its input
wires. The output-wires of the circuit are the last ones. For the sake of simplicity, we
assume that each party obtains n output bits, and that the j-th output bit of the i-th
party corresponds to wire N − (m + 1 − i) · n + j, where N denotes the size of the
circuit.
705
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Construction 7.5.10 (privately reducing any deterministic m-ary functionality to the

functionality of Eq. (7.44) – (7.45), for any m ≥ 2): For simplicity, we assume that the
circuit is either fixed or can be determined in poly(n + m)-time as a function of n and
m, where n denotes the length of the input to each party.

Inputs: Party i holds the bit string xi = xi1 · · · xin ∈ {0, 1}n , for i = 1, ..., m.
Step 1 – Sharing the Inputs: Each party splits and shares each of its input bits with
all other parties. That is, for every i = 1, ..., m and j = 1, ..., n, and every k = i,
(i−1)n+ j
Party i uniformly selects a bit rk and sends it to Party k as the party’s share
of input-wire (i − 1) · n + j. Party i sets its own share of the (i − 1) · n + j th input
j (i−1)n+ j
wire to xi + k =i rk .
Step 2 – Circuit Emulation: Proceeding by the order of wires, the parties use their
shares of the two input wires to a gate in order to privately compute shares for the
Output-wire of the gate. Suppose that the parties hold shares to the two input-wires
of some gate; that is, for i = 1, ..., m, Party i holds the shares ai , bi , where a1 , ..., am
are the shares of the first wire and b1 , ..., bm are the shares of the second wire. We
consider two cases:
Emulation of an addition gate: Each party, i, just sets its share of the output-wire of
the gate to be ai + bi .
Emulation of a multiplication gate: Shares of the output-wire of the gate are ob-
tained by invoking the oracle for the functionality of Eq. (7.44) – (7.45), where
Party i supplies the input (query-part) (ai , bi ). When the oracle responds, each
party sets its share of the output-wire of the gate to equal its part of the oracle
answer.
Step 3 – Recovering the Output Bits: Once the shares of the circuit-output wires are
computed, each party sends its share of each such wire to the party with which the
wire is associated. That is, for i = 1, ..., m and j = 1, ..., n, each party sends its share
of wire N − (m + 1 − i) · n + j to Party i. Each party recovers the corresponding
output bits by adding up the corresponding m shares; that is, it adds the share it had
obtained in Step 2 to the m − 1 shares it has obtained in the current step.
Outputs: Each party locally outputs the bits recovered in Step 3.
As in the two-party case, one can easily verify that the output of the protocol is indeed
correct. Specifically, by using induction on the wires of the circuits, one can show that
the shares of each wire sum up to the correct value of the wire. Indeed, for m = 2, Con-
struction 7.5.10 coincides with Construction 7.3.9. The privacy of Construction 7.5.10
is also shown by extending the analysis of the two-party case; that is, analogously to
Proposition 7.3.10, one can show that Construction 7.5.10 privately reduces the com-
putation of a circuit to the multiplication-gate emulation.

Proposition 7.5.11: Construction 7.5.10 privately reduces the evaluation of arithmetic

circuits over GF(2), representing an m-ary deterministic functionality, to the function-
ality of Eq. (7.44) – (7.45).

706

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Proof Sketch: Just follow the proof of Proposition 7.3.10, treating the parties in I
analogously to the way that Party 1 is treated there. In treating the output wires of
parties in I (i.e., Step 3 in the simulation), note that the shares of parties in I and the
known output value uniquely determine the shares received in Step 3 of the protocol
only if |I | = m − 1 (as was the case in the proof of Proposition 7.3.10). Otherwise (i.e.,
for |I | < m − 1), the shares sent (in Step 3 of the protocol) by parties in I¯ should be
selected uniformly among all sequences that (together with the shares of parties in I )
add up to the given output value.

7.5.2.4. Conclusion: Private Computation of Any Functionality

As in Section 7.3, we may privately reduce the computation of a general (randomized)
m-ary functionality, g, to the computation of the deterministic m-ary functionality, f ,
defined by

def
f ((x 1 , r1 ), ..., (xm , rm )) = g(⊕i=1
m
ri , (x1 , ..., x m )) (7.48)

where g(r, x) denotes the value of g(x) when using coin-tosses r ∈ {0, 1}poly(|x|)
(i.e., g(x) is the randomized process consisting of uniformly selecting r ∈
{0, 1}poly(|x|) , and deterministically computing g(r, x)). Combining this fact with Propo-
sitions 7.5.11, 7.5.9, and 7.3.8 (and using the transitivity of privacy reductions),
we obtain:

Theorem 7.5.12: Any m-ary functionality is privately reducible to OT41 .

Combining Theorem 7.5.12 and Proposition 7.3.6 with the Composition Theorem (The-
orem 7.5.7), we conclude that if enhanced trapdoor permutations exist, then any m-ary
functionality is privately computable. As in the two-party case, we wish to highlight a
useful property of the protocols underlying the latter fact. Indeed, we refer to a notion
of canonical m-party computation that extends Definition 7.3.13.

Definition 7.5.13 (canonical semi-honest multi-party protocols): A protocol
for

privately computing the m-ary functionality f is called canonical if it proceeds by
executing the following two stages:

Stage 1: The parties privately compute the functionality x → ((r11 , ..., rm1 ), ...,
(r1m , ..., rmm )), where the r ij ’s are uniformly distributed among all possibilities that
satisfy (⊕i=1 m
r1i , ..., ⊕i=1
m
rmi ) = f (x).
Stage 2: For i = 2, ..., m and j ∈ [m] \ {i}, Party i sends r ij to Party j. Next, Party 1
sends r 1j to Party j, for j = 2..., m. Finally, each party computes its own output; that
is, for j = 1..., m, Party j outputs ⊕i=1m
r ij .

707
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Indeed, the protocols underlying the proof of Theorem 7.5.12 are essentially
canonical.64 Hence,

Theorem 7.5.14: Suppose that there exist collections of enhanced trapdoor permuta-
tions. Then any functionality can be privately computable by a canonical protocol.

We comment that the said protocols happen to maintain their security even if the adver-
sary can wire-tap all communication lines. This follows from the fact that privacy with
respect to wire-tapping adversaries happens to hold for all privacy reductions presented
in the current section, as well as for the protocols presented in Section 7.3.

7.5.3. The Malicious Models: Overview and Preliminaries

Our aim is to use Theorem 7.5.14 in order to establish the main result of this section;
that is,

Theorem 7.5.15 (main result for the multi-party case): Suppose that there exist collec-
tions of enhanced trapdoor permutations. Then any m-ary functionality can be securely
computable in each of the two malicious models, provided that a public-key infrastruc-
ture exists in the network.65

The theorem will be established in two steps. First, we compile any protocol for the
semi-honest model into an “equivalent” protocol for the first malicious model. This
compiler is very similar to the one used in the two-party case. Next, we compile any
protocol for the first malicious model into an “equivalent” protocol for the second
malicious model. The heart of the second compiler is a primitive, which is alien to
the two-party case, called Verifiable Secret Sharing (VSS). For simplicity, we again
think of the number of parties m as being fixed. The reader may again verify that the
dependence of our constructions on m is at most polynomial.
To simplify the exposition of the multi-party compilers, we describe them as pro-
ducing protocols for a communication model consisting of a single broadcast channel
(and no point-to-point links). In this model, in each communication round, only one
(predetermined) party may send a message, and this message arrives to all parties. We
stress that only this predetermined party may send a message in the said round (i.e.,
the message is “authenticated” in the sense that each other party can verify that, in-
deed, the message was sent by the designated sender). Such a broadcast channel can
be implemented via an (authenticated) Byzantine Agreement protocol, thus providing
an emulation of the broadcast model on the standard point-to-point model (in which a
broadcast channel does not exist).

64 This assertion depends on the exact implementation of Step 3 of Construction 7.5.10, and holds provided that
Party 1 is the last party to send its shares to all other parties.
65 That is, we assume that each party has generated a pair of keys for a signature scheme and has publicized its

verification-key (so that it is known to all other parties). This set-up assumption can be avoided if the network
is augmented with a broadcast channel.

708

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Recall that our goal is to transform protocols that are secure in the semi-honest
point-to-point model into protocols that are secure in the two malicious broadcast
models. Starting with (semi-honestly secure) protocols that operate in the point-to-
point communication model, we first derive equivalent protocols for the broadcast-
channel model, and only next we apply the two compilers, where each compiler takes
and produces protocols in the broadcast-channel model (which are secure with respect
to a corresponding type of adversaries). Thus, the full sequence of transformations
establishing Theorem 7.5.15 (based on Theorem 7.5.14) is as follows:
r We first use the pre-compiler (of Section 7.5.3.1) to transform a protocol
0 that
privately computes a functionality f in the (private-channel) point-to-point model
into a protocol

0 that privately computes f in the broadcast model (where no private
point-to-point channels exist).
Note that, since we refer to semi-honest behavior, we do not gain by having a broad-
cast channel, and we may only lose by the elimination of the private point-to-point
channels (because this allows the adversary to obtain all messages sent). However,
the protocols presented in Section 7.5.2 happen to be secure in the semi-honest
broadcast model,66 and so this pre-compiler is actually not needed (provided we
start with these specific protocols, rather than with arbitrary semi-honestly secure
protocols).
r Using the first compiler (of Section 7.5.4), we transform

(which is secure in
0
the semi-honest model) into a protocol

1 that is secure in the first malicious
model.
We stress that both

0 and

1 operate and are evaluated for security in a communi-
cation model consisting of a single broadcast channel. The same holds also for

2
mentioned next.
r Using the second compiler (of Section 7.5.5), we transform

(which is secure in
1
the first malicious model) into a protocol

2 that is secure in the second malicious
model.
r Finally, we use the post-compiler (of Section 7.5.3.2) to transform each of the pro-
tocols

1 and

2 , which are secure in the first and second malicious models when
communication is via a broadcast channel, into corresponding protocols,
1 and
2 ,
for the standard point-to-point model. That is,
1 (resp.,
2 ) securely computes f
in the first (resp., second) malicious model in which communication is via standard
point-to-point channels.
We stress that security holds even if the adversary is allowed to wire-tap the (point-
to-point) communication lines between honest parties.
We start by discussing the security definitions for the broadcast communication model
and by presenting the aforementioned pre-compiler and the post-compiler. Once this is

66 As noted at the very end of Section 7.5.2, these protocols also happen to be secure against semi-honest adversaries
that do wire-tape all communication channels. These protocols can be trivially converted to work in the broadcast
model by letting the honest parties ignore broadcast messages that are not intended for them. Indeed, in the
resulting protocol, the adversary receives all messages (including those intended for other parties), but it could
also obtain these messages in the original protocol by wire-tapping all point-to-point channels.

709
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

done, we turn to the real core of this section: the two compilers (which are applied to
protocols that operate in the broadcast model).

Definitions. Indeed, security in the broadcast model was not defined so far. However,
the three relevant definitions for the broadcast communication model are easily de-
rived from the corresponding definitions given in Section 7.5.1, where a point-to-point
communication model was used. Specifically, in defining security in the semi-honest
model, one merely includes the entire transcript of the communication over the (single)
broadcast channel in each party’s view. Similarly, when defining security in the two
malicious models, one merely notes that the “real execution model” (i.e., real
, I, A )
changes (since the protocol is now executed over a different communication media),
(1) (2)
whereas the “ideal model” (i.e., ideal f, I, B or ideal f, I, B ) remains intact.

7.5.3.1. Pre-Compiler (Emulating Private Channels)

It is easy to (securely) emulate a set of (private) point-to-point communication channels
over a (single) broadcast channel. All that one needs to do is use a secure public-key en-
cryption scheme. Specifically, a protocol
that operates in the (private) point-to-point
communication model is emulated as follows. First, each party randomly generates a
pair of encryption/decryption keys, posts the encryption-key on the broadcast chan-
nel, and keeps the decryption-key secret. Next, any party instructed (by
) to send a
message, msg, to Party i encrypts msg using the encryption-key posted by Party i, and
places the resulting ciphertext on the broadcast channel (indicating that it is intended
for Party i). Party i recovers msg by using its decryption-key and proceeds as directed
by
. Denote the resulting protocol by

. In the following, we merely consider the
effect of this transformation in the semi-honest model.

Proposition 7.5.16 (pre-compiler): Suppose that there exist collections of enhanced

trapdoor permutations. Then any m-ary functionality is privately computable in the
broadcast communication model. Furthermore, the protocol is canonical.

Proof Sketch: Let f be an m-ary functionality, and
be a protocol (guaranteed

by Theorem 7.5.14) for privately computing f in the (private-channel) point-to-point
communication model. Given a trapdoor permutation, we construct a secure public-key
encryption scheme and use it to transform
into

as described previously.
To simulate the view of parties in an execution of

(taking place in the broadcast
communication model), we first simulate their view in an execution of
(taking place
in the point-to-point communication model). We then encrypt each message sent by a
party that belongs to the semi-honest coalition, as this would be done in an execution
of

. Note that we know both the message and the corresponding encryption-key.
We do the same for messages received by semi-honest parties. All that remains is to
deal with messages, which we do not know, sent between two honest parties. Here, we
merely place an encryption of an arbitrary message. This concludes the description of
the “broadcast-model” simulator.
710

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

The analysis of the latter simulator combines the guarantee given for the “point-to-
point simulator” and the guarantee that the encryption scheme is secure. That is, the
ability to distinguish the output of the broadcast-model simulator from the execution
view (in the broadcast model) yields either (1) the ability to distinguish the output of
the “point-to-point” simulator from the execution view (in the point-to-point model) or
(2) the ability to distinguish encryptions under the public-key encryption scheme being
used. In both cases we reach contradiction to our hypothesis.

7.5.3.2. Post-Compiler (Emulating a Broadcast Channel)

Here we go the other way around (i.e., from the broadcast model to the point-to-point
model). We are given a protocol that securely computes (in one of the two malicious
models) some functionality, where the protocol uses a broadcast channel. We wish to
convert this protocol into an equivalent one that works in a point-to-point communi-
cation model. (Actually, we do not go all the way back, because we do not assume
these point-to-point lines to provide private communication.) Thus, all we need to do
is emulate a broadcast channel over a point-to-point network and do so in the presence
of malicious parties, which reduces to solving the celebrated Byzantine Agreement
problem. However, we have signature schemes at our disposal, and so we merely need
to solve the much easier problem known as authenticated Byzantine Agreement. For
the sake of self-containment, we define the problem and present a solution.
Authenticated Byzantine Agreement. We presuppose a synchronous point-to-point
model of communication and a signature scheme infrastructure. That is, each party
knows the verification-key of all other parties. Party 1 has an input bit, denoted σ , and
its objective is to let all honest parties agree on the value of this bit. In case Party 1 is
honest, the other parties must agree on its actual input, but otherwise they may agree
on any value (as long as they agree).

Construction 7.5.17 (Authenticated Byzantine Agreement): Let m denote the number

of parties. We assume that the signature scheme in use has signature of length that
depends only on the security parameter, and not on the length of the message to be
signed.67
Phase 1: Party 1 signs its input and sends the resulting input-signature pair to all
parties. Party 1 may terminate at this point.
Definition: A message is called (v, i)-authentic if it has the form (v, s p1 , ..., s pi ), where
p1 = 1, all p j ’s are distinct, and for every j = 1, ..., i, the string s p j is accepted as
a signature to (v, s p1 , ..., s p j−1 ) relative to the verification-key of party p j .
Observe that when Party 1 follows the protocol with input v, at Phase 1 it sends a
(v, 1)-authentic message to each party. For every i ≥ 2, if (v, s p1 , ..., s pi ) is (v, i)-
authentic, then (v, s p1 , ..., s pi−1 ) is (v, i − 1)-authentic.

67 Such a signature scheme can be constructed given any one-way function. In particular, one may use Construc-
tion 6.4.30. Maintaining short signatures is important in this application, because we are going to iteratively
sign messages consisting of (the concatenation of an original message and) prior signatures.

711
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Phase i = 2, ..., m: Each honest party (other than Party 1) inspects the messages it
has received at Phase i − 1, and forwards signed versions of the (·, i − 1)-authentic
messages that it has received. Specifically, for every v ∈ {0, 1}, if Party j has received
a (v, i − 1)-authentic message (v, s p1 , ..., s pi−1 ) such that all pk ’s are different from
j, then it appends its signature to the message and sends the resulting (v, i)-authentic
message to all parties.
We stress that for each value of v, Party j sends at most one (v, i)-authentic message
to all parties. Actually, it may refrain from sending (v, i)-authentic messages if it has
already sent (v, i
)-authentic messages for some i
< i.
Termination: Each honest party (other than Party 1) evaluates the situation as follows:

1. If, for some i 0 , i 1 ∈ [m] (which are not necessarily different), it has received both
a (0, i 0 )-authentic message and a (1, i 1 )-authentic message, then it decides that
Party 1 is malicious and outputs an error symbol, say ⊥.
2. If, for a single v ∈ {0, 1} and some i, it has received a (v, i)-authentic message,
then it outputs the value v.
3. If it has never received a (v, i)-authentic message, for any v ∈ {0, 1} and i, then
it decides that Party 1 is malicious and outputs an error symbol, say ⊥.
We comment that in the Distributed Computing literature, an alternative presentation
is preferred in which if a party detects cheating by Party 1 (i.e., in Cases 1 and 3),
then the party outputs a default value, say 0, rather than the error symbol ⊥.
The protocol can be easily adapted to handle non-binary input values. For the sake of
efficiency, one may instruct honest parties to forward at most two authentic messages
that refer to different values (because this suffices to establish that Party 1 is malicious).

Proposition 7.5.18: Assuming that the signature scheme in use is unforgeable, Con-
struction 7.5.17 satisfies the following two conditions:
1. It is infeasible to make any two honest parties output different values.
2. If Party 1 is honest, then it is infeasible to make any honest party output a value
different from the input of Party 1.
The claim holds regardless of the number of dishonest parties and even if dishonest
parties abort the execution.

In other words, Proposition 7.5.18 asserts that Construction 7.5.17 is essentially a

secure implementation of the (“broadcast”) functionality (v, λ, ..., λ) → (v, v, ..., v).
In particular, the case in which the honest parties output ⊥ can be accounted for by
the abort of an ideal-model adversary playing Party 1. We note that security as used
here is incomparable to security in either of the two malicious models. On the one
hand, we do not provide security with respect to an external adversary that only taps
the communication lines while not controlling any of the parties. That is, we do not
provide secrecy with respect to an external adversary, and indeed, this feature is not
required by the post-compiler (presented in the proof of Proposition 7.5.19). On the
other hand, we do provide security in the (stronger) sense of the second malicious model
712

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

but do so without limiting the number of dishonest parties. That is, for any number of
dishonest parties, the protocol effectively prevents dishonest parties from aborting (be-
cause abort is treated as sending some illegal message). In particular, the case in which
Party 1 does not even enter the execution is treated as the case in which it sent illegal
messages.
Proof Sketch: Fixing any j and v, suppose that in Phase i − 1, Party j receives a
(v, i − 1)-authentic message, and assume that i is the smallest integer for which this
happens. For this event to happen, it must be that i ≤ m, because the message must
contain i − 1 signatures from different parties (other than Party j itself).68 In such
a case, if Party j is honest, then it will send an authentic (v, i)-message in Phase i
(i ≤ m), and so all parties will receive an authentic (v, i)-message in Phase i. Thus, for
every v, if an honest party sees a (v, ·)-authentic message, then so do all other honest
parties, and Part 1 follows. Part 2 follows by observing that if Party 1 is honest and has
input v, then all honest parties see a (v, 1)-authentic message. Furthermore, none can
see a (v
, i)-authentic message, for v
= v and any i.

Proposition 7.5.19 (post-compiler): Suppose that one-way functions exist. Then any
m-ary functionality that is securely computable in the first (resp., second) malicious
broadcast model is also securely computable in the first (resp., second) malicious point-
to-point model, provided that a public-key infrastructure exists in the network.

Proof Sketch: The idea is to replace any broadcast message sent in the original protocol
by an execution of the Authenticated Byzantine Agreement (AuthBA) protocol. This
idea needs to be carefully implemented because it is not clear that the security of
AuthBA is preserved under multiple executions, and thus applying Proposition 7.5.18
per se will not do. The problem is that the adversary may use authenticated messages
sent in one execution of the protocol in order to fool some parties in a different execution.
This attack can be avoided in the current context by using identifiers (which can be
assigned consistently by the higher-level protocol) for each of the executions of the
AuthBA protocol. That is, authentic messages will be required to bear the distinct
identifier of the corresponding AuthBA execution (and all signatures will be applied to
that identifier as well). Thus, authentic messages of one AuthBA execution will not be
authentic in any other AuthBA execution. It follows that the proof of Proposition 7.5.18
can be extended to our context, where sequential executions of AuthBA (with externally
assigned distinct identifiers) take place.
The proof of security transforms any real-model adversary for the point-to-point
protocol to a real-model adversary for the broadcast-channel protocol. In particular,
the latter determines the messages posted on the broadcast channel exactly as an hon-
est party determines the values delivered by the various executions of AuthBA. In
the transformation, we assume that each instance of the AuthBA sub-protocol satis-
fies the conditions stated in Proposition 7.5.18 (i.e., it delivers the same value to all

68 Note that the said message cannot contain a signature of Party j due to the minimality of i: If the (v, i − 1)-
authentic message had contained a signature of Party j, then for some i
< i, Party j would have received a
(v, i
− 1)-authentic message in Phase i
− 1.

713
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

honest parties, and this value equals the one entered by the honest sender). In case the
assumption does not hold, we derive a forger for the underlying signature scheme.

7.5.4. The First Compiler: Forcing Semi-Honest Behavior

We follow the basic structure of the compiler presented in Section 7.4 for the two-party
case. Adapting that compiler to the multi-party setting merely requires generalizing
the implementation of each of the three phases (of the compiled two-party protocols).
Following is a high-level description of the multi-party protocols generated by the
corresponding compiler. Recall that all communication, both in the input protocol as
well as in the one resulting from the compilation, is conducted merely by posting
messages on a single broadcast channel.
Input-commitment phase: Each of the parties commits to its input bits. This will be
done using a multi-party version of the input-commitment functionality of Eq. (7.36).
Intuitively, malicious parties may (abort or) substitute their inputs during this phase,
but they may do so depending only on the value of the inputs held by malicious
parties.
Coin-generation phase: The parties generate random-tapes for each of the parties.
These random-tapes are intended to serve as the coins of the corresponding parties
in their emulation of the semi-honest protocol. Each party obtains the random-tape
to be held by it, whereas the other parties obtain commitments to this value. This
will be done using a multi-party version of the augmented coin-tossing functionality
of Eq. (7.35).
Intuitively, malicious parties may abort during this phase, but otherwise they end up
with a uniformly distributed random-tape.
Protocol emulation phase: The parties emulate the execution of the semi-honest pro-
tocol with respect to the inputs committed in the first phase and the random-tapes
selected in the second phase. This will be done using a multi-party version of the
authenticated-computation functionality of Eq. (7.33). The fact that the original pro-
tocol is executed over a broadcast channel is used here.
Intuitively, malicious parties may abort during this phase, but otherwise they end up
sending messages as directed by the semi-honest protocol.
In order to implement these phases, we define natural generalizations of the input-
commitment, coin-tossing, and authenticated-computation functionalities (of the two-
party case), and present secure implementations of them in the current (first malicious)
multi-party model. The original definitions and constructions are obtained by setting
m = 2. We start again by defining an adequate notion of reducibility, which allows a
modular presentation of the compiled protocols.

7.5.4.1. Security Reductions and a Composition Theorem

Analogously to Section 7.5.2.1, we now define what we mean by saying that one func-
tionality is securely reducible to another functionality. We use the same definition of an
714

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

oracle-aided protocol (i.e., Definition 7.5.5), but require such a protocol to be secure
in the first malicious model (rather than be secure in the semi-honest model). As in the
two-party case, we require that the length of each oracle-query can be determined from
the length of the initial input to the oracle-aided protocol.

Definition 7.5.20 (Security Reductions in the First Malicious Model):

r As in Definition 7.5.6, an m-party oracle-aided protocol is said to be using the k-
party oracle-functionality f if the oracle answers are according to f . However, in
accordance with the definition of the (first) ideal model (for the invoked functionality),
the oracle does not answer all parties concurrently, but rather answers first the real-
model party that requested this specific oracle call (in the oracle-aided protocol).
When receiving its part of the oracle answer, the party that requested the oracle call
instructs the oracle whether or not to respond to the other parties.
We consider only protocols in which the length of each oracle-query is a polynomial-
time computable function of the length of the initial input to the protocol. Further-
more, the length of each query must be polynomially related to the length of the
initial input.
Analogously to Definition 7.5.3, the joint execution of an oracle-aided protocol

with oracle f under (I, A) in the real model (on input sequence x = (x 1 , ..., x m ) and
f
auxiliary input z), denoted real
, I, A(z) (x), is defined as the output sequence resulting
from the interaction between the m parties, where the messages of parties in I are
computed according to A(x I , z), the messages of parties not in I are computed
according to
, and the oracle calls are answered according to f .
r An oracle-aided protocol
, using the oracle-functionality f , is said to securely
compute g (in the first malicious model) if a condition analogous to the one in
Definition 7.5.3 holds. That is, the effect of any efficient real-model adversary as in
the previous item can be simulated by a corresponding ideal-model adversary, where
the ideal model for computing g is exactly as in Definition 7.5.2.

More specifically, the oracle-aided protocol
(using oracle f ) is said to securely

compute g (in the first malicious model) if for every probabilistic polynomial-
time A, there exists a probabilistic polynomial-time B such that for every I ⊆ [m]

(1) c f
{idealg, I, B(z) (x)}x,z ≡ {real
, I, A(z) (x)}x,z

r An oracle-aided protocol is said to securely reduce g to f (in the first malicious

model) if it securely computes g (in the first malicious model) when using the oracle-
functionality f . In such a case, we say that g is securely reducible to f ,
Indeed, when it is clear from the context, we often omit the qualifier “in the first malicious
model.”

We are now ready to state a composition theorem for the first multi-party malicious
model.
715
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Theorem 7.5.21 (Composition Theorem for the first multi-party malicious model):
Suppose that the m-ary functionality g is securely reducible to the k-ary functionality
f and that there exists a k-party protocol for securely computing f . Then there exists
an m-party protocol for securely computing g.

Recall that the syntax of oracle-aided protocols disallows concurrent oracle calls, and
thus Theorem 7.5.21 is actually a sequential composition theorem. As in the two-
party case, the Composition Theorem can be generalized to yield transitivity of secure
reductions and to account for reductions that use several oracles rather than one.
Proof Sketch: Analogously to the proof of previous composition theorems, we are given
an oracle-aided protocol, denoted
g| f , that securely reduces g to f , and an ordinary
protocol
f that securely computes f . Again, we construct a protocol
for computing
g in the natural manner; that is, starting with
g| f , we replace each invocation of the
oracle (i.e., of f ) by an execution of the protocol
f . Clearly,
computes g, and
we need to show that
securely computes g. This is proven by merely generalizing
the proof of Theorem 7.4.3 (i.e., the two-party case). The only point that is worthwhile
stressing is that the real-model adversary for
f , derived from the real-model adversary
for
, is constructed obliviously of the set of parties I that the adversary controls.69 As
in the proof of Theorem 7.5.7, we determine the set of parties for every such invocation
of
f , and rely on the fact that security holds with respect to adversaries controlling
any subset of the k parties participating in an execution of
f . In particular, the security
of an invocation of
f by parties P = { p1 , ..., pk } holds also in case I ∩ P = ∅, where
it means that a real-model adversary (which controls no party in P) learns nothing by
merely tapping the broadcast channel.70

7.5.4.2. Secret Broadcast

In order to facilitate the implementation of some functionalities, we introduce the
following secret-broadcast functionality:
(α, 1|α| , ..., 1|α| ) → (α, α, ..., α) (7.49)
At first glance, it seems that Eq. (7.49) is trivially implementable by Party 1 posting α on
the broadcast channel. This solution is “secure” as long as the (real-model) adversary
controls a non-empty set of parties, but fails in case the adversary controls none of
the parties and yet can tap the broadcast channel. That is, the trivial solution does
not provide secrecy with respect to an external adversary (which taps the channel but
controls none of the parties and thus is not supposed to learn the value sent by Party 1 to
all other parties). Note that secrecy with respect to an external adversary also arises in a
subtle way when we do not care about it a priori (e.g., see the proof of Theorem 7.5.21).

Proposition 7.5.22: Assuming the existence of trapdoor permutations, there exists a

secure implementation of Eq. (7.49) in the first malicious model.

69 Unlike in the two-party case, here we cannot afford to consider a designated adversary for each subset of parties.
70 Security holds also in the other extreme case, where I ∩ P = P, but it is not meaningful in that case.

716

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Proof Sketch: The first idea that comes to mind is to let each party generate a pair of
keys for a public-key encryption scheme and broadcast the encryption-key, and then let
Party 1 broadcast the encryption of its input under each of these encryption-keys. The
problem with this protocol is that it is no longer guaranteed that all parties receive the
same value. One solution is to let Party 1 provide zero-knowledge proofs (to each of
the parties) that the posted ciphertexts are consistent (i.e., encrypt the same value), but
the implementation of this solution is not straightforward (cf. Construction 7.5.24). An
alternative solution, adopted here, is to use the encryption scheme in order to emulate a
set of private (point-to-point) channels, as in Section 7.5.3.1, and run an authenticated
Byzantine Agreement on this network. Since we have an ordinary broadcast channel at
our disposal, we do not need to assume an initial set-up that corresponds to a public-key
infrastructure, but can rather generate it on the fly. The resulting protocol is as follows:

1. Each party generates a pair of keys for a signature scheme and posts the verification-
key on the broadcast channel. This establishes the public-key infrastructure as relied
upon in Construction 7.5.17.
2. Each party generates a pair of keys for a public-key encryption scheme and posts
the encryption-key on the broadcast channel. This effectively establishes a network
of private (point-to-point) channels to be used in Step 3.
3. The parties invoke the authenticated Byzantine Agreement protocol of Construc-
tion 7.5.17 in order to let Party 1 broadcast its input to all other parties. All messages
of this protocol are sent in encrypted form, where each message is encrypted using
the encryption-key posted in Step 2 by the designated receiver.
Combining the ideas underlying the proofs of Propositions 7.5.16 and 7.5.18 (and con-
sidering two cases corresponding to whether I is empty or not), the current proposition
follows.

7.5.4.3. Multi-Party Authenticated Computation

We start our assembly of multi-party functionalities by presenting and implement-
ing a multi-party generalization of the authenticated-computation functionality of
Eq. (7.33).

Definition 7.5.23 (authenticated computation, multi-party version): Let f : {0, 1}∗ ×

{0, 1}∗ → {0, 1}∗ and h : {0, 1}∗ → {0, 1}∗ be polynomial-time computable. The
h-authenticated f -computation m-party functionality is defined by

(α, β2 , ..., βm ) → (λ, v2 , ..., vm ) (7.50)

def def
where vi = f (α) if βi = h(α) and vi = (h(α), f (α)) otherwise, for each i.71

71 Indeed, an alternative multi-party generalization may require that all vi ’s equal f (α) if β2 = · · · = βm = h(α)
and equal (h(α), f (α)) otherwise. However, this alternative generalization seems harder to implement, whereas
Eq. (7.50) suffices for our application.

717
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Note that the obvious reduction of Eq. (7.50) to the two-party case (i.e., to Eq. (7.33))
does not work (see Exercise 16). As in the two-party case, we will securely reduce
Eq. (7.50) to an adequate multi-party generalization of the image-transmission func-
tionality and provide a secure implementation of the latter. We start by implementing the
adequate multi-party generalization of the image-transmission functionality, defined as
follows:

(α, 1|α| , ..., 1|α| ) → (λ, f (α), ..., f (α)) (7.51)

Indeed, Eq. (7.51) is essentially a special case of Eq. (7.50). We stress that in a secure
implementation of Eq. (7.51), either all parties obtain the same f -image or they all
obtain an indication that Party 1 has misbehaved. Thus, the honest parties must be in
agreement regarding whether or not Party 1 has misbehaved, which makes the gener-
alization of the two-party protocol less obvious than it may seem. In particular, the fact
that we use a proof system of perfect completeness plays a central role in the analysis of
the multi-party protocol. The same holds with respect to the fact that all messages are
sent using (secret) broadcast (and so the honest parties agree on their value). Together,
these two facts imply that any party can determine whether some other party has “jus-
tifiably rejected” some claim, and this ability enables the parties to reach agreement
regarding whether or not Party 1 has misbehaved.

def
Construction 7.5.24 (image-transmission protocol, multi-party version): Let R =
{(v, w) : v = f (w)}. For simplicity, we assume that f is length-regular; that is,
| f (x)| = | f (y)| for every |x| = |y|.

Inputs: Party 1 gets input α ∈ {0, 1}∗ , and each other party gets input 1n , where n = |α|.
def
Step C1: Party 1 secretly broadcasts v = f (α). That is, Party 1 invokes Eq. (7.49) with
input v, whereas each other party enters the input 1| f (1 )| and receives the output v.
n

Step C2: For i = 2, ..., m, Parties 1 and i invoke a zero-knowledge strong-proof-of-

knowledge system for R such that Party 1 plays the prover and Party i plays the
verifier. The common input to the proof system is v, the prover gets α as auxiliary
input, and its objective is to prove that it knows a w such that (v, w) ∈ R (i.e.,
v = f (w)). In case the verifier rejects the proof, Party i sends the coins used by the
verifier so that all other parties can be convinced of its justifiable rejection, where the
latter corresponds to the view of the verifier in a rejecting interaction. All messages
of the proof system are sent using the secret broadcast functionality.
Outputs: For i = 2, ..., m, if Party i sees some justifiable rejection, then it outputs ⊥;
otherwise it outputs v. (Party 1 has no output.)
Agreement on whether or not Party 1 has misbehaved is reduced to the decision whether
or not some verifier has justifiably rejected in Step C2, where of the latter decision
depends on information available to all parties. A key observation is that if Party 1
is honest, then no party can justifiably reject its proof in Step C2, because the proof
system has perfect completeness (which means that there exists no random-tape that
makes the verifier reject a claim by an honest prover). Note that Construction 7.5.24 is
718

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

actually an oracle-aided protocol, using the secret broadcast oracle. Consequently, if the
real-model adversary controls none of the parties, then it learns nothing (as opposed
to what might have happened if we were to use an ordinary broadcast in Steps C1
or C2).

Proposition 7.5.25: Suppose that the proof system, (P, V ), used in Step C2 is indeed a
zero-knowledge strong-proof-of-knowledge for the relation R. Then Construction 7.5.24
securely reduces Eq. (7.51) to Eq. (7.49).

Proof Sketch: The proof extends the two-party case treated in Proposition 7.4.12. Here,
we transform any real-model adversary A into a corresponding ideal-model adversary
B, where both get the set I as auxiliary input. The case I = ∅ is handled by relying
on the secret broadcast functionality (which implies that in this case, the real-model
adversary, which refers to an oracle-aided protocol in which all messages are sent using
Eq. (7.49), gets nothing). Otherwise, the operation of B depends on whether or not
1 ∈ I , which corresponds to the cases handled in the two-party case.
As in the two-party case, when transforming real-model adversaries to ideal-model
adversaries, we sometimes allow the latter to halt before invoking the trusted party.
This can be viewed as invoking the trusted party with a special abort symbol, where in
this case, the latter responds to all parties with a special abort symbol.
We start with the case where the first party is honest, which means here that 1 ∈ I .
In this case, the input to B consists essentially of 1n , where n = |α|, and it operates as
follows (assuming I = ∅):

1. Acting on behalf of each party in I , the ideal-model adversary B sends 1|α| to

the trusted party and obtains the answer v, which equals f (α) for α handed (to
the trusted party) by (the honest) Party 1. Thus, indeed, (v, α) ∈ R. (Recall that
Party 1 always obtains λ from the trusted party, but the other parties in I¯ = [m] \ I
obtain v.)
2. For i = 2, ..., m, machine B invokes the simulator guaranteed for the zero-knowledge
proof system (P, V ), on input v, using (the residual) A as a possible malicious verifier.
Note that we are simulating the actions of the prescribed prover P, which in the real
protocol is played by the honest Party 1. Once one simulation is finished, its transcript
becomes part of the history fed to A in subsequent simulations. Denote the obtained
sequence of simulation transcripts by S = S(v).
3. Finally, B feeds A with the alleged execution view (v, S) and outputs whatever A
does.

The computational indistinguishability of the output of the real-model adversary under

(A, I ) and the output of the ideal-model adversary under (B, I ) follows from the guar-
anteed quality of the zero-knowledge simulator. In addition, we need to consider the
outputs of the honest parties (i.e., the parties in I¯), and specifically the outputs of parties
in I¯ \ {1} (since Party 1 has no output). (Indeed, this is an issue only if I¯ \ {1} = ∅,
which is the reason that this issue did not arise in the two-party case.) In the ideal-model
execution, each party in I¯ \ {1} outputs v = f (α), and we have to prove that the same
719
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

occurs in the real-model execution (when Party 1 is honest). This follows from the
perfect completeness of (P, V ), as discussed earlier.
We now turn to the case where the first party is dishonest (i.e., 1 ∈ I ). In this
case, the input to B includes α, and it operates as follows (ignoring the easy case
I = [m]):
1. B invokes A on input α, and obtains the Step C1 message, denoted v, that A instructs
Party 1 to send (i.e., v = A(α)). As (implicit) in the protocol, any action of A in
Step C1 (including abort) is interpreted as sending a string.
2. B tries to obtain a pre-image of v under f . Toward this end, B uses the (strong)
knowledge-extractor associated with (P, V ). Specifically, providing the strong
knowledge-extractor with oracle access to (the residual prover) A(α), machine B
tries to extract (from A) a string w such that f (w) = v. This is done for each of the
| I¯| executions of the proof system in which the verifier is played by an honest party,
while updating the history of A accordingly.72 In case the extractor succeeds (in one
def def
of these | I¯| attempts), machine B sets α
= w. Otherwise, B sets α
= ⊥.
3. B now emulates an execution of Step C2. Specifically, for each i ∈ I¯, machine B lets
the adequate residual A play the prover, and emulates by itself the (honest) verifier
interacting with A (i.e., B behaves as a honest Party i). (The emulation of the proofs
given to parties in I is performed in the straightforward manner.) Next, B decides
whether or not to invoke the trusted party and let it respond to the honest parties.
This decision is based on all the m − 1 emulated proofs.
r In case any of the m − 1 emulated verifiers rejects justifiably, machine B aborts
(without invoking the trusted party), and outputs whatever A does (when fed with
these emulated proof transcripts).
r Otherwise (i.e., no verifier rejects justifiably), we consider two sub-cases:

(a) If α
= ⊥, then B sends α
(on behalf of Party 1) to the trusted party and
allows it to respond the honest parties. (The response will be f (α
), which
by Step 2 must equal v.)
(b) Otherwise (i.e., α
= ⊥, indicating that extraction has failed), B fails. (Note
that this means that in Step 3 the verifier was convinced, while in Step 2
the extraction attempt has failed.)
4. Finally, B feeds A with the execution view, which contains the prover’s view of the
emulation of Step C2 (produced in Step 3), and outputs whatever A does.
As in the two-party case (see proof of Proposition 7.4.12), the real-model execution
differs from the ideal-model execution only in case the real-model adversary A succeeds
in convincing the knowledge-verifier (which is properly emulated for any i ∈ I¯) that it
knows a pre-image of v under f , and yet the knowledge-extractor failed to find such a
pre-image. By definition of strong knowledge-verifiers, such an event may occur only
with negligible probability.

72 If necessary (i.e., | I¯| = {2, ..., | I¯| + 1}), we also emulate the interleaved proofs that are given to parties in I .
This is performed in the straightforward manner (i.e., by letting A emulate both parties in the interaction).

720

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Securely Reducing Authenticated Computation to Image Transmission. Analo-

gously to the two-party case, we securely reduce Eq. (7.50) to Eq. (7.51).

Construction 7.5.26 (multi-party authenticated computation, oracle-aided protocol):

Inputs: Party 1 gets input α ∈ {0, 1}∗ , and Party i = 1 gets input βi ∈ {0, 1}|α| .
Step C1: Party 1 uses the (multi-party) image-transmission functionality to send the
def
pair (u, v) = (h(α), f (α)) to the other parties. That is, the parties invoke the func-
tionality of Eq. (7.51), where Party 1 enters the input α and Party i is to obtain
def
g(α) = (h(α), f (α)).
Step C2: Assuming that Step C1 was not aborted by Party 1 and that Party i receives
the pair (u, v) in Step C2, Party i outputs v if u = βi and (u, v) otherwise.
Outputs: If not aborted (with output ⊥), Party i = 1 sets its local output as directed in
Step C2. (Party 1 has no output.)
Extending the proof of Proposition 7.4.15 (to apply to Construction 7.5.26), and using
Propositions 7.5.25 and 7.5.22, we obtain:

Proposition 7.5.27: Assuming the existence of trapdoor permutations, the h-

authenticated f -computation m-party functionality of Eq. (7.50) can be securely im-
plemented in the first malicious model.

Proof Sketch: We focus on the analysis of Construction 7.5.26, which extends the
proof of Proposition 7.4.15. As in the proof of Proposition 7.5.25, when extending
the proof of the two-party setting, the two cases (in the proof) correspond to whether
or not Party 1 is honest (resp., 1 ∈ I or 1 ∈ I ). Again, we discard the case I = ∅,
where here the justification is that the oracle-aided protocol does not use the broadcast
channel at all (and so no information is available to the real-model adversary in this
case). The case 1 ∈ I = ∅ is handled exactly as the case that Party 1 is honest in the
proof of Proposition 7.4.15 (i.e., B sends the βi ’s it holds to the trusted party, obtains
h(α) (either explicitly or implicitly) and f (α), where α is the input of Party 1, and
uses (h(α), f (α)) to emulate the real execution). In case 1 ∈ I , we need to extend the
two-party treatment a little, because we also have to emulate the oracle answer given (in
Step C1) to dishonest parties (different from Party 1, which gets no answer). However,
this answer is determined by the query α
made in Step C1 by Party 1, and indeed, we
merely need to feed A with the corresponding oracle answer (h(α
), f (α
)). The rest of
the treatment is exactly as in the two-party case. The proposition follows.

Comment: Pure Oracle-Aided Protocols. Note that Construction 7.5.26 makes no

direct use of its communication channel, but is rather confined to the invocation of
oracles and local computations. Such an oracle-aided protocol is called pure. Note that
most oracle-aided protocols presented in Section 7.4 are pure. An important property
of pure oracle-aided protocols is that an adversary that controls none of the parties
and only wire-taps the communication channel gets no information, and so this case
721
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

can be discarded (as done in the proof of Proposition 7.5.27).73 In fact, Construc-
tion 7.5.24 is also a pure oracle-aided protocol (by virtue of its use of the secret broadcast
functionality).

7.5.4.4. Multi-Party Augmented Coin-Tossing

In this section, we generalize the augmented coin-tossing functionality (of Sec-
tion 7.4.3.5) to the multi-party setting. More generally, for any positive polynomial
 : N → N and a polynomial-time computable function g, we consider the randomized
m-ary functionality

(1n , ..., 1n ) → (r, g(r ), ..., g(r )), (7.52)

where r is uniformly distributed in {0, 1}(n) . We securely reduce Eq. (7.52) to the multi-
party authenticated-computation functionality. We note that the following construction
is different from the one used in the two-party case:

Construction 7.5.28 (an oracle-aided protocol for Eq. (7.52)): Let C be a commitment
scheme and C r1 ,...,r (σ1 , ..., σ ) = (Cr1 (σ1 ), ..., Cr (σ )) be as in Construction 7.4.16.
def
Inputs: Each party gets input 1n and sets  = (n).
Step C1: For i = 1, .., m, Party i uniformly selects ri ∈ {0, 1} and si ∈ {0, 1}·n .
Step C2: For i = 1, .., m, Party i uses the image-transmission functionality to send
def
ci = C si (ri ) to all parties. Actually, Party i enters Eq. (7.50) with input (ri , si ); each
def
other party enters with input 1+·n , which is supposed to equal h (C2) (ri , si ) = 1|ri |+|si | ,
def
and is supposed to obtain f (C2) (ri , si ) = C si (ri ). Abusing notation, let us denote by ci
the answer received by each party, where ci may equal ⊥ in case Party i has aborted
the i-th oracle call. Thus, in Steps C1–C2, each party commits to a random string.
Without loss of generality, we assume that no party aborts these steps (i.e., we treat
abort as if it were some legitimate default action).
Step C3: For i = 2, .., m (but not for i = 1), Party i uses the authenticated-
computation functionality to send ri to all parties. That is, Party i enters Eq. (7.50)
with input (ri , si ); each other party enters with input ci , where ci is supposed to equal
def def
h (C3) (ri , si ) = C si (ri ), and is supposed to obtain f (C3) (ri , si ) = ri . If Party i aborts the
oracle call (that it has invoked) or some Party j obtains an answer of a different
format, which happens in case the inputs of these two parties do not match, then
Party j halts with output ⊥. Otherwise, Party j obtains f (C3) (ri , si ) = ri and sets
j j def
ri = ri . (For simplicity, let r j = r j .)
Thus, in this step, each party (except Party 1), reveals the -bit long string to which
it has committed in Step C2. The correctness of the revealed value is guaranteed
by the definition of the authenticated-computation functionality, which is used here

73 Recall that in Section 7.4 we did not consider such external adversaries, and thus the notion of pure oracle-aided
protocols was neither discussed nor used.

722

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

instead of the straightforward way of disclosing the decommitment information. It

j
follows that for every j ∈ [m], if Party j is honest and did not halt, then ri = ri for
every i ∈ [m] \ {1}, where ri is the value committed by Party i in Step C3.
Step C4: In case Party 1 did not halt (and so ri1 = ri for every i ∈ [m]), Party 1 uses
m
the authenticated-computation functionality to send g(⊕i=1 ri1 ) to all parties. Details
follow:
def j
For j = 1, ..., m, Party j sets r j = ⊕i=2 m
ri . Note that in case Party j did not halt,
j
it holds that ri = ri (for every i ∈ [m]), and so r j = ⊕i=2 m
ri = r 1 . Thus, ⊕i=1
m
ri1 =
r1 ⊕ r . 1
def
Party 1 sets r = r1 ⊕ r 1 and uses the authenticated-computation functional-
ity to send g(r ) to all parties. Specifically, Party 1 enters Eq. (7.50) with in-
put (r1 , s1 , r 1 ); each (other) Party j enters with input (c1 , r j ), where (c1 , r j )
def
is supposed to equal h (C4) (r1 , s1 , r 1 ) = (C s1 (r1 ), r 1 ), and is supposed to obtain
def
f (C4) (r1 , s1 , r 1 ) = g(r1 ⊕ r 1 ), which equals g(r ). In case Party 1 aborts or Party j ob-
tains an answer of a different format, which happens if the inputs to the functionality
do not match, Party j halts with output ⊥ (indicating that Party 1 misbehaved).
Outputs: Unless halted in Step C3 (with output ⊥), Party 1 outputs r , and Party j = 1
outputs the value determined in Step C4, which is either g(r ) or ⊥.
In case m = 2, Construction 7.5.28 yields an alternative protocol for Eq. (7.34), that
is, a protocol that is fundamentally different from the one in Construction 7.4.16.

Proposition 7.5.29: Construction 7.5.28 securely reduces Eq. (7.52) to Eq. (7.50).

Proof Sketch:74 We transform any real-model adversary A (for the oracle-aided exe-
cution) into a corresponding ideal-model adversary B. The operation of B depends on
whether or not Party 1 is honest (i.e., 1 ∈ I¯), and we ignore the trivial cases of I = ∅
and I = [m]. In case 1 ∈ I¯ (i.e., Party 1 is honest), machine B proceeds as follows:

1. Machine B emulates the local actions of the honest parties in Step C1. In particular,
it uniformly selects (ri , si ) for each i ∈ I¯ (including i = 1).
2. For every i ∈ I¯, machine B emulates the i-th sub-step of Step C2 by feeding A with
the corresponding ci = C si (ri ) (as if it were the answer of the i-th oracle call). For
every i ∈ I , machine B obtains the input (ri , si ) that A enters (on behalf of Party i)
to the i-th oracle call of Step C2, and feeds A with adequate emulations of the oracle
answers (to other parties in I ).
3. For every i ∈ I¯ \ {1}, machine B emulates the i-th sub-step of Step C3 by feeding A
with a sequence in {ri , (ci , ri )}|I | that corresponds to whether or not each Party j ∈ I
has entered the input ci (defined in Step 2). For every i ∈ I , machine B obtains the
input (ri
, si
) that A enters (on behalf of Party i) to the i-th oracle call of Step C3,
records whether or not C si (ri ) = C si
(ri
), and feeds A with adequate emulations of
the oracle answers.

74 As in the proof of Proposition 7.5.25, we sometimes present ideal-model adversaries that halt before invoking
the trusted party. This can be viewed as invoking the trusted party with a special abort symbol.

723
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

For every i ∈ I¯, machine B sets ri1 = ri . For every i ∈ I , machine B sets ri1 = ri if
C si (ri ) = C si
(ri
) and aborts otherwise (while outputting whatever A outputs [when
Party 1 halts in Step C3]). Note that for every i, this setting of ri1 agrees with the
setting of ri1 in the protocol. In particular, B aborts if and only if (the honest) Party 1
would have halted in the corresponding (emulated) execution of Step C3.75
4. In case B did not abort, it invokes the trusted party with input 1n and obtains the
answer g(r ), where r is the uniformly distributed -bit string handed to Party 1.
Next, machine B emulates Step C4 by feeding each dishonest party with either
def
g(r ) or ((c1 , r 1 ), g(r )), where r 1 = ⊕i=2
m
ri1 . The choice is determined by whether
or not (in Step C4) this party has entered the input (c1 , r 1 ). (Note that we cheat in
the emulation of the oracle answer in Step C4; specifically, we use g(r ) rather than
g(r1 ⊕ r 1 ).) Finally, machine B outputs whatever A does.
We stress that in this case (i.e., 1 ∈ I ), machine B may possibly abort only before
invoking the trusted party (which satisfies the security definition). Observe that the
only difference between the ideal-model execution, under B and the real-model exe-
cution under A is that in the ideal-model execution, an independently and uniformly
distributed r ∈ {0, 1} is used (in the emulation of Step C4), whereas in the real-model
execution, r (as used in Step C4) is set to ⊕i=1
m
ri1 = r1 ⊕ r 1 . That is, in the ideal-model,
r1 is independent of r and r , whereas in the real-model, r1 = r ⊕ r 1 , where g(r ) and
1

r 1 = r i (for every i) are known to the adversary (and r appears in the joint-view). Thus,
in addition to its possible affect on r (in the real model), the only (other) affect that r 1
has on the joint-view is that the latter contains c1 = C(r1 ). In other words, (the joint-
views in) the real model and the ideal model differ only in whether c1 is a commitment
to r ⊕ r 1 or to a uniformly and independently distributed string, where r and r 1 are
explicit in the joint-view. By the hiding property of C, this difference is undetectable.
We now turn to the case that 1 ∈ I (i.e., Party 1 is dishonest). The treatment of this
case differs in two main aspects. First, unlike in the previous case, here the real-model
adversary (which controls Party 1) obtains all ri ’s, and so we must guarantee that in the
ideal-model execution, the trusted party’s answer (to Party 1) equals ⊕i=1 m
ri . Second,
unlike in the previous case, here the real-model adversary may effectively abort Step C4
(i.e., abort after obtaining the outcome), but this is easy to handle using the ideal-
model adversary’s ability to instruct the trusted party not to respond the honest parties.
Returning to the first issue, we present a different way of emulating the real-model
execution.76 Specifically, we will cheat in our emulation of the honest parties and use (in
Step 1–2) commitments to the value 0 , rather than commitments to the corresponding
ri ’s, which will be determined only at the end of Step 2. Details follow:
1. Machine B starts by invoking the trusted party and obtains a uniformly distributed
r ∈ {0, 1} . At this time, B does not decide whether or not to allow the trusted party
to answer the honest parties.

75 Indeed, in Step C3, Party 1 halts if and only if for some i, the input that Party 1 enters to the i-th sub-step (which
in turn equals the value ci = C si (ri ) that Party 1 has obtained in the i-th sub-step of Step C2) does not fit the
input (ri
, si
) that is entered by Party i (i.e., iff ci = C s
(ri
)).
i
76 We comment that the alternative emulation strategy can also be used in case Party 1 is honest.

724

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

In addition, B emulates the local actions of the honest parties in Step C1 by uniformly
selecting only the si ’s, for each i ∈ I¯.
2. For every i ∈ I¯, machine B emulates the i-th sub-step of Step C2 by feeding A with
ci = C si (0 ). For every i ∈ I , machine B obtains the input (ri , si ) that A enters (on
behalf of Party i) to the i-th oracle call of Step C2. Finally, B uniformly selects
all other ri ’s (i.e., for i’s in I¯) such that ⊕i=1
m
ri = r holds; for example, for each

i ∈ I \ {1}, select ri ∈ {0, 1} uniformly, and set r1 = r ⊕ (⊕i=2
¯ m
ri ).
3. For every i ∈ I , machine B emulates the i-th sub-step of Step C3 by feeding A with
¯
a sequence in {ri , (ci , ri )}|I | that corresponds to whether or not each Party j ∈ I has
entered the input ci . Note that the fact that ci is unlikely to be a commitment to ri is
irrelevant here. The rest of this step (i.e., the determination of the ri1 ’s) is as in the
case that Party 1 is honest. In particular, we let B halt if some Party i ∈ I behaves
improperly (i.e., invokes the corresponding oracle with input that does not fit ci as
recorded in the emulation of Step C2).
The next step is performed only in case B did not abort. In this case, ri1 = ri holds
for every i = 2, ..., m, and r = r1 ⊕ (⊕i=2 m
ri1 ) follows.
4. Next, machine B emulates Step C4 and determines whether or not A instructs Party 1
to abort its oracle call (in Step C4). The decision is based on whether or not the oracle
query (q1 , q2 , q3 ) of Party 1 (in Step C4) matches the oracle query (ri , si ) it made
in Step C2 and the value of ⊕i=2 m
ri1 as determined in Step 3 (i.e., whether or not
C q2 (q1 ) = C si (ri ) and q3 = ⊕i=2ri1 ). If Party 1 aborts, then B prevents the trusted
m

party from answering the honest parties, and otherwise B allows the trusted party to
answer. (Indeed, in case the trusted party answers Party i = 1, the answer is g(r )). In
addition, B emulates the answers of the Step C4 oracle call to the dishonest parties
(as in the case that Party 1 is honest). Finally, machine B outputs whatever A does.
Observe that the only difference between of the ideal-model execution under B and the
real-model execution under A is that in the former, commitments to 0 (rather than to
the ri ’s, for i ∈ I¯) are delivered in Step C2. However, by the hiding property of C, this
difference is undetectable.

An Important Special Case. An important special case of Eq. (7.52) is the case that
g(r, s) = C s (r ), where |s| = n · |r |. This special case will be called the augmented
(m-party) coin-tossing functionality. That is, for some fixed commitment scheme, C,
and a positive polynomial , we consider the m-ary functionality:

(1n , ..., 1n ) → ((r, s), C s (r ), ..., C s (r )) (7.53)

where (r, s) is uniformly distributed in {0, 1}(n) × {0, 1}(n)·n . Combining Proposi-
tions 7.5.27 and 7.5.29, we get:

Proposition 7.5.30: Assuming the existence of trapdoor permutations, the augmented

coin-tossing functionality of Eq. (7.53) can be securely implemented in the first mali-
cious model.

725
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

7.5.4.5. Multi-Party Input Commitment

The last functionality needed for the first multi-party compiler is a multi-party gener-
alization of the input-commitment functionality of Section 7.4.3.6. Specifically, for C
and C as in Section 7.5.4.4, we consider the m-party input-commitment functionality

(x, 1|x| , ..., 1|x| ) → (r, C r (x), ..., C r (x)), (7.54)

where r is uniformly distributed in {0, 1}|x| . By combining a straightforward general-

2

ization of Construction 7.4.20 with Propositions 7.5.27 and 7.5.30, we get:

Proposition 7.5.31: Assuming the existence of trapdoor permutations, the input-

commitment functionality of Eq. (7.54) can be securely implemented in the first mali-
cious model.

Proof Sketch: Starting from Construction 7.4.20, we replace each oracle call to a two-
party functionality by a call to the corresponding multi-party functionality. That is, in
Step C2 Party 1 uses the image-transmission (or rather the authenticated-computation)
def
functionality to send c
= C r
(x) to all other parties, in Step C3 an augmented coin-
tossing is used to provide Party 1 with a random pair (r, r

) whereas each other party gets
def
c

= C r

(r ), and in Step C4 Party 1 uses the authenticated-computation functionality
to send C r (x) to all other parties. Each of the other parties acts exactly as Party 2 acts
in Construction 7.4.20.
The security of the resulting multi-party oracle-aided protocol is established as in the
two-party case (treated in Proposition 7.4.21). As in the previous analysis of multi-party
protocols that generalize two-party ones, the two cases here are according to whether
or not Party 1 is honest (resp., 1 ∈ I or 1 ∈ I ). Finally, composing the oracle-aided
protocol with secure implementations of the adequate multi-party functionalities (as
provided by Propositions 7.5.27 and 7.5.30), the proposition follows.

7.5.4.6. The Compiler Itself

We are now ready to present the first multi-party compiler. Given a multi-party protocol,

, for the semi-honest model, we want to generate an “equivalent” protocol

for the
first malicious model. Recall that the given protocol operates in a communication model
consisting of a single broadcast channel. The compiled protocol will operate in the
same communication model. As in the two-party case, we first present an oracle-aided
version of the compiled protocol (which will actually be a pure oracle-aided protocol,
and thus the communication model is actually irrelevant for discussing the oracle-
aided version of the compiled protocol). The compiled protocol is a generalization of
the one presented in Construction 7.4.23 (for m = 2), and the reader is referred there
for additional clarifications.

Construction 7.5.32 (The first multi-party compiler, oracle-aided version): Given an

m-party protocol,
, for the semi-honest model (using a single broadcast channel), the

726

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

compiler produces the following oracle-aided m-party protocol, denoted

, for the
first malicious model:

Inputs: Party i gets input x i ∈ {0, 1}n .

Input-Commitment Phase: Each of the parties commits to its input by using the input-
commitment functionality of Eq. (7.54). That is, for i = 1, ..., m, Party i invokes
Eq. (7.54), playing the role of the first party with input x i , and obtains the output ρ i ,
def
whereas each other party obtains γ i = C ρ i (x i ).
Coin-Generation Phase: The parties generate random-tapes for the emulation of
.
Each party obtains the random-tape to be held by it, whereas each other party
obtains a commitment to this value. This is done by invoking the augmented coin-
tossing functionality of Eq. (7.53). That is, for i = 1, ..., m, Party i invokes Eq. (7.53),
playing the role of the first party, and obtains the output (r i , ωi ), whereas each other
def
party obtains δ i = C ωi (r i ).
Protocol-Emulation Phase: The parties use the authenticated-computation function-
ality of Eq. (7.50) in order to emulate each step of protocol
. The party that
is supposed to send (i.e., broadcast) a message plays the role of the first party in
Eq. (7.50), and the other parties play the other roles. Suppose that the current mes-
sage in
is to be sent by Party j. Then the functions h, f and the inputs α, β2 , ..., βm ,
for the functionality of Eq. (7.50), are set as follows (analogously to their setting in
Construction 7.4.23):

r The string α is set to equal (α1 , α2 , α3 ), where α1 = (x j , ρ j ) is the query and

answer of Party j in the oracle call that it initiated in the input-commitment
phase, α2 = (r j , ω j ) is the answer that Party j obtained in the oracle call that
it initiated in the coin-generation phase, and α3 is the sequence of messages that
def
Party j obtained so far in the emulation of
. Each βi equals β = (γ j , δ j , α3 ),
where γ and δ are the answers that the other parties obtained in the same oracle
j j

calls in the first two phases (and α3 is as previously).

Note that since
operates over a single broadcast channel, all parties receive
exactly the same messages.
r The function h is defined such that h((v1 , s1 ), (v2 , s2 ), v3 ) equals (C s (v1 ),
1
C s2 (v2 ), v3 ). Indeed, it holds that h(α1 , α2 , α3 ) = β.
r The function f equals the computation that determines the message to be
sent in
. Note that this message is computable in polynomial-time from
the party’s input (denoted x j and being part of α1 ), its random-tape (de-
noted r j and being part of α2 ), and the messages it has received so far (i.e.,
α3 ). Indeed, it holds that f (α1 , α2 , α3 ) is the message that Party j should
send in
.
Recall that each party that plays a receiver in the current oracle call obtains either
f (α) or (h(α), f (α)). It treats the second case as if the sending party has aborted,
which is also possible per se.

727
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Aborting: In case any of the functionalities invoked in any of the above phases termi-
nates in an abort state, the parties obtaining this indication abort the execution and
set their output to ⊥. Otherwise, outputs are as follows.
Outputs: At the end of the emulation phase, each party holds the corresponding output
of the party in protocol
. The party just locally outputs this value.
We note that both the compiler and the protocols produced by it are efficient, and that
their dependence on m is polynomially bounded.

7.5.4.7. Analysis of the Compiler

The effect of Construction 7.5.32 is analyzed analogously to the effect of Construc-
tion 7.4.23. In view of this similarity, we combine the two main steps (in the analysis)
and state only the end result:

Theorem 7.5.33 (Restating half of Theorem 7.5.15): Suppose that there exist collec-
tions of enhanced trapdoor permutations. Then any m-ary functionality can be securely
computable in the first malicious model (using only point-to-point communication
lines), provided that a public-key infrastructure exists in the network. Furthermore,
security holds even if the adversary can read all communication among honest parties.

Proof Sketch: We start by noting that the definition of the augmented semi-honest
model (i.e., Definition 7.4.24) applies without any change to the multi-party context
(also in case the communication is via a single broadcast channel). Recall that the
augmented semi-honest model allows parties to enter the protocol with modified inputs
(rather than the original ones) and abort the execution at any point in time. We stress
that in the multi-party augmented semi-honest model, an adversary controls all non-
honest parties and coordinates their input modifications and abort decisions. As in the
two-party case, other than these non-proper actions, the non-honest parties follow the
protocol (as in the semi-honest model).
The first significant part of the proof is showing that the compiler of Construc-
tion 7.5.32 transforms any protocol
into a protocol

such that executions of


in the first malicious real model can be emulated by executions of
in the aug-
mented semi-honest model (over a single broadcast channel). This part is analogous
to Proposition 7.4.25, and its proof is analogous to the proof presented in the two-
party case. That is, we transform any real-model adversary ( A, I ) for

into a cor-
responding augmented semi-honest adversary, (B, I ), for
. The construction of B
out of A in analogous to the construction of Bmal out of Amal (carried out in the proof
of Proposition 7.4.25): Specifically, B modifies inputs according to the queries that
A makes in the input-committing phase, uniformly selects random-tapes (in accor-
dance with the coin-generation phase), and aborts in case the emulated machine does
so. Thus, B, which is an augmented semi-honest adversary, emulates the malicious
adversary A.
The second significant part of the proof is showing that canonical protocols (as
provided by Theorem 7.5.14) have the property that their execution in the augmented
728

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

semi-honest model can be emulated in the (first) malicious ideal model of Defini-
tion 7.5.2. This part is analogous to Proposition 7.4.27, and its proof is analogous to
the proof presented in the two-party case.
Thus, given any m-ary functionality f , we first (use Theorem 7.5.14 to) obtain a
canonical protocol
that privately computes f . (Actually, we use the version of

that operates over a single broadcast channel, as provided by the pre-compiler [i.e.,
Proposition 7.5.16].) Combining the two parts, we conclude that when feeding
to
the compiler of Construction 7.5.32, the result is an oracle-aided protocol

such that
executions of

in the (first) malicious real model can be emulated in the ideal model
of Definition 7.5.2. Thus,

securely computes f in the first malicious model.
We are almost done, but there are two relatively minor issues to address. First,


is an oracle-aided protocol rather than an ordinary one. However, an ordinary protocol
that securely computes f can be derived by using secure implementations of the oracles
used by

(as provided by Propositions 7.5.27, 7.5.30, and 7.5.31). Second,

operates
in the broadcast-channel communication model, whereas we claimed a protocol in the
point-to-point communication model. This gap is bridged by using the post-compiler
(i.e., Proposition 7.5.19).

7.5.5. The Second Compiler: Effectively Preventing Abort

We now show how to transform any protocol for securely computing some functionality
in the first malicious model into a protocol that securely computes the same functionality
in the second malicious model. We stress that again, all communication, both in the
input protocol as well as in the one resulting from the compilation, is conducted by
posting messages on a single broadcast channel.
The current compiler has little to do with anything done in the two-party case.
The only similarity is at a technical level; that is, in using a secure implementation of
the authenticated-computation functionality. The main novelty is in the use of a new
ingredient, called Verifiable Secret Sharing (VSS).
It is interesting to note that we use implementations of the authenticated-computation
functionality (of Eq. (7.50)) and of VSS that are (“only”) secure in the first malicious
model. It is what we add on top of these implementations that makes the resulting
protocol secure in the second malicious model. Following is a high-level description of
the multi-party protocols generated by the current compiler. Recall that the input to the
compiler is a protocol secure in the first malicious model (and so the random-tape and
actions mentioned here refer to this protocol).77

The sharing phase: Each party shares its input and random-tape with all the parties
such that any strict majority of parties can retrieve their value, whereas no minority
group can obtain any knowledge of these values. This is done by using Verifiable
Secret Sharing (VSS).

77 In our application, we feed the current compiler with a protocol generated by the first compiler. Still, the random-
tape and protocol actions mentioned here refer to the secure protocol compiled by the first compiler, not to the
semi-honest protocol from which it was derived.

729
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Intuitively, the malicious parties (which are in a strict minority) are effectively pre-
vented from aborting the protocol by the following conventions:
r If a party aborts the execution prior to completion of the sharing phase, then the
honest parties (which are in the majority) will set its input and random-tape to
some default value and will carry out the execution (“on its behalf”).
r If a party aborts the execution after the completion of the sharing phase, then
the honest (majority) parties will reconstruct its input and random-tape and will
carry out the execution (“on its behalf”). The ability of the majority parties to
reconstruct the party’s input and random-tape relies on the properties of VSS.
The fact that communication is conducted over a broadcast channel, as well as the
abovementioned conventions, guarantee that the (honest) majority parties will always
be in consensus as to which parties have aborted (and what messages were sent).
Protocol-emulation phase: The parties emulate the execution of the original protocol
with respect to the input and random-tapes shared in the first phase. This will be done
using a secure (in the first malicious model) implementation of the authenticated-
computation functionality of Eq. (7.50).
We start by defining and implementing the only new tool needed; that is, Verifiable
Secret Sharing.

7.5.5.1. Verifiable Secret Sharing

Loosely speaking, a Verifiable Secret Sharing scheme is (merely) a secure (in the first
malicious model) implementation of a secret-sharing functionality. Thus, we first define
the latter functionality.

Definition 7.5.34 (Secret-Sharing Schemes): Let t ≤ m be positive integers. A t-out-

of-m Secret-Sharing Scheme is a pair of algorithms, G m,t and Rm,t , satisfying the
following conditions:78
Syntax: The share-generation algorithm, G m,t , is a probabilistic mapping of secret bits
to m-sequences of shares; that is, for every σ ∈ {0, 1}, the random variable G m,t (σ )
is distributed over ({0, 1}∗ )m . The recovering algorithm, Rm,t , maps t-sequences of
def
pairs in [m] × {0, 1}∗ into a single bit, where [m] = {1, ..., m}.
The recovery condition: For any σ ∈ {0, 1}, any sequence (s1 , ..., sm ) in the range of
G m,t (σ ), and any t-subset {i 1 , ..., i t } ⊆ [m], it holds that
Rm,t ((i 1 , si1 ), ..., (i t , sit )) = σ
The secrecy condition: For any (t − 1)-subset I ⊂ [m], the distribution of the I -
components of G m,t (σ ) is independent of σ . That is, for any I = {i 1 , ..., i t−1 } ⊂ [m],
let g I (σ ) be defined to equal ((i1 , si1 ), ..., (i t−1 , sit−1 )), where (s1 , ..., sm ) ← G m,t (σ ).

78 At this point, we place no computational requirements on G m,t and Rm,t . Typically, when m is treated as a
parameter, these algorithms will operate in time that is polynomial in m.

730

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Then we require that for any such I , the random variables g I (0) and g I (1) are
identically distributed.
Indeed, an m-out-of-m secret-sharing scheme is implicit in the construction presented
in Section 7.5.2: To share a bit σ , one just generates m random bits that sum up to
σ (mod 2). Efficient t-out-of-m secret-sharing schemes do exist for any value of t ≤ m.
The most popular one, which uses low-degree polynomials over finite fields, is presented
next.

Construction 7.5.35 (Shamir’s t-out-of-m secret-sharing scheme): Find the smallest

prime number, denoted p, that is bigger than m, and consider arithmetic over the finite
field GF( p).79 The share-generating algorithm consists of uniformly selecting a degree
t − 1 polynomial over GF( p) with free term equal to σ , and setting the i-th share
to be the value of this polynomial at i. The recovering algorithm consists of finding
(by interpolation) the unique degree t − 1 polynomial that fits the given values and
outputting its free term.

Construction 7.5.35 is analyzed in Exercise 17. Getting back to our subject matter, we
derive the basic definition of verifiable secret sharing.

Definition 7.5.36 (Verifiable Secret Sharing, basic version): A verifiable secret shar-
ing (VSS) scheme with parameters (m, t) is an m-party protocol that implements
(i.e., securely computes in the first malicious model) the share-generation func-
tionality of some t-out-of-m secret-sharing scheme. That is, let G m,t be a share-
generation algorithm of some t-out-of-m secret-sharing scheme. Then the correspond-
ing share-generation functionality that the VSS securely computes (in the first malicious
model) is
((σ, 1n ), 1n , ..., 1n ) → G m,t (σ ) (7.55)

Actually, it will be more convenient to use an augmented notion of Verifiable Secret

Sharing. The augmentation provides each party with an auxiliary input that determines
the secret σ (as in a commitment scheme), and allows Party 1 to later conduct au-
thenticated computations relative to the auxiliary inputs given to the other parties.
Furthermore, each party is provided with a certificate of the validity of its own share
(relative to the auxiliary inputs given to the other parties). We seize the opportunity to
generalize the definition, such that it refers to the sharing of strings (of a priori known
length), rather than to the sharing of single bits. From this point on, when we say Veri-
fiable Secret Sharing (or VSS), we mean the notion defined next (rather the the weaker
form in Definition 7.5.36).

Definition 7.5.37 (Verifiable Secret Sharing, revised): Given a share-generation al-

gorithm G m,t of some t-out-of-m secret-sharing scheme, we extend it to handle

79 By the Fundamental Theorem of Number Theory, p ≤ 2m. Thus, p can be found by merely (brute-force)
factoring all integers between m + 1 and 2m.

731
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

def
n-bit long strings; that is, G m,t (σ1 , ..., σn ) = (s1 , ..., sm ), where si = si,1 · · · si,n
and (s1, j , ..., sm, j ) ← G m,t (σ j ) for every i = 1, ..., m and j = 1, ..., n. Suppose
that G m,t (α) ∈ ({0, 1}(|α|) )m , and let C be a commitment scheme and C be as
in Construction 7.5.28. Consider the corresponding (augmented) share-generation
functionality

(α, 1|α| , ..., 1|α| ) → ((s, ρ), (s2 , ρ2 , c), ..., (sm , ρm , c)) (7.56)
def
where s = (s1 , ..., sm ) ← G m,t (α), (7.57)
def 2
ρ = (ρ1 , ..., ρm ) ∈ {0, 1}m·(|α|)
(7.58)
is uniformly distributed,
def
and c = (C ρ1 (s1 ), ..., C ρm (sm )). (7.59)

Then any m-party protocol that securely computes Eq. (7.56) – (7.59) in the first ma-
licious model is called a verifiable secret sharing (VSS) scheme with parameters
(m, t).

Observe that each party may demonstrate (to each other party) the validity of its
“primary” share (i.e., the si ) with respect to the globally held c, by revealing the
corresponding ρi . We shall be particularly interested in VSS schemes with param-
eters (m, m/2); that is, t = m/2. The reason for this focus is that we assume
throughout this section that the malicious parties are in strict minority. Thus, by the
secrecy requirement, setting t ≥ m/2 guarantees that the (less than m/2) dishonest
parties are not able to obtain any information about the secret from their shares. On
the other hand, by the recovery requirement, setting t ≤ m/2 guarantees that the
(more than m/2) honest parties are able to efficiently recover the secret from their
shares. Thus, in the sequel, whenever we mention VSS without specifying the param-
eters, we mean the VSS with parameters (m, m/2), where m is understood from the
context.
Clearly, by Theorem 7.5.33, verifiable secret sharing schemes exist, provided that
enhanced trapdoor permutations exist. Actually, to establish the existence of VSS, we
merely need to apply the first compiler to the straightforward protocol that privately
computes Eq. (7.56) – (7.59); see Exercise 10. For the sake of subsequent reference, we
state the latter result.

Proposition 7.5.38: Suppose that trapdoor permutations exist. Then for every t ≤ m,
there exists a verifiable secret-sharing scheme with parameters (m, t).

Note that the assumption used in Proposition 7.5.38 is (merely) the one needed for the
operation of the first compiler, which amounts to the assumption needed for imple-
menting the functionalities used in Construction 7.5.32.

732

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

7.5.5.2. The Compiler Itself

We are now ready to present the second compiler. Recall that we are given a multi-
party protocol,
, that is secure in the first malicious model, and we want to generate
an “equivalent” protocol

for the second malicious model. Also recall that both the
given protocol and the one generated by the compiler operate in a communication model
consisting of a single broadcast channel. Finally, we note that the generated protocol
uses sub-protocols that are secure with respect to the first malicious model (and yet the
entire protocol will be analyzed with respect to the second malicious model).80

def
Construction 7.5.39 (The second multi-party compiler): Let t = m/2. Given an m-
party protocol,
, for the first malicious model, the compiler produces the following
m-party protocol, denoted

, for the second malicious model.
Inputs: Party i gets input x i ∈ {0, 1}n .
Random-Tape: Party i uniformly selects a random-tape, denoted r i ∈ {0, 1}c(n) , for
the emulation of
.
The Sharing Phase: Each party shares its input and random-tape with all the parties,
using a Verifiable Secret Sharing scheme. That is, for i = 1, ..., m, Party i invokes
the VSS scheme playing the first party with input x i r i , while the other parties play
the roles of the other parties in Eq. (7.56) – (7.59) with input 1n+c(n) .
Regarding the i-th VSS invocation,81 we denote the output that Party i ob-
tains by (s i , ρ i ), and the outputs that each other Party j obtains by (s ij , ρ ij , ci ),
where s i = (s1i , ..., smi ) ← G m,t (x i r i ), ρ i = (ρ1i , ..., ρmi ) is uniformly distributed,
ci = (c1i , ..., cmi ), and cki = C ρki (ski ). Note that either all honest parties get the correct
outcome or they all detect that Party i is cheating and set their outcome to ⊥.
Handling Abort: If Party i aborts the i-th VSS invocation, which means that all honest
parties received the outcome ⊥, then the honest parties set its input and random-tape
to some default value; that is, they set their record of the input and random-tape of
Party i (which are otherwise unknown to them) to some default value. Note that by
definition, the VSS scheme is secure in the first malicious model, and thus all honest
parties agree on whether or not the VSS initiator (i.e., Party i) has aborted.82

80 For this reason, we cannot utilize a composition theorem for the second malicious model. We comment that such
a composition theorem would anyhow be more restricted than Theorem 7.5.21. One issue is that the second mali-
cious model depends on a bound on the fraction of dishonest parties. Thus, if the m-party oracle-aided protocol in-
vokes a k-ary functionality with k < m, then the bound (on the fraction of dishonest parties) may be violated in the
sub-protocol that replaces the latter. For this reason, when dealing with the second malicious model, one should
confine the treatment to m-party oracle-aided protocols that use m-ary (rather than k-ary) functionalities.
81 Indeed, this notation is slightly inconsistent with the one used in Definition 7.5.37. Here, Party i plays the first

party in the VSS, and being consistent with Definition 7.5.37 would required calling its share s1i rather than sii .
Consequently, the share of Party j in this invocation would have been denoted sπi i ( j) , where πi ( j) is the role that
Party j plays in this invocation. However, such notation would have made our exposition more cumbersome.
82 This is reflected in the corresponding ideal-model adversary that either makes all honest parties detect abort

(i.e., output ⊥) or allows all of them to obtain (and output) the corresponding entries in a valid m-sequence.

733
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

We stress that in case Party i aborts the i-th VSS invocation, its (default) input and
random-tape become known to all parties. Since the entire execution takes place
over a broadcast channel, each party can determine by itself what messages Party i
should send in a corresponding execution of
. Thus, there is actually no need to
send actual messages on behalf of Party i.
Protocol-Emulation Phase: The parties emulate the execution of protocol
with re-
spect to the input and random-tapes shared in the first phase. This will be done by
using a secure (in the first malicious model) implementation of the authenticated-
computation functionality of Eq. (7.50).

That is, Party i, which is supposed to send a message in
, plays the role of the first
party in Eq. (7.50), and the other parties play the other roles. The inputs α, β2 , ..., βm
and the functions h, f , for the functionality of Eq. (7.50), are set as follows:
r The string α = (α1 , α2 ) is set such that α1 = (x i r i , s i , ρ i ) and α2 equals the
concatenation of all previous messages sent in the emulation of previous steps
of
. Recall that (x i r i , (s i , ρ i )) is the input–output pair of Party i in the i-th
invocation of the VSS.
r The string β j equals β def = (ci , α2 ), where α2 is as in previous item. Recall that ci
is part of the output that each other party got in the i-th invocation of the VSS.
r The function h is defined such that h((z, (s1 , ..., sm ), (r1 , ..., rm )), γ ) =
((C r1 (s1 ), ..., C rm (sm )), γ ). Indeed, h(α1 , α2 ) = β.
r The function f is set to be the computation that determines the message to be sent
in
. Note that this message is computable in polynomial-time from the party’s
input (denoted x i ), its random-tape (denoted r i ), and the previous messages posted
so far (i.e., α2 ).
As a result of the execution of the authenticated-computation sub-protocol, each
party either gets an indication that Party i aborted or determines the message that
Party i should have sent in a corresponding execution of
. By the definition of
security in the first malicious model, all honest parties agree on whether or not
Party i aborted, and in case it did not abort, they also agree on the message it sent.
Handling Abort: If a party aborts when playing the role of the first party in an
invocation of Eq. (7.50) during the emulation phase, then the majority parties re-
cover its (actual) input and random-tape, and carry out the execution on its behalf.
Specifically, if Party j detects that Party i has aborted, then it broadcasts the pair
(s ij , ρ ij ) that it has obtained in the sharing phase, and each party uses the correctly
decomitted shares (i.e., the s ij ’s) to reconstruct x i r i .

We note that the completion of the sharing phase (and the definition of VSS) guarantee
that the majority parties hold shares that yield the input and random-tape of any party.
Furthermore, the correct shares are verifiable by each of the other parties, and so
reconstruction of the initial secret is efficiently implementable whenever a majority
of parties wishes to do so.

734

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

Outputs: At the end of the emulation phase, each party holds the corresponding output
of the party in protocol
. The party just locally outputs this value.
Note that the VSS scheme is implicitly used as a commitment scheme for the value
of x i r i ; that is, ci = (c1i , ..., cmi ) serves as a commitment to the sequence of shares
(s1i , ..., smi ), which in turn determine x i r i . Actually, the main steps in the emulation
phase only refer to this aspect of the VSS, whereas only the abort-handling procedure
refers to the additional aspects (e.g., the fact that Party j holds the value of the share s ij
that is determined by the commitment cij , as well as the corresponding decommitment
information).

Comment. Applying the two (multi-party protocol) compilers one after the other is
indeed wasteful. For example, we enforce proper emulation (via the authenticated-
computation functionality) twice: first with respect to the semi-honest protocol, and
next with respect to the protocol resulting from the first compiler. Indeed, more ef-
ficient protocols for the second malicious model could be derived by omitting the
authenticated-computation protocols generated by the first compiler (and having the
second compiler refer to the actions of the semi-honest protocol). Similarly, one can
omit the input-commit phase in the first compiler. In general, feeding the second com-
piler with protocols that are secure in the first malicious model is an overkill; see further
discussion subsequent to Proposition 7.5.42.

7.5.5.3. Analysis of the Compiler

Our aim is to establish the following:

Theorem 7.5.40 (Restating the second half of Theorem 7.5.15): Suppose that there
exist collections of enhanced trapdoor permutations. Then any m-ary functionality
can be securely computable in the second malicious model (using only point-to-point
communication lines), provided that a public-key infrastructure exists in the network.
Furthermore, security holds even if the adversary can read all communication among
honest parties.

As will be shown here, given a protocol as guaranteed by Theorem 7.5.33, the second
compiler produces a protocol that securely computes (in the second malicious model)
the same functionality. Thus, for any functionality f , the compiler transforms proto-
cols for securely computing f in the first malicious model into protocols for securely
computing f in the second malicious model. This suffices to establish Theorem 7.5.40,
yet it does not say what the compiler does when given an arbitrary protocol (i.e., one
not provided by Theorem 7.5.33). In order to analyze the action of the second compiler,
in general, we introduce the following model that is a hybrid of the semi-honest and
the two malicious models. We call this new model the second-augmented semi-honest
model. Unlike the (first) augmented semi-honest model (used in the analysis of the first
compiler [see proof of Theorem 7.5.33]), the new model allows a dishonest party to
select its random-tape arbitrarily, but does not allow it to abort.

735
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Definition 7.5.41 (the second-augmented semi-honest model): Let
be a multi-party

protocol. A coordinated strategy for parties I is admissible as a second-augmented
semi-honest behavior (with respect to
) if the following holds:

Entering the execution: Depending on their initial inputs and in coordination with each
other, the parties in I may enter the execution of
with any input of their choice.
Selection of random-tape: Depending on their inputs and in coordination with each
other, the parties in I may arbitrarily select their random-tapes for the execution
of
.
Here and in the previous step, the parties in I may employ randomized procedures,
but the randomization in their procedures is not to be confused with the random-tapes
for
selected in the current step.
Proper message transmission: In each step of
, depending on its view so far, the
designated (by
) party sends a message as instructed by
. We stress that the
message is computed as
instructs based on the party’s (possibly modified) input,
its (possibly non-uniformly selected) random-tape, and the messages received so far,
where the input and random-tape are as set in the previous two steps.
Output: At the end of the interaction, the parties in I produce outputs depending on
their entire view of the interaction. We stress that the view contains their initial inputs
and all messages sent over all channels.83
Intuitively, the compiler transforms any protocol
into a protocol

so that executions
of

in the second malicious model correspond to executions of
in the second
augmented semi-honest model. That is:

Proposition 7.5.42 (general analysis of the second multi-party compiler): Let

be

the m-party protocol produced by the compiler of Construction 7.5.39, when given
the protocol
. Then for every probabilistic polynomial-time adversary A for the sec-
ond malicious model, there exists a probabilistic polynomial-time strategy B that is
admissible (with respect to
) in the second-augmented semi-honest model (of Defini-
tion 7.5.41), such that for every I ⊂ [m] with |I | < m/2
c
{real
, I, B(z) (x)}x,z ≡ {real

, I, A(z) (x)}x,z

Proposition 7.5.42 can be viewed as asserting that if
is secure in the second-augmented

semi-honest model, then

is secure in the second malicious model, where by the former
term we mean that for every real-model adversary B that is admissible (with respect to

) in the second-augmented semi-honest model, there exists an ideal-model adversary
(2) c
C as per Definition 7.5.4 such that {ideal f, I,C(z) (x)}x,z ≡ {real
, I, B(z) (x)}x,z (for every
I ). Proposition 7.5.42 will be applied to protocols that securely compute a functionality
in the first malicious model. As we shall see, for such protocols, the second augmented
semi-honest model (of Definition 7.5.41) can be emulated by the second ideal malicious

83 This model is applicable both when the communication is via a single broadcast channel and when the commu-
nication is via point-to-point channels that can be wire-tapped by the adversary.

736

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

model (of Definition 7.5.4). Thus, Theorem 7.5.40 will follow. We start by establishing
Proposition 7.5.42:
Proof Sketch: Given a real-model adversary A (for

), we present a corresponding
adversary B that is admissible with respect to
for the second augmented semi-honest
model. We stress two points. First, whereas A may abort some parties, the adversary B
may not do so (as per Definition 7.5.41). Second, we may assume that the number of
parties controlled by A (and thus by B) is less than m/2 (because nothing is required
otherwise).
Machine B will use A as well as the ideal-model adversaries derived (as per Def-
inition 7.5.3) from the behavior of A in the various sub-protocols invoked by

. We
stress that these ideal-model adversaries are of the first malicious model. Furthermore,
machine B will also emulate the behavior of the trusted party in these ideal-model em-
ulations (without communicating with any trusted party; there is no trusted party in the
augmented semi-honest model). Thus, the following description contains an implicit
special-purpose composition theorem (in which sub-protocols that are secure in the
first malicious model are used to implement the oracles of an oracle-aided protocol that
is secure in the second malicious model):
Entering the execution and selecting a random-tape: B invokes A (on the very input
supplied to it) and decides with what input and random-tape to enter the execution
of
. Toward this end, machine B emulates the execution of the sharing phase of


, using A (as subroutine). Machine B supplies A with the messages it expects to
see, thus emulating the honest parties in

, and obtains the messages sent by the
parties in I (i.e., those controlled by A). We stress that this activity is internal to B
and involves no real interaction (of B in
).
Specifically, B emulates the executions of the VSS protocol in an attempt to obtain
the values that the parties in I share with all parties. The emulation of each such
VSS-execution is done by using the ideal-model adversary derived from (the residual
real-model malicious adversary) A. We stress that in accordance with the definition
of VSS (i.e., security in the first malicious model), the ideal-model adversary derived
from (the residual) A is in the first malicious model and may abort some parties.
Note that (by Definitions 7.5.3 and 7.5.2) this may happen only if the initiator of
the VSS is dishonest. In case the execution initiated by some party aborts, its input
and random-tape are set to the default value (as in the corresponding abort-handling
procedure of

). Details follow:
r In an execution of VSS initiated by an honest party (i.e., in which an honest party
plays the role of the first party in VSS), machine B obtains the corresponding
augmented shares (available to I ).84 Machine B will use an arbitrary value, say
0n+c(n) , as the first party’s input for the current emulation of the VSS (because
the real value is unknown to B). In emulating the VSS, machine B will use
the ideal-model adversary, denoted A
, that emulates the behavior of A in this
VSS (in

), when given the history so far. We stress that since the initiating

84 These will be used in the emulation of future message-transmission steps.

737
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

party of the VSS is honest, this ideal-model adversary (i.e., A
) cannot abort any
party.
Invoking the ideal-model adversary A
, and emulating both the honest (ideal-
model) parties and the trusted party, machine B obtains the outputs of all parties
(i.e., and in particular, the output of the initiating party). That is, machine B emu-
lates the sharing of value 0n+c(n) by the initiating party and emulates the response
of the trusted oracle (i.e., by setting s ← G m,t (0n+c(n) ), uniformly selecting ρ of
adequate length, and computing the outputs as in Eq. (7.56) – (7.59)).
r In an execution of VSS initiated by a party in I (i.e., a dishonest party plays the
role of the first party in VSS), machine B obtains the corresponding input and
random-tape of the initiator, as well as the randomization used in the commitment
to it. As before, machine B uses the derived ideal-model adversary, denoted A
,
to emulate the execution of the VSS. Recall that A
emulates the behavior of A in
the corresponding execution of the VSS.
Suppose that we are currently emulating the instance of VSS initiated by Party i,
where i ∈ I . Then B invokes A
on input x i r i (i.e., the initial input and random-
tape of Party i), and emulating both the honest (ideal-model) parties and the
trusted party, machine B obtains the outputs of all parties (including the “VSS-
randomization” (i.e., (s i , ρ i )) handed to Party i which is in I ). A key point is
that machine B has obtained, while emulating the trusted party, the input handed
by A
to the trusted party. This value is recorded as the modified input and
random-tape of Party i .
In case the emulated machine did not abort the initiator (i.e., Party i), machine
B records the previous value, as well as the randomization used by B (as trusted
party) in the execution of VSS. Otherwise (i.e., A aborts Party i in the invocation
of VSS initiated by it), the input and random-tape of Party i are set to the default
value (as in

). In either case, B concatenates the emulation of the VSS to the
history of the execution of A.
Thus, inputs and random-tapes are determined for all parties in I , depending only
on their initial inputs. (All this is done before entering the actual execution of
.)
Furthermore, the view of machine A in the sharing phase of

has been emulated,
and the VSS-randomizations (i.e., the pairs (s i , ρ i )) used in the sharing of all values
have been recorded by B. (Actually, it suffices to record the VSS-randomization
handed to dishonest parties and the commitments made on behalf of honest ones;
these will be used in the emulation of the message-transmission steps of

,
where the VSS-randomization will be used only in case the corresponding party
aborts.)
Subsequent steps – message transmission: Machine B now enters the actual execution
of
(with inputs and random-tapes for I -parties as determined earlier). It proceeds
in this real execution of
, along with emulating the corresponding executions of
the authenticated computation of Eq. (7.50) (which are invoked in

).
In a message-transmission step by an honest party in
, machine B obtains a mes-
sage from this honest party (in the real execution of
) and emulates an execution
738

www.Ebook777.com
 7.5* EXTENSION TO THE MULTI-PARTY CASE

of the authenticated-computation protocol resulting in this message as output. In

a message-transmission step by a dishonest party in
, machine B computes the
message to be sent as instructed by
, based on the input and random-tape de-
termined in the previous stage and the messages obtained so far (in
). In addi-
tion, B emulates an execution of the authenticated-computation protocol resulting
in this message as output. The emulation of each execution of the authenticated-
computation protocol, which securely computes (in the first malicious model) the
functionality Eq. (7.50), is done by using the malicious ideal-model adversary de-
rived from A. The fact that in these emulations machine B also emulates the trusted
party allows it to set the outcome of the authenticated-computation protocol to fit
the message being delivered. We stress that the fact that a (dishonest) party may
abort some parties in these emulations of

does not result in aborting the real
execution of
(and is merely reflected in the transcript of these emulations). Details
follow:
r In a message-transmission step by an honest party in
, machine B first obtains
from this party (in the real execution of
) a message, denoted msg. This completes
all that is done in this step with respect to communication in
.
Next, machine B proceeds in emulating the corresponding message-transmission
sub-protocol of

. Firstly, machine B derives the ideal-model adversary, denoted
A
, which corresponds to the behavior of A in the corresponding execution of the
authenticated-computation sub-protocol (executed by protocol

). Invoking the
ideal-model adversary A
, and emulating both the honest (ideal-model) parties
and the trusted party, machine B sets the trusted party’s replies (to parties in I ) to
equal either msg or (β, msg), where β is as in Construction 7.5.39.85 The decision
concerning which reply to deliver to each party in I depends on the input that this
party hands to the trusted party (or rather the input that A
hands on its behalf):
If the party hands the correct value β, then it receives msg; otherwise it receives
(β, msg).
Note that the emulation of the authenticated-computation sub-protocol is carried
out so as to produce an output in {msg, (β, msg)}|I | , where msg does not neces-
sarily fit the output of the authenticated-computation functionality of Eq. (7.50)
on the corresponding dummy inputs. However, the machine A
used in the em-
ulation cannot detect that we are cheating because the inputs that A
gets (i.e.,
commitments to dummy values) are computationally indistinguishable from the
correct inputs (i.e., commitments to values that correspond to the unknown input
and random-tape of the corresponding honest party).
Finally, B concatenates the emulation of the authenticated-computation sub-
protocol to the history of the execution of A. (Note that since the initiator of
the authenticated-computation sub-protocol is honest, abort is not possible here,
by definition of the first ideal model.)

85 Recall that β = (ci , α2 ), where ci is the commitment produced by the VSS that was invoked by Party i, which
is assumed to be the sender in the current message-transmission step, and α2 equals the sequence of messages
sent so far in the emulated execution of
.

739
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

r In a message-transmission step by a dishonest party in
, machine B first com-

putes the message to be sent according to
. This message is computed based
on the input and random-tape determined (and recorded) in the emulation of the
sharing phase of

and the messages received so far (in the execution of
).
Denote the resulting message by msg. Machine B completes the execution of this
step in
by posting msg on the broadcast channel.
Next, machine B proceeds in emulating the corresponding authenticated-
computation sub-protocol of

. Firstly, machine B derives the ideal-model
adversary, denoted A
. Invoking A
and emulating both the honest (ideal-
model) parties and the trusted party, machine B produces an emulation of
the corresponding execution of the authenticated-computation protocol. The in-
put (for the transmitting party) used by B in this emulation equals the value
recorded in the emulation of (the corresponding VSS in) the sharing phase
of

.
Note that this emulation (of the corresponding authenticated-computation sub-
protocol of

) either produces the very same message msg (or a pair (β, msg)
as above) or aborts the sender. In the latter case, we emulate the abort-handling
procedure of

, by using the corresponding VSS-randomization (as recorded
in the sharing phase of

). In both cases, B concatenates the emulation of the
authenticated-computation protocol (and possibly also the abort-handling proce-
dure) to the history of the execution of A.
Note that each message-transmission step is implemented in polynomial-time, and
each message posted is computed exactly as instructed by
. (We stress again that
the emulation of an aborting event in

does not result in aborting the execution of
any party in
.)
Output: Machine B just outputs whatever machine A outputs given the execution history
composed (or actually emulated) as in the previous steps.
Clearly, machine B (described here) implements a second-augmented semi-honest be-
havior with respect to
. It is left to show that
c
{real

, I, A (x)}x ≡ {real
, I, B (x)}x (7.60)
There are two differences between the two ensembles referred to in Eq. (7.60):
1. In the first distribution (i.e., real

,(A, I ) (x)), secure (in the first malicious model)
protocols implementing VSS and authenticated computation (of Eq. (7.56) – (7.59)
and Eq. (7.50), respectively) are executed; whereas in the second distribution (i.e.,
real
,(B, I ) (x)), these executions are emulated using the corresponding ideal-model
adversaries.
2. The emulation of Eq. (7.50) in real
,(B, I ) (x) is performed with a potentially wrong
input, specifically, with commitments to dummy values, rather than to the correct
values.
However, these differences are computationally undetectable.

740

www.Ebook777.com
 7.6* PERFECT SECURITY IN THE PRIVATE CHANNEL MODEL

Proof of Theorem 7.5.40: Given an m-ary functionality f , let
be an m-party proto-

col, as guaranteed by Theorem 7.5.33, for securely computing f in the first malicious
model. (Actually, we merely need a protocol operating in the broadcast channel [rather
than point-to-point] communication model.) We now apply the compiler of Construc-
tion 7.5.39 to
and derive a protocol

. By Proposition 7.5.42, for any efficient
real-model adversary A (for

), there exists an efficient admissible behavior (with
respect to
) in the second-augmented semi-honest model, denoted B, such that for
every I ⊂ [m] with |I | < m/2
c
{real

, I, A (x)}x ≡ {real
, I, B (x)}x (7.61)
One key observation is that B constitutes a benign form of a real-model adversar-
ial behavior with respect to
(which is certainly allowed by the first malicious
model). Specifically, the malicious behavior of B amounts to replacing inputs and
random-tapes arbitrarily, and executing
with these replaced values and without
aborting any party. Thus, by the security of
(in the first malicious model), the
real-model adversary B can be emulated by an ideal-model adversary C that oper-
ates in the first ideal model (and so may potentially abort parties). However, since
B does not abort parties, then neither does C (except with negligible probabil-
ity). It follows that C is essentially an admissible ideal-model adversary for the
second malicious party, or, more accurately, C behaves in a way that is statisti-
cally close to a second ideal-model adversary C
(which behaves as C except that
it never aborts). Combining Eq. (7.61) with the latter observations, we obtain (for
every |I | < m/2)
c
{real

, I, A (x)}x ≡ {real
, I, B (x)}x
c (1)
≡ {ideal f, I,C (x)}x
s (2)
≡ {ideal f, I,C
(x)}x
We are almost done. The only problem is that

operates in the communication model
of a single broadcast channel. As in the proof of Theorem 7.5.33, this problem is
resolved by applying the post-compiler (i.e., Proposition 7.5.19).

7.6.* Perfect Security in the Private Channel Model

In this section, we present an alternative treatment of general secure multi-party

protocols. Specifically, we assume the existence of private channels between each
pair of parties and present protocols that are “perfectly secure” (i.e., perfectly em-
ulate a trusted party), and we do so without relying on any intractability assump-
tions. However, security holds only in case the honest parties are in a strict ma-
jority, and thus the current treatment is not meaningful for the two-party case. Let
us summarize the Pros and Cons of the current treatment in comparison to the
treatment offered in Section 7.5:

741
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

pros: Abstracting away computational issues. In particular:

1. Making no intractability assumptions.
2. Emulating a trusted party in a perfect sense (rather than in a computation-
ally indistinguishable sense), even with respect to computationally unbounded
adversaries.
cons: Limited applicability. In particular:
1. A strict majority of honest parties is required (even for withstanding semi-
honest adversaries). Thus, the current treatment is inapplicable to the two-
party case.
2. Perfectly private channels are postulated to exist.
Again, our ultimate goal is to design protocols that withstand any feasible adversarial
behavior, and again we proceed in two steps: first dealing with the semi-honest model
and next with the malicious model. However, here, protocols for the malicious model
are derived by extending the ideas that underlie the semi-honest protocols, rather than
by compiling the latter.

7.6.1. Definitions
We consider both the semi-honest and the malicious models, where in both cases
we refer to explicit bounds on the number of dishonest parties. Furthermore, in both
cases, we consider a communication network consisting of point-to-point channels
that cannot be wire-taped by the adversary. Finally, in both models, we require the
relevant probability ensembles to be statistically indistinguishable, rather than (only)
computationally indistinguishable.
Security in the Semi-Honest Model. The following definition is derived from Defi-
nition 7.5.1 by restricting the number of dishonest parties and strengthening the indis-
tinguishability requirement.

Definition 7.6.1 (t-privacy of m-party protocols): Let f be an m-ary functionality, and

be an m-party protocol for computing f . As in Definition 7.5.1, we denote the joint-
view of the parties in I ⊆ [m] by view
I (x), and the corresponding output sequence
of all parties by output
(x). We say that
t-privately computes f if there exists
a probabilistic polynomial-time algorithm, denoted S, such that for every I ⊂ [m] of
cardinality at most t, it holds that
{(S(I, x I , f I (x)) , f (x))}x∈({0,1}∗ )m
s
≡ {(view


I (x) , output (x))}x∈({0,1}∗ )m (7.62)
where x I and f I denote projections of the corresponding m-ary sequence on the coor-
dinates in I . In case the ensembles in Eq. (7.62) are identically distributed, we say that
the emulation is perfect.

We stress that Eq. (7.62) requires statistical indistinguishability, whereas the ana-
logue requirement in Definition 7.5.1 is of computational indistinguishability. As in
742

www.Ebook777.com
 7.6* PERFECT SECURITY IN THE PRIVATE CHANNEL MODEL

Definition 7.5.1, the view of parties in I does not include messages sent among parties
def
in I¯ = [m] \ I .

Security in the Malicious Model. Analogously, the following definition is derived

from Definition 7.5.4 by restricting the number of dishonest parties, disallowing wire-
tapping, and strengthening the indistinguishability requirement. Recall that Defini-
tion 7.5.4 refers to security in the second malicious model, which is reflected in the
choice of the ideal model. We further strengthen the definition by allowing computation-
ally unbounded real-model adversaries, and by requiring the corresponding ideal-model
adversaries to be of “comparable complexity” (i.e., have polynomially related running
time). Specifically, we say that algorithm B has comparable complexity to algorithm A
if there exists a polynomial p such that for every y, it holds that time B (y) ≤ p(time A (y)),
where time A (y) (resp., time B (y)) denotes the number of steps taken by A (resp., B) on
input y.
(2)
Definition 7.6.2 (t-security of m-party protocols): Let f ,
, and ideal f, I, B(z) (x) be
exactly as in Definition 7.5.4. The real-model adversary is defined as in Definitions 7.5.3
and 7.5.4, except that here, the real-model adversary A does not see messages sent
among honest parties (i.e., parties in I¯). We say that
t-securely computes f if
for every probabilistic algorithm A (representing a real-model adversary strategy),
there exists a probabilistic algorithm of comparable complexity B (representing an
ideal-model adversary strategy), such that for every I ⊂ [m] of cardinality at most t,
it holds that
(2) s
{ideal f, I, B(z) (x)}x,z ≡ {real
, I, A(z) (x)}x,z (7.63)

In case the ensembles in Eq. (7.63) are identically distributed, we say that the emulation
is perfect.

We stress that Eq. (7.63) requires statistical indistinguishability, whereas the analogue
requirement in Definition 7.5.4 is of computational indistinguishability. More impor-
tantly, we make no computational restrictions regarding the real-model adversary, and
require the corresponding ideal-model adversary to be of comparable complexity. The
latter requirement is very important: It prevents obviously bad protocols (see Exer-
cise 18), and it guarantees that Definition 7.6.2 is actually a strengthening of Defini-
tion 7.5.4 (see Exercise 19).

7.6.2. Security in the Semi-Honest Model

The following construction of t-private m-party protocols, for t < m/2, is a modifi-
cation of the construction presented in Section 7.5.2 (which in turn generalized the
construction presented in the two-party case [i.e., Section 7.3]). Recall that the core of
these constructions is the privately computed propagation of shares of bits through a
circuit that represents the desired computation. In the previous cases (see Sections 7.3
and 7.5.2), we used a very simple m-out-of-m secret-sharing scheme (i.e., a bit was
shared by m random bits that sum up to the value of the secret bit). Here, we use the
743
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

more sophisticated (t + 1)-out-of-m secret-sharing scheme of Construction 7.5.35 (i.e.,

a bit is shared by the values of a random degree t polynomial with free term that equals
the value of the secret bit). Thus, our focus is on propagating these types of shares
through the circuit, and on doing so via a t-private computation. Again, the heart of the
construction is performing the propagation through a single multiplication gate.
Let us clarify this discussion by being more specific about the details. We fix some
prime p > m and consider polynomials of degree t over GF( p).86 Recall that the value
of such a polynomial at t + 1 arbitrary (known) points allows for recovery of the
polynomial and specifically its free term. On the other hand, the value of a random
(degree t) polynomial at t arbitrary (known) points does not reveal information about
the value of the free term of the polynomial. Thus, each party will share each of its input
bits with all other parties, by uniformly selecting a random (degree t) polynomial with
free term equal to the value of this bit, and hand to Party i the value of this polynomial
at point i.
Suppose that the parties hold the shares of two GF( p) values and wish to derive
shares of the sum of these values, where all arithmetic operations refer to GF( p). Then
letting each party add the two shares it holds yields the desired shares. That is, suppose
that the values u and v are shared using the (degree t) polynomials a(·) and b(·), such
that u = a(0) and v = b(0), and Party i holds the shares ai = a(i) and bi = b(i). Then
the ai + bi ’s are shares of a polynomial c(·) that has free term u + v (i.e., letting c(z) =
a(z) + b(z), it holds that c(i) = ai + bi and c(0) = u + v). Furthermore, the degree of
c(·) is at most t. Thus, we are able to propagate shares through an addition gate, and
we do so in a totally private manner (because only local computations are used).
It is appealing to try to do the same in case of multiplication (rather than addition).
Indeed, the entire argument goes through, except that the corresponding polynomial
c may have a degree greater than t (but not more than 2t). Thus, we need a more
sophisticated way of propagating shares through multiplication gates. Using the same
notations (as previously), we consider the following (randomized) process:

Construction 7.6.3 (t-private m-party protocol for propagating shares through a mul-
tiplication gate): Recall that t < m/2, and so 2t ≤ m − 1.

Input: Party i enters with input (ai , bi ), where ai = a(i) and bi = b(i) for degree t
polynomials a(·) and b(·).
The protocol itself proceeds as follows:
1. For every i, Party i (locally) computes ci ← ai · bi .
def
Indeed, these ci ’s are the values of the polynomial c(z) = a(z) · b(z) at the corre-
sponding i’s, and c(0) = u · v. However, c may have degree 2t (rather than at most t).
2. For every i, Party i shares ci with all other parties. That is, Party i selects uniformly
a polynomial qi of degree t such that qi (0) = ci , and sends qi ( j) to Party j, for
every j.

86 Here and in the following, when we say a degree d polynomial, we actually mean a polynomial of degree at
most d.

744

www.Ebook777.com
 7.6* PERFECT SECURITY IN THE PRIVATE CHANNEL MODEL

Motivation: Extrapolation of the ci ’s yields the value of c(0) = u · v. In the following

we will let each party perform the corresponding operation on the shares it obtained.
We will show that this will yield shares with the desired properties.
Recall that by the Extrapolation Theorem, there exist constants γ1 , ..., γm such that
for every polynomial q of degree m − 1 it holds that

m
q(0) = γi q(i) (7.64)
i=1
m 
(Specifically, γi = −(1) i+1
· i .)
m
3. For every j, Party j (locally) computes d j ← i=1 γi qi ( j), where γ1 , ..., γm are
the extrapolation constants satisfying Eq. (7.64), and qi ( j) is the share that Party j
received from Party i in Step 2.

Output: Party i exits with output di .

It is quite clear that Construction 7.6.3 yields no information about u and v to any
coalition of t (or fewer) parties. The reason is that the only new information obtained
by t parties (i.e., t shares of each of the other ci ’s) yields no information about the
polynomials a and b (because it yields no information about these ci ’s). It is also clear
that every sequence of t of the di ’s is uniformly distributed (because the values of the
qi ’s at any t points are uniformly distributed). What is less clear is that the di ’s are indeed
admissible shares of the desired value (i.e., di = d(i) for some degree t polynomial d
having free term u · v). This fact will be established next.

Fact 7.6.4: Let the di ’s be defined as in Construction 7.6.3, and t < m/2. Then there
exists a degree t polynomial, d, such that d(0) = a(0) · b(0) and d(i) = di for i =
1, ..., m.

def m
Proof: Consider the formal polynomial q(z) = i=1 γi qi (z), where the qi ’s are the
polynomials selected at Step 2. Since each qi  has degree t, this holds also for q. For
m
every j = 1, ..., m, by Step 3, we have d j = i=1 γi qi ( j) = q( j), where the second
equality is due to the definition of q. Finally, note that


m
q(0) = γi qi (0)
i=1


m
= γi ci
i=1


m
= γi · a(i) · b(i)
i=1
= a(0) · b(0)
where the second equality is by Step 2, the third equality is by Step 1, and the last
equality is by the Extrapolation Theorem (applied to the 2t ≤ m − 1 degree polynomial
a(z) · b(z)).
745
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Conclusion. Using Fact 7.6.4, for t < m/2, one can show (see Exercise 23)
that Construction 7.6.3 constitutes a t-private computation of the (partial) m-ary
functionality

((a(1), b(1)), ..., (a(m), b(m))) → (r (1), ..., r (m)) (7.65)

where a and b are degree t polynomials and r is a uniformly distributed degree t

polynomial with free term equal to a(0) · b(0). By a straightforward adaptation of
Construction 7.5.10 and its analysis, it follows that any m-ary functionality can be
t-privately reduced to Eq. (7.65). Finally, by using a suitable Composition Theorem,
we obtain:

Theorem 7.6.5: For t < m/2, any m-ary functionality is t-privately computable. Fur-
thermore, the emulation is perfect.

In contrast, very few m-ary functionalities are t-privately computable for t ≥ m/2.
In particular, the only m-ary Boolean-valued functions that are m/2-privately com-
putable are linear combinations
m of Boolean-valued functions of the individual inputs
(i.e., f (x 1 , ..., x m ) = i=1 ci f (i) (xi ) mod 2).

7.6.3. Security in the Malicious Model

In order to deal with the malicious model, we replace Construction 7.6.3 with a more
robust protocol that t-securely computes Eq. (7.65). In particular, the protocol should
withstand a possible modification of t of the inputs (which, in particular, may not fit the
domain of the functionality as partially defined earlier). This turns out to be possible,
provided t < m/3, and so we get:

Theorem 7.6.6: For t < m/3, any m-ary functionality is t-securely computable. Fur-
thermore, the emulation is perfect.

We briefly sketch the ideas that underlie the proof of Theorem 7.6.6. Let us first assume
that t < m/4, and note that Steps 2–3 of Construction 7.6.3 constitute a t-private
computation of the (partial) m-ary functionality

(c(1), ..., c(m)) → (r (1), ..., r (m)) (7.66)

where c is a degree 2t polynomial and r is a uniformly distributed degree t polyno-

mial with free term equal to c(0). We wish to t-securely compute Eq. (7.66). Let us
first consider the related task of t-securely computing c(0). Construction 7.5.10 sug-
gests that c(0) can be computed by extrapolation of the c(i)’s, and that extrapolation
is a linear function, which (as such) can be t-privately computed (see Exercise 20).
However, when some parties are malicious, simple extrapolation will fail. What we
need is a “robust extrapolation” procedure, which corresponds to error correction
of Reed-Solomon codes, which in turn is a linear function of the given sequence.

746

www.Ebook777.com
 7.7 MISCELLANEOUS

Specifically, this task is to find the free term of the unique degree 2t polynomial (i.e.,
c) that fits at least m − t of the inputs (i.e., the correct c(i)’s), and we can perform
this task in a t-secure manner. (The desired polynomial is indeed unique, because
otherwise we get two different degree 2t polynomials that agree on m − 2t ≥ 2t + 1
of the inputs.) Finally, observe that the parties can t-securely generate shares of a
random degree t polynomial with free term equal to zero. Combining the two linear
computations, one obtains the desired t-secure implementation of Eq. (7.66), provided
that t < m/4.
In order to handle the case m/4 ≤ t < m/3, we have to work directly with Eq. (7.65),
rather than with Eq. (7.66); that is, we use the fact that the parties actually hold the
shares of two degree t polynomials, rather than only the product of these shares (which
corresponds to shares of a degree 2t polynomial).

7.7. Miscellaneous

7.7.1.* Three Deferred Issues

In this section, we briefly discuss three important issues that were avoided (for the sake
of simplicity) in previous sections.

7.7.1.1. Partial Fairness, or On Exchanging Secrets

As commented upon in Section 7.2.3, in general, no two-party protocol can guarantee
perfect fairness; that is, it cannot be guaranteed that one party obtains its desired output
if and only if the other party obtains its own desired output. Intuitively, an adversary
may always abort at the first possible time at which it obtains its output, and this means
that one of the parties may obtain the desired output while the other party does not quite
get its own output. In fact, in the specific (two-party and multi-party) protocols that
we have presented, this phenomenon occurs in an extreme sense; that is, Party 1 gets
the output before any other party gains any knowledge regarding its own output. As
we will show, the severity of this phenomenon can be reduced (but, as shown in [65],
cannot be totally eliminated). That is, “partial fairness” (alas, not “perfect fairness”)
may be achieved in some sense. In the rest of this section, we focus on two-party pro-
tocols, but similar treatment can be applied to multi-party protocols (lacking an honest
majority).
A general framework for obtaining partial fairness consists of first computing shares
of both desired outputs, and next gradually revealing pieces of these shares, such that
a party reveals the next piece only if its counterpart has revealed the previous piece.
The parties should be able to verify the correctness of the revealed pieces, which can
be achieved by generating also commitments to these pieces (and asking the reveal-
ing party to also provide the corresponding decommitment information). Thus, for

747
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

a functionality f , which without loss of generality satisfies | f 1 (x, y)| = | f 2 (x, y)|,
we may proceed in two stages:

1. The parties securely compute shares of the desired outputs of f . Specifically, the
parties securely compute the functionality

(x, y) → ((v1 ⊕ s1 , s2 , r1 , c) , (s1 , v2 ⊕ s2 , r2 , c))

where (v1 , v2 ) ← f (x, y), the si ’s are uniformly distributed in {0, 1}|vi | , and
c ← C r1 ⊕r2 (v1 , v2 ), for uniformly distributed r1 , r2 ∈ {0, 1}|v1 ,v2 | . Note that at this
2

stage, each individual party obtains no knowledge of the desired outputs, but to-
gether they hold (verifiable) secrets (i.e., the vi ⊕ si ’s and si ’s) that yield both
outputs.
2. The parties gradually exchange the secrets that they hold. That is, Party 1 re-
veals pieces of s2 in exchange for pieces of s1 (revealed by Party 2), where one
piece of s2 is revealed per one piece of s1 . The pieces are revealed by using
a secure computation of an adequate functionality. Suppose that Party i is sup-
posed to obtain the piece πi (si ), where πi may be a (predetermined) Boolean
function or a randomized process. Then the parties securely compute the func-
tionality that maps ((a1 , a2 , ρ1 , γ1 ) , (b1 , b2 , ρ2 , γ2 )) to (π1 (b1 ), π2 (a2 )) if γ1 =
γ2 = C ρ1 ⊕ρ2 (a1 ⊕ b1 , a2 ⊕ b2 ) and to (λ, λ) otherwise. Indeed, each party en-
ters this secure computation with the input it has received in the first stage;
that is, Party 1 (resp., Party 2) enters with input (v1 ⊕ s1 , s2 , r1 , c) (resp.,
(s1 , v2 ⊕ s2 , r2 , c)).
The entire approach (and, in particular, the gradual exchange of secrets) depends
on a satisfactory definition of a piece of a secret. Such a definition should satisfy
two properties: (1) Given sufficiently many pieces of a secret, one should be able to
recover the secret, whereas (2) getting yet another piece of the secret contributes little
to the knowledge of the secret. We admit that we do not know of a definition (of a
piece of a secret) that is “uncontroversially satisfactory”; still, some suggestions (for
what these pieces of information may be) seem quite appealing. For example, consider
the randomized process π that maps the n-bit long secret σ1 · · · σn to the n-bit long
string τ1 · · · τn , such that τi = σi with probability 12 + ε and τi = 1 − σi otherwise, for
every i, independently.87 Then each piece carries O(nε 2 ) bits of information, whereas
after seeing t such pieces of the secret, one can guess it with success probability at
least 1 − n · exp(−tε 2 ), which for t = O(n/ε 2 ) means practically obtaining the secret.
However, if Party 1 knows that s1 ∈ {0n , 1n }, whereas Party 2 only knows that s2 ∈
{0, 1}n , then π (s1 ) seems more meaningful to Party 1 than π(s2 ) is to Party 2. Is it really
so or is the proposed exchange actually fair? Note that things are even more complex
(than they seem), because the uncertainty of the parties is actually not information-
theoretic but rather computational.

87 An alternative randomized process π maps the n-bit string s to the random pair (r, b), such that r is uniformly
distributed in {0, 1}n and b ∈ {0, 1} equals the inner product (mod 2) of s and r with probability 12 + ε (and the
complementary value otherwise). In this case, each piece carries O(ε 2 ) bits of information about s, whereas
after seeing O(n/ε2 ) such pieces, one practically obtains s.

748

www.Ebook777.com
 7.7 MISCELLANEOUS

7.7.1.2. The Adaptive Model

The definitions presented in Section 7.5.1 referred to adversaries, called non-adaptive,
that control a predetermined set of parties (which, of course, is not known to the honest
parties).88 In this section, we consider a stronger type of adversaries, called adaptive,
that can select the parties that they control as the execution proceeds. To demonstrate the
power of adaptive adversaries, consider an m-party protocol in which Party 1 uniformly
selects an m/3-subset J of the parties, publicizes J , and shares its own input  with the
parties in J as a whole (i.e., it hands each Party j in J a random r j such that j∈J r j
equals its own input). Treating m as a parameter, this protocol (for computing nothing)
is secure with respect to Definition 7.5.4, essentially because for every set I of fewer
than m/2 parties, it holds that the probability that a random m/3-subset J is contained in
I is exponentially vanishing in m. However, an adaptive adversary that selects the set of
parties that it controls to equal the publicized set J obtains the input of Party 1 without
controlling it (and, hence, demonstrates that the protocol is insecure with respect to
adaptive adversaries).
In general, an adaptive adversary is one that can decide which parties to corrupt (i.e.,
seize control of) during the course of the execution of the protocol. Potentially, such an
adaptive decision may be more beneficial to the adversary than an oblivious decision.
Security in the adaptive model means that even an adaptive adversary cannot gain from
the execution more than what is unavoidable (even in the presence of a trusted party).
To actually define security with respect to adaptive adversaries, we should first define
an adequate ideal model, which corresponds to what is unavoidable when considering
adaptive adversaries. The crucial point is that even in an ideal-model execution, the
adversary may select the parties that it controls adaptively and based on the information
it has gathered so far (i.e., the inputs of the parties controlled so far).89 We stress that once
the adversary seizes control of a party, it learns the party’s initial input (and in the real
model, it also learns its random-tape and the messages that this party has received so far).
When defining the result of such an ideal-model execution, we also include in it
the set of parties that the adversary controls. The same is done when defining the
result of the real-model execution. Consequently, when we require that the ideal-model
execution can emulate the real-model execution, the executions must refer to the same
(or computationally indistinguishable) sets of controlled parties. Actually, one should
also consider the order in which the controlled parties are selected. To clarify this
discussion, let use consider an extension of Definition 7.5.4 (i.e., the second malicious
model) to the adaptive model.

Definition 7.7.1 (security in the malicious adaptive model, a sketch): Let f and
be
as in Section 7.5.1, and t be a bound on the number of parties that the adversary is
allowed to control (e.g., t < m/2).

88 The issue of adaptivity also arises, but in a more subtle way, in the case of two-party protocols.
89 The non-adaptive model can be viewed as a special case in which the adversary selects the parties that it controls
up-front, before learning any information regarding the current execution. But in general (in the adaptive model),
only the choice of the first controlled party is oblivious of the execution.

749
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

r A t-adaptive ideal-model adversary is a randomized process that operates in up-to

t + 1 steps, which are partitioned into two main phases. In each step of the first
phase, based on the information available to it, the adversary decides whether to
seize control of another party or to move to the second phase. In the first case,
the adversary also determines the identity of the new party to be controlled and
obtains its local input. In the second case, the adversary invokes the trusted party
and supplies the trusted party with inputs of its choice corresponding to the parties
that it currently controls. At this point, the other parties supply the trusted party with
their original inputs, and the trusted party determines the corresponding outputs
and provides each party with its corresponding output, where the adversary receives
all the outputs of parties that it controls.90
In each step of the second phase, based on the information available to it, the
adversary decides whether or not to seize control of another party and if so also
determines its identity. Consequently, the adversary receives the local input and
output of this party. The joint computation in the ideal model, under an adaptive
adversary, is defined as the concatenation of the outputs of the uncontrolled parties,
the adversary’s output, and the sequence of the parties on which the adversary gained
control.
r A t-adaptive real-model adversary is a randomized strategy that corresponds to
an attack on the actual execution of the protocol. Such an adversary may adaptively
select up to t parties it wishes to control, obtain their current view of the execution
(as per Definition 7.2.1), and determine their actions. The adversary may select some
parties before the actual execution starts, some parties during the actual execution,
and some after it has terminated, as long as the total number of selected parties
is at most t. The joint computation in the real model, under an adaptive adver-
sary, is defined as the concatenation of the outputs of the uncontrolled parties, the
adversary’s output, and the sequence of the parties on which the adversary gained
control.
r Protocol
for computing f is called t-adaptively secure if for every feasi-
ble t-adaptive real-model adversary A, there exists a feasible t-adaptive ideal-
model adversary B, such that the joint computation in the real model under A
is computationally indistinguishable from the joint computation in the ideal model
under B.
We stress that in the real model, when the adversary seizes control of a party, it gets
the party’s view of the execution so far (where the party’s view is as defined in Defini-
tion 7.2.1). In particular, the protocol’s possible instructions to erase certain data does
not affect the party’s view, which always contains its input, its random-tape, and all
messages it has received so far. A weaker notion of security postulates that when the
adversary seizes control of a party, it only gets the current values of the party’s local
variables as determined by the protocol (in which case, the adversary does not obtain

90 As in Definition 7.5.4 (and unlike in Definition 7.5.2), the trusted party always answers all parties; that is, the
adversary has no option of preventing the trusted party from answering the honest parties. Recall that here the
trusted party is invoked (by the adversary) at the time the adversary decides that it controls enough parties.

750

www.Ebook777.com
 7.7 MISCELLANEOUS

data that was explicitly erased by an instruction of the protocol). Our definitional choice
is motivated by the fear that the past values of the party’s local variables (i.e., the party’s
view as per Definition 7.2.1) may be available somewhere on its computing system; see
analogous discussion in Section 7.2.2 (regarding the semi-honest model).

Theorem 7.7.2 (main results regarding adaptively secure protocols):

1. In the private channel model, any m-ary functionality can be computed in a

(m − 1)/3 -adaptively secure manner. Furthermore, as in Theorem 7.6.6, the em-
ulation is perfect.
2. Assuming the intractability of inverting RSA (or of the DLP), any m-ary function-
ality can be computed in a (m − 1)/3 -adaptively secure manner, even when the
adversary can tap all communication lines.
Part 1 follows by extending the proof of Theorem 7.6.6, that is, by observing that the
protocols used toward proving the latter result are in fact adaptively secure. Proving
Part 2 is more problematic. In particular, a straightforward application of the pre-
compiler described in Section 7.5.3.1 seems to fail. The source of trouble is that standard
encryption schemes, which may be used to emulate private (point-to-point) channels
over ordinary (point-to-point) channels, effectively “commit” to the single value that
was sent (which is a problem because of messages sent between honest parties that
are later corrupted by the adversary). Intuitively, the solution is to use non-standard
encryption schemes (i.e., “non-committing” ones). The latter can be constructed using
trapdoor permutations with certain additional properties.

7.7.1.3. Reactive Systems

Our treatment so far has focused on functionalities that represent standard (multi-
party) computations, mapping (sequences of) inputs to (sequences of) outputs. A more
general treatment may refer to (multi-party) reactive systems that iteratively respond
to inputs presented from the outside. Furthermore, the functionalities of these reactive
systems may depend on a (global) state that they maintain and update. This global state
may not be known to any individual party (but is rather the concatenation of the local
states that the individual parties maintain and update). Thus, we view (multi-party)
reactive systems as iterating the following steps (for an a priori unbounded number of
times):91

r Parties are given inputs for the current iteration; that is, in the j-th iteration Party i is
( j)
given input xi . In addition, there is a global state: The global state at the beginning of
the j-th iteration is denoted s ( j) , where the initial global state is empty (i.e., s (1) = λ).
r Depending on the current inputs and the global state, the parties are supposed to
compute outputs for the current iteration, as well as update the global state. That is,

91 As usual, the number of iterations (and the length of the inputs) must be polynomial in the security parameter.
Furthermore, the length of the global state (at any time) must also be polynomial in the security parameter.

751
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

( j)
the outputs in iteration j are determined by the xi ’s, for all i’s, and s ( j) . The new
( j)
global state, s ( j+1) , is determined similarly (i.e., also based on xi ’s and s ( j) ).
As it is an abstraction, one may think of the global state as being held by a trusted
party. In other words, reactive systems are captured by reactive functionalities in
which the trusted party maintains a state and interacts with the actual parties in iter-
ations. Indeed, in each iteration, the trusted party obtains an input from each party,
responds (as directed by the reactive functionality) with corresponding outputs, de-
pending also on its state, and updates its state. Note that the latter formulation fits
a definition of an ideal model (for computing the reactive functionality), whereas a
(real-model) reactive protocol must emulate this augmented notion of a trusted party.
Thus, the reactive protocol should emulate the iterative computation of outputs while
maintaining the state of the imaginary trusted party. Indeed, it is natural to have the
real-model parties use a secret-sharing scheme in order to maintain the latter state
(such that the state remains unknown to individual parties and even to a bounded num-
ber of dishonest parties). In fact, we need to use a verifiable secret-sharing scheme (see
Section 7.5.5.1), because dishonest parties should be prevented from (illegally) modi-
fying the (system’s) state (except from the predetermined effect of the choice of their
own inputs).
This discussion suggests that the secure implementation of reactive functionalities
can be reduced to the secure implementation of ordinary (i.e., non-reactive) function-
alities. For example, we refer to security in the second malicious model, as defined
in Definition 7.5.4 (for ordinary functionalities). That is, we postulate that a major-
ity of the parties are honest and require that the dishonest parties cannot (effectively)
abort the execution. In such a case, we use a verifiable secret-sharing scheme in which
only a majority of the pieces yield the secret. Once a verifiable secret-sharing scheme
is fixed and the (system’s) state is shared using it, the computation of each iteration
of the reactive system can be cast as an ordinary functionality. The latter maps se-
quences of the form ((x1 , s1 ), ..., (xm , sm )), where xi denotes the current input of Party i
and si denotes its share of the current state, to the sequence ((y1 , r1 ), ..., (ym , rm )),
where yi denotes the next output of Party i and ri denotes its share of the updated
state.
We conclude that the results regarding secure computation of ordinary (i.e., non-
reactive) computations can be extended to reactive systems (thus obtaining secure
implementations of the latter).

7.7.2.* Concurrent Executions

A natural problem regarding cryptographic protocol is whether (or to what extent) they
preserve their security when executed concurrently. The problems that arise with respect
to the preservation of zero-knowledge (see Section C.5.1) are merely an indication to
the type of problems that we may encounter. The lesson to be learned (even from that
brief discussion) is that an adversary attacking several concurrent executions of the
same protocol may be able to cause more harm than by attacking a single execution
(or several sequential executions) of the same protocol.
752

www.Ebook777.com
 7.7 MISCELLANEOUS

7.7.2.1. Definitions
One may say that a protocol is concurrently secure if whatever the adversary may ob-
tain by invoking and controlling parties in real concurrent executions of the protocol is
also obtainable by a corresponding adversary that controls corresponding parties mak-
ing concurrent functionality calls to a trusted party (in a corresponding ideal model).
More generally, one may consider concurrent executions of many sessions of several
protocols, and say that a set of protocols is concurrently secure if whatever the ad-
versary may obtain by invoking and controlling such real concurrent executions is also
obtainable by a corresponding adversary that invokes and controls concurrent calls to
a trusted party (in a corresponding ideal model). Consequently, a protocol is said to be
secure with respect to concurrent compositions if adding this protocol to any set of
concurrently secure protocols yields a set of concurrently secure protocols.
A much more appealing approach has been recently suggested by Canetti [51].
Loosely speaking, he suggests considering a protocol to be secure (hereafter referred
to as environmentally secure)92 only if it remains secure when executed within any
(feasible) environment. The notion of an environment is a generalization of the notion
of an auxiliary-input; in a sense, the environment is an auxiliary oracle (or rather a
state-dependent oracle) that the adversary may access. In particular, the environment
may represent other executions of various protocols that are taking place concurrently
(to the execution that we consider). We stress that the environment is not supposed to
assist the proper execution of the protocol (and, in fact, honest parties merely obtain
their inputs from it and return their outputs to it). In contrast, potentially, the envi-
ronment may assist the adversary in attacking the execution. Following the simulation
paradigm, we say that a protocol is environmentally secure if any feasible real-model
adversary attacking the protocol, with the assistance of any feasible environment, can
be emulated by a corresponding ideal-model adversary that uses the same environment,
while making similar queries to the environment. In the following formulation, the en-
vironment is implemented by a (non-uniform) family of polynomial-size circuits, and
is also responsible for providing the parties with inputs and for trying to distinguish the
real-model execution from the ideal-model execution.

Definition 7.7.3 (Environmentally Secure Protocols, a rough sketch): Let f be an m-

ary functionality and
be an m-party protocol, and consider the following real and
ideal models:
r As usual, a real-model adversary controls some of the parties in an execution of
the protocol
. In addition to executing
, all parties can communicate with an
arbitrary interactive process, which is called an environment. Honest parties only
communicate with the environment before the execution starts and when it ends;
they merely obtain their inputs from the environment and pass their outputs to it.
In contrast, dishonest parties (controlled by the adversary) may communicate freely
with the environment and do so concurrently with the entire execution of
.

92 The term used in [51] is Universally Composable, but we believe that a reasonable sense of “universal compos-
ability” is only a corollary of the suggested definition.

753
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

r An ideal-model adversary controls some of the parties in an ideal computation

assisted by a trusted party that behaves according to the functionality f . In addition,
all parties can communicate with an environment (as in the real model). Indeed,
the dishonest parties may communicate extensively with the environment before and
after their single communication with the trusted party, whereas the honest parties
merely obtain their inputs from the environment and pass their outputs to it.
We say that
is an environmentally secure protocol for computing f if for ev-
ery probabilistic polynomial-time real-model adversary A there exists a probabilistic
polynomial-time ideal-model adversary B, such that for any subset I ⊂ [m] of ade-
quate cardinality,93 no family of polynomial-size circuits E = {E n }n∈N can distinguish
the case in which it interacts with parties in the real-model execution of
under
adversary (I, A) from the case in which it interacts with parties in the ideal-model
computation of f under adversary (I, B). Schematically,
c
{ideal f, I, B(1n ), E n }n∈N ≡ {real
, I, A(1n ), En }n∈N

where ideal f, I, B(1n ), En (resp., real
, I, A(1n ), E n ) denotes the output of E n after interacting
with the ideal-model (resp., real-model) execution under (I, B) (resp., (I, A)).

As hinted earlier, the environment may account for other executions of various protocols
that are taking place concurrently with the main execution being considered. Defini-
tion 7.7.3 implies that such environments cannot distinguish the real execution from an
ideal one. This means that anything that the real-model adversary gains from the exe-
cution of the protocol and any environment (representing other concurrent executions)
can also be obtained by an adversary operating in the ideal model and having access to
the same environment. Thus, each single execution of an environmentally secure pro-
tocol can be replaced by an ideal oracle call to the corresponding functionality, without
affecting the other concurrent executions. Furthermore, one can simultaneously replace
all these concurrent executions by ideal oracle calls and use a hybrid argument to show
that the behavior is maintained. (One needs to use the fact that a single replacement does
not affect the other concurrent executions, even in case some of the other executions
are in the real model and the rest are in the ideal model.) It follows that environmentally
secure protocols are secure with respect to concurrent compositions [51]. We wonder
whether the reverse direction holds.

7.7.2.2. Constructions
The main positive result currently known is that environmentally secure protocols
for any functionality can be constructed for settings in which more than two-thirds
of the active parties are honest (cf. [51]). This holds unconditionally for the private-
channel model and under standard assumptions (e.g., allowing the construction of
public-key encryption schemes) for the standard model (i.e., without private channel).

93 Thus, the definition should actually specify an additional parameter bounding the number of parties that may
be controlled by the adversary.

754

www.Ebook777.com
 7.7 MISCELLANEOUS

The immediate consequence of this result is that general environmentally secure multi-
party computation is possible, provided that more than two-thirds of the parties are
honest.
In contrast, general environmentally secure two-party computation is not possible
(in the standard sense).94 Still, one can salvage general environmentally secure two-
party computation in the following reasonable model: Consider a network that contains
servers that are willing to participate (as “helpers,” possibly for a payment) in compu-
tations initiated by a set of (two or more) users. Now, suppose that two users wishing to
conduct a secure computation can agree on a set of servers such that each user believes
that more than two-thirds of the servers (in this set) are honest. Then, with the active
participation of this set of servers, the two users can compute any functionality in an
environmentally secure manner.
Another reasonable model where general environmentally secure two-party com-
putation is possible is the shared random-string model [59]. In this model, all parties
have access to a universal random string (of length related to the security parameter).
We stress that the entity trusted to post this universal random string is not required to
take part in any execution of any protocol, and that all executions of all protocols may
use the same universal random string.

7.7.3. Concluding Remarks

In this chapter, we have presented a host of definitions of security for multi-party
protocols (especially for the case of more than two parties). Furthermore, some of
these definitions are incomparable to others (i.e., they neither imply the others nor are
implied by them), and there seems to be no single definition that may be crowned as
the central one.
For example, as stated in Section 7.5.1.3, the security definitions for the two ma-
licious (multi-party) models (i.e., Definitions 7.5.3 and 7.5.4) are incomparable and
there is no generic reason to prefer one over the other. Actually, one could formulate a
natural definition that implies both Definitions 7.5.3 and 7.5.4. We refer to waiving the
bound on the number of dishonest parties in Definition 7.5.4 (or, equivalently, eliminat-
ing the abort as an admissible option in the ideal model underlying Definition 7.5.3).
That is, the resulting definition is free of the annoying restrictions (or provisions) that
were introduced in each of the previous definitions. The “only” problem with the re-
sulting definition is that it cannot be satisfied (in general), whereas each of the previous
definitions could be satisfied. Thus, for the first time in this work, we have reached
a situation in which a natural (and general) definition cannot be satisfied, and we are
forced to choose between two weaker alternatives, where each of these alternatives car-
ries fundamental disadvantages (beyond the fact that security holds only with respect
to probabilistic polynomial-time adversaries).
In general, the current chapter carries a stronger flavor of compromise (i.e., recog-
nizing inherent limitations and settling for a restricted meaningful goal) than previous

94 Of course, some specific two-party computations do have environmentally secure protocols. See [51] for several
important examples (e.g., key exchange).

755
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

chapters. In contrast to the impression given in other parts of this work, it is now obvious
that we cannot get all that we may want. Instead, we should study the alternatives and
go for the one that best suits our real needs.
Indeed, as stated in the preface, the fact that we can define a cryptographic goal
does not mean that we can satisfy it as defined. In case we cannot satisfy the initial
definition, we should search for acceptable relaxations that can be satisfied. These
relaxations should be defined in a clear manner so that it would be obvious what
they achieve and what they fail to achieve. Doing so will allow a sound choice of
the relaxation to be used in a specific application. That is, the choice will have to be a
circumstantial rather than a generic one. This seems to be a good point at which to end the
current work.

A good compromise is one in which

the most important interests of all parties are satisfied.
Adv. Klara Goldreich-Ingwer (1912–2004)

7.7.4. Historical Notes

The main results presented in this chapter (i.e., Theorems 7.4.1 and 7.5.15) are due
to Yao [191] and to Goldreich, Micali, and Wigderson [117, 118], treating the two-
party and multi-party cases, respectively. Unfortunately, the original papers do not
provide a satisfactory presentation of these results. In particular, these papers lack
adequate definitions of security (which were developed only later) and provide only
rough sketches of the constructions and no proofs of their security. Still, the conference
version of [117] provides a rough sketch of the compilation of protocols for the semi-
honest model into protocols for the malicious model, by using zero-knowledge proofs
(which are the main focus of [117]) to “force” malicious parties to behave in a semi-
honest manner. Yao’s work [191] presents a construction that can be used to derive
two-party protocols for privately computing any desirable functionality, whereas the
second paper of Goldreich et. al. [118] presents a different construction for the multi-
party case.
Our presentation reverses the chronological order (in which these results were dis-
covered). Firstly, our treatment of the two-party case is derived, via some degeneration,
from the treatment of the multi-party case (in [117, 118]). Secondly, we start by treating
the semi-honest models and only later compile protocols for this model into protocols
for the malicious models. We note that (following [118]) our presentation of the main
protocols is essentially symmetric, whereas Yao’s original protocol for the two-party
case [191] is asymmetric (with respect to the two parties). The latter asymmetry has its
own merits as demonstrated in [27, 165, 143].
In treating the semi-honest model, we follow the framework of Goldreich, Micali,
and Wigderson [118], while adapting important simplifications due to [129] and [120].
In presenting the “semi-honest to malicious” compilers (or the paradigm of “forcing”
semi-honest behavior), we follow the outline provided in [117, FOCS Version, Sec. 4]

756

www.Ebook777.com
 7.7 MISCELLANEOUS

and [118, Sec. 5]. We comment that the original sources (i.e., [117, 118]) are very terse,
and that full details were only provided in [107]. Our treatment differs from [107] in
using a higher level of modularity, which is supported by composition theorems for the
malicious models.
As stated earlier, a satisfactory definitional treatment of secure multi-party compu-
tation was provided after the presentation of the constructions of [117, 118, 191]. The
basic approach was developed by Micali and Rogaway [157] and Beaver [10, 11],95 and
reached maturity in Canetti’s work [50], which provides a relatively simple, flexible,
and comprehensive treatment of the (basic) definitions of secure multi-party com-
putation. In particular, the composition theorems that we use are essentially taken
from [50].
A variety of cryptographic tools is used in establishing the main results of this chapter.
Firstly, we mention the prominent role of Oblivious Transfer in the protocols developed
for the semi-honest model.96 An Oblivious Transfer protocol was first suggested by
Rabin [172], but our actual definition and implementation follow the ideas of Even,
Goldreich, and Lempel [84] (as further developed in the proceedings version of [117]).
Several ingredients play a major role in the compilation of protocols secure in the semi-
honest model into generally secure protocols (for the malicious models). These include
commitment schemes, zero-knowledge proofs-of-knowledge, verifiable secret sharing
(introduced by Chor, Goldwasser, Micali, and Awerbuch [63]), and secure coin-flipping
(introduced by Blum [37]).

The Private Channel Model. In contrast to the bulk of this chapter (as well as the
bulk of the entire work), the private channel model (treated in Section 7.6) allows
the presentation of results that do not rely on intractability assumptions. These re-
sults (e.g., Theorem 7.6.6) were obtained by Ben-Or, Goldwasser, and Wigderson [34]
and Chaum, Crépeau, and Damgård [62]. These works were done after the results of
Yao [191] and Goldreich, Micali, and Wigderson [117, 118] were known, with the ex-
plicit motivation of obtaining results that do not rely on intractability assumptions. Our
presentation is based on [34] (cf. [97]). The essential role of the bound on the number of
dishonest parties (even in the semi-honest model) was studied in [64] and subsequent
works.

7.7.5. Suggestions for Further Reading

As hinted, Yao’s alternative treatment of the two-party case offers some advantages
over the treatment presented in Section 7.3. A sketch of Yao’s construction is provided
in Section 7.1.3.2. For more details, see [177].

95 The approach of Goldwasser and Levin [121] is more general: It avoids the definition of security (with respect
to a given functionality) and defines instead a notion of protocol robustness. Loosely speaking, a protocol is
robust if whatever an arbitrary malicious adversary can obtain by attacking it can also be obtained by a very
benign adversarial behavior.
96 Subsequent results by Kilian [137] further demonstrate the importance of Oblivious Transfer in this context.

757
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

The aforementioned results were originally obtained using protocols that use a
polynomial number of rounds. In some cases, subsequent works obtained secure
constant-round protocols (e.g., in the case of multi-party computations with honest
majority [27], and in the case of two-party computations allowing abort [143]).
We have mentioned (e.g., in Section 7.7.1.1) the impossibility of obtaining fairness
in secure computations without an honest majority. These statements are backed by the
impossibility of implementing a fair two-party coin-toss, as proven in [65].
We have briefly discussed the notion of adaptive adversaries. A more detailed dis-
cussion of the definitions is provided in [50], which builds on [49]. For a proof of
Theorem 7.7.2, the reader is referred to [49, 53]. For a study of adaptive versus non-
adaptive security, the reader is referred to [52].
Our treatment of multi-party protocols assumes a synchronous network with point-
to-point channels between every pair of parties. Results for asynchronous communi-
cation and arbitrary networks of point-to-point channels were presented in [33, 49]
and [78], respectively.
General secure multi-party computation in a model of transient adversarial behavior
was considered in [166]. In this model, the adversary may seize control of each party
during the protocol’s execution, but can never control more than (say) 10 percent of the
parties at any point in time. We comment that schemes secure in this model were later
termed “proactive” (cf., [57]).
Whenever we have restricted the adversary’s control of parties, we have done so by
bounding the cardinality of the set of controlled parties. It is quite natural to consider
arbitrary restrictions on the set of controlled parties (i.e., that this set belongs to a
family of sets against which security is guaranteed). The interested reader is referred
to [131].
For further discussion of Byzantine Agreement, see any standard textbook on Dis-
tributed Computing (e.g., [3, 147]). We mention that whereas plain m-party Byzantine
Agreement can tolerate at most (m − 1)/3 malicious parties, Authenticated Byzan-
tine Agreement can tolerate any number of malicious parties (see Construction 7.5.17,
which follows [80]). The problems arising when composing Authenticated Byzantine
Agreement are investigated in [144].

7.7.6. Open Problems

Recall that by Theorem 7.5.12, one can privately reduce any functionality to Oblivious
Transfer. Furthermore, the compilation of protocols that are secure in the semi-honest
model into protocols that are secure in the malicious models only requires one-way
functions and private channels, whereas the latter can be emulated using secure com-
munication protocols (e.g., secure public-key encryption schemes). Since Oblivious
Transfer implies the existence of the latter (see Exercise 7), general secure computa-
tion is reducible to Oblivious Transfer. Thus, determining the complexity assumptions
required for the implementation of Oblivious Transfer seems to be of key importance.
In particular, we have shown that Oblivious Transfer can be implemented using en-
hanced trapdoor permutations (see Proposition 7.3.6). We wonder whether the en-
hanced requirement can be omitted (i.e., whether ordinary trapdoor permutations may
758

www.Ebook777.com
 7.7 MISCELLANEOUS

suffice).97 For further discussion of enhanced trapdoor permutations, see Section C.1
in Appendix C.

7.7.7. Exercises
Exercise 1: Oblivious sampling: Suppose that both parties hold a function (or circuit)
that defines a distribution in the natural way and wish to obtain a sample from
this distribution without letting any party learn the corresponding pre-image. Cast
this problem as one of securely computing a corresponding functionality, treating
differently the case in which the function (or circuit) is fixed and the case in which it
is given as input to both parties. Consider also the case in which only the first party
is to obtain the output.
Exercise 2: Oblivious signing: In continuation of Exercise 1, consider the case in
which the distribution to be sampled is determined by the inputs of both parties. For
example, consider the task of oblivious signing in which one party wishes to obtain
the signature of the second party to some document without revealing the document
to the signer (i.e., the document is the input of the first party, whereas the signing-key
is the input of the second party).
Exercise 3: Privacy and Correctness: Referring to the discussion that follows Defini-
tion 7.2.6, consider the following definitions of (partial) privacy and correctness (with
respect to malicious adversaries). Partial privacy is defined as a restriction of Defi-
nition 7.2.6 to the adversary’s component of the random variables real
, A(z) (x, y)
and ideal f, B(z) (x, y), whereas partial correctness coincides with a restriction of
Definition 7.2.6 to the honest party’s component of these random variables.

1. Show that both properties are implied by Definition 7.2.6, but that even their
combination does not imply Definition 7.2.6.
2. Why were both properties qualified by the term “partial”?

Guideline (Item 1): Note that computational indistinguishability of ensembles

of pairs implies computational indistinguishability of the ensembles resulting by
projection to each coordinate, but the converse does not necessarily hold.

Guideline (Item 2): This is related to the need to use the general formulation
of Definition 7.2.1 for randomized functionalities; see the discussion that follows
Definition 7.2.1.

Exercise 4: On the importance of the length convention: Show that if the equal-length
convention is omitted from definitions like Definition 7.2.1 and 7.2.6, then they
cannot be satisfied for many natural functionalities. That is, consider these definitions
when the ensembles are indexed by the set of all pairs of strings, rather than by the
set of pairs of equal-length strings.

97 Partial progress toward this goal is reported in Haitner’s work “Implementing Oblivious Transfer using collection
of dense trapdoor permutations” (proceedings of the first Theory of Cryptography Conference, 2004).

759
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Guideline: (Here, privacy and security refer to the notions obtained when
omitting the equal-length convention.) Show that the functionality (x, y) →
def def
( f (x, y), f (x, y)), where f (x, y) = 1 if |x| = |y| and f (x, y) = 0 otherwise, can-
not be privately computed. Show that (x, y) → (|y|, |x|) can be privately computed
but tthat he simple protocol in which Party 1 sends |x| to Party 2 (and Party 2 sends
|y| to Party 1) fails to securely compute it. Challenge: Try to show that the latter
functionality cannot be securely computed.

Exercise 5: Transitivity of privacy reductions: Show that if f is privately reducible to

f
, and f
is privately reducible to f

, then f is privately reducible to f

. Note that
Theorem 7.3.3 is obtained as a special case (e.g., by setting f

to be the identity
mapping).


Guideline: Generalize the proof of Theorem 7.3.3. Specifically, let
f | f (resp.,





f | f ) be an oracle-aided protocol for f (resp., f
) using oracle f
(resp., f

).
Composing these two protocols, obtain and analyze the oracle-aided protocol
=




f|f .

Exercise 6: Variants of Oblivious Transfer: In continuation of Section 7.3.2, consider

the following two variants of Oblivious Transfer.

1. For functions k,  : N → N, consider the extension of 1-out-of-k Oblivious Trans-

fer to k(n) secrets each of length (n), where n is the security parameter.
2. For a function  : N → N, consider the Oblivious Transfer of a single (n)-bit
long secret (denoted σ ) that is to be delivered with probability 1/2; that is, the
randomized functionality that maps (σ, λ) to (λ, σ ) with probability 1/2 and to
(λ, λ) otherwise.

Assuming that k and  are polynomially bounded and efficiently computable, present
privacy reductions between all these variants. Specifically, show a privacy reduction
of the extended 1-out-of-k Oblivious Transfer to the original 1-out-of-2 Oblivious
Transfer of bits, and between 1-out-of-2 Oblivious Transfer of -bit long secrets and
Oblivious Transfer of a single (n)-bit long secret.
Guideline: Note that you are asked only to present oracle-aided protocols that are
secure in the semi-honest model. The only non-obvious reduction is from 1-out-
of-2 Oblivious Transfer to single-secret Oblivious Transfer (OT), presented next.
The first party randomly selects r1 , r2 ∈ {0, 1}(n) , and the parties invoke OT twice
where the first party inputs r1 in the first time and r2 in the second time. If the second
party wishes to obtain the i-th secret, for i ∈ {1, 2}, then it says OK if and only if
it has obtained ri but not r3−i . Otherwise, the parties repeat the experiment. Once
the second party says OK, the first party sends it the pair (σ1 ⊕ r1 , σ2 ⊕ r2 ), where
the σ j ’s are the actual secrets.

Exercise 7: Oblivious Transfer implies secure communication protocols: A secure

communication protocol is a two-party protocol that allows the parties to com-
municate in secrecy even when the communication line is tapped by an adversary
(see Exercise 1 of Chapter 5). Show that any 1-out-of-2 Oblivious Transfer (with
security with respect to the semi-honest model) implies the existence of a secure
760

www.Ebook777.com
 7.7 MISCELLANEOUS

communication protocol. Recall that the latter implies the existence of one-way
functions.
Guideline: To transmit a bit σ , the sender invokes the 1-out-of-2 Oblivious Transfer
with input (σ, 0), while the receiver sets its input to 1 and gets σ (i.e., the sender’s
first bit in the OT). Observe that “privacy with respect to the sender” implies that
(the sender and thus also) the adversary cannot distinguish the case where the
receiver enters 1 from the case where it enters 2. Likewise, “privacy with respect
to the receiver” implies that, in the (fictitious) case where the receiver enters 2, the
adversary (like the receiver) cannot tell whether the sender enters (0, 0) or (1, 0).
Thus, also in the (real) case where the receiver enters 1, the adversary cannot tell
whether the sender enters (0, 0) or (1, 0).

Exercise 8: Privately reducing OT to the functionality of Eq. (7.17) – (7.18): Show

that 1-out-of-2 Oblivious Transfer can be privately reduced to the functionality of
Eq. (7.17) – (7.18).
Guideline: Reduce 1-out-of-2 OT to ((s1 , s2 ), (r1 , r2 )) → (λ, s1 r1 + s2 r2 ), and re-
duce the latter to Eq. (7.17) – (7.18).

Exercise 9: Alternative analysis of Construction 7.3.7: The said construction can be de-
coupled into two reductions. First, the functionality of Eq. (7.17) – (7.18) is reduced to
the deterministic functionality ((a1 , b1 , c1 ), (a2 , b2 )) → (λ, f a2 ,b2 (a1 , b1 , c1 )), where
def
f a,b (x, y, z) = z + (x + a) · (y + b), and next the latter is reduced to OT41 . Present
each of these reductions and prove that each is a privacy reduction.
Guideline: When analyzing the second reduction, use the fact that it is used to com-
pute a deterministic functionality and that thus, the simpler form of Definition 7.2.1
can be used.

Exercise 10: Some functionalities that are trivial to privately compute: Show that each
of the following types of functionalities has a trivial protocol for privately computing
it (i.e., using a single message):

1. Each deterministic functionality that only depends on the input of one party (i.e.,
(x, 1|x| ) → ( f 1 (x), f 2 (x)) for arbitrary functions f 1 and f 2 ).
2. Each randomized functionality of the form (x, 1|x| ) → (g(x), f (x, g(x))), where
g is any randomized process and f is a function.

Generalize these functionality types and their treatment to the multi-party case.
Exercise 11: In continuation of Exercise 10, show that all six functionalities introduced
in Section 7.4.3 are trivial to compute in a private manner.
Guideline: Note that the restricted authenticated-computation functionality of
Eq. (7.27) and the image-transmission functionality of Eq. (7.31) fit Item 1,
whereas the basic and augmented coin-tossing functionalities, as well as the input-
commitment functionality, fit Item 2. What about Eq. (7.33)?

761
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Exercise 12: On the difficulty of implementing more natural versions of authenticated

computation: Consider the functionality (α, β) → (λ , v), where v = f (α) if β =
h(α) and v = λ otherwise. We call this functionality the natural auth-comp.

1. Show that the equality functionality (i.e., (α, β) → (χ , χ), where χ = 1 if β = α

and χ = 0 otherwise) is privately reducible to a special case of natural auth-comp.
2. Show that Oblivious Transfer is privately reducible to a special case of natural
auth-comp. Conclude that there can be no trivial protocol for privately computing
the latter (e.g., a protocol that privately computes natural auth-comp implies the
existence of one-way functions).

Guideline (Part 2): Privately reduce the single-secret (bit) version of Oblivious
Transfer to the special case of natural auth-comp in which h(α) (resp., f (α)) equals
the first (resp., second) bit of α. On input a secret bit σ , Party 1 sets its oracle-query
to 1σ and Party 2 sets its query to a uniformly selected bit (and so if the latter equals
h(1σ ) = 1, then Party 2 gets f (1σ ) = σ , and otherwise it gets λ).

Exercise 13: Transitivity of security reductions: Show that if f is securely reducible

to f
, and f
is securely reducible to f

, then f is securely reducible to f

. Note
that Theorem 7.4.3 is obtained as a special case (e.g., by setting f

to be the identity
mapping).
Guideline: See Exercise 5.

Exercise 14: Voting, Elections, and Lottery: Write a specification for some social pro-
cedure (e.g., voting, elections, or lottery), and cast it as a multi-party functional-
ity. Note that allowing appeals and various forms of interaction requires a reactive
functionality (see Section 7.7.1.3), which in turn can be reduced to a standard (non-
reactive) functionality.
Exercise 15: Threshold Cryptography: Loosely speaking, Threshold Cryptography is
concerned with allowing a set of parties to share the ability to perform certain
(cryptographic) operations (cf. [74, 96]). For example, suppose that we wish m
parties to hold shares of a signing-key (with respect to some signature scheme), such
that every t of these parties (but not fewer) can generate signatures to documents of
their choice. Cast this example as a multi-party functionality. (The same holds for
other versions of Threshold Cryptography.)
Exercise 16: Failure of a simple protocol for multi-party authenticated computation:
Consider the m-party oracle-aided protocol for computing Eq. (7.50) in which, for
i = 2, ..., m, Parties 1 and i invoke Eq. (7.33), with Party 1 entering the input α and
Party i entering the input βi . Show that this oracle-aided protocol does not constitute
a secure implementation of Eq. (7.50).
Exercise 17: Analysis of Shamir’s Secret-Sharing Scheme: Prove that Construc-
tion 7.5.35 satisfies the conditions of Definition 7.5.34.
Guideline: For every sequence (u 1 , v1 ), ..., (u  , v ), where the u i ’s are distinct,
consider the set of degree d ≥  − 1 polynomials q that satisfy q(u i ) = vi for

762

www.Ebook777.com
 7.7 MISCELLANEOUS

i = 1, ..., . Denoting the unknown coefficients of q by q j ’s, observe that each

 j
equality q(u i ) = vi yields a linear equation for the q j ’s (i.e., dj=0 u i · q j = vi ).
Furthermore, the equations are linearly independent, and so the solution space has
cardinality p d+1− . Indeed, it is important to consider these equations as referring
to the variables q j ’s and the constants u i ’s, rather than the other way around.

Exercise 18: On the importance of comparable complexity in the definition of per-

fect security: Consider a modification of Definition 7.6.2 such that the ideal-model
adversary is not required to be of comparable complexity to the real-model adversary.
Present protocols that are deemed “secure” under the modified definition although
they are insecure under the original definition (as well as under Definition 7.5.4).

Guideline: Consider any protocol for computing the functionality

(x, 1|x| , ..., 1|x| ) → ( f (x), f (x), ..., f (x)), where f is a one-way permutation.
Show that such a protocol, even the trivial (and bad) one in which Party 1 sends
its input (in the clear) to everybody, is deemed “secure” under the modified defini-
tion. In particular, an ideal-model adversary that does not control (resp., controls)
Party 1 can recover the input of Party 1 (resp., the substituted input of Party 1) in
exponential time by inverting f on any party’s output.

Exercise 19: Perfect security implies ordinary security: Show that Definition 7.6.2
implies Definition 7.5.4.

Guideline: Observe that if A is polynomial-time and B has complexity comparable

to A, then B is polynomial-time.

Exercise 20: Private computation of linear functions: For any fixed m-by-m matrix
M, over a finite field, show that the m-ary functionality x → x M can be m-privately
computed (as per Definition 7.6.1).

Guideline: For starters, consider first the functionality (x1 , ..., x m ) →

m
( i=1 ci x i , 0, ..., 0), where the ci ’s are fixed constants. Show that the following
protocol is m-private: First, each party shares its input with all other parties (by uni-
formly selecting shares that sum up to its input and sending a share to each party).
Next, each party computes the linear combination of the shares it has received.
Finally, each party sends the result to Party 1. Note that this computation would be
t-private if we were to use sharing via a degree t polynomial.

Exercise 21: Private generation of vectors in a linear subspace: For M as in Ex-

ercise 20, show that the m-ary functionality (λ, ..., λ) → r , such that r is a
random m-ary vector satisfying r M = 0m can be m-privately computed (as per
Definition 7.6.1).
Guideline: Consider the generating matrix, denoted G, of the subspace defined by
(the parity-check) matrix M. Suppose that G has rank k (i.e., G is a full-rank k-
by-m matrix). Show that, without loss of generality, the k-by-k left submatrix of G
equals the identity matrix. Privately reduce the generation task to the functionality
of Exercise 20.

763
 Free ebooks ==> www.Ebook777.com
GENERAL CRYPTOGRAPHIC PROTOCOLS

Exercise 22: Alternative presentation of t-private computation of Eq. (7.66):

1. In continuation of Exercises 20 and 21, given M1 and M2 , consider the m-ary
functionality x → x M1 + r , such that r is a random m-ary vector satisfying
r M2 = 0m . Show that this functionality can be m-privately computed.
2. Show that the functionality of Eq. (7.66) is a special case of the class of function-
alities considered in Item 1.
Guideline (Item 1): Note that privately computing each of the two terms of
x M1 + r , separately, will not do. Instead, one has to combine the ideas under-
lying these constructions. Defining G as in Exercise 21, our aim is to privately
compute x → x M1 + sG, where s is a uniformly distributed k-element long se-
quence. Assuming, without loss of generality, that the k-by-k left submatrix of G
is of full rank, we can rewrite the functionality as x → x M
+ sG
, where the k-
by-k left submatrix of G
is the identity matrix and the m-by-k left submatrix of
M
is the all-zero matrix. Recall that we know how to privately compute each of
the two terms of x M
+ sG
, separately; but again this will not do. Instead, we
combine these two computations so as not to yield information about the value of
the individual terms. As a warm-up, consider privately computing the functional-
ity ((x 1 , y1 ), ..., (x m , ym )) → (x 1 , y1 , ..., xm , ym )M, where M is a fixed 2m-by-m
matrix.

Guideline (Item 2): Show that the computation of the free term of the polynomial
c can be captured by an adequate M1 , whereas the generation of the values of a
random degree t polynomial with free-term equal to zero can be captured by an
adequate M2 .

Exercise 23: Analysis of Construction 7.6.3: For t < m/2, show that Construc-
tion 7.6.3 constitutes a protocol that t-privately computes Eq. (7.65).
Guideline: Consider, without loss of generality, I = {1, ..., t}. The simulator is
given an input sequence ((a1 , b1 ), ..., (at , bt )) and an output sequence (r1 , ..., rt ),
and needs to emulate the messages that the parties in I obtain at Step 2. This can
be done by randomly selecting degree t polynomials q
j ’s that are consistent with
these sequences and letting the messages that Party i obtains equal q1
(i), ..., qm
(i).
Specifically, for i = 1, .., t, the polynomial qi
is selected like qi (i.e., uniformly
among the t polynomials having free-term ai bi ); for i = t + 1, .., m − 1, the poly-
nomial qi
is selected uniformly among all t polynomials, and qm
is selected such

that mj=1 γ j q
j (i) = ri holds for all i ∈ [t].

764

www.Ebook777.com
 APPENDIX C

Corrections and Additions

to Volume 1

There is no 100% guarantee in the world;

whoever wants 100% guarantee should not build anything.
Eng. Isidor Goldreich (1906–1995)

In this appendix, we list a few corrections and additions to the previous chapters of this
work (which appeared in [108]).

C.1. Enhanced Trapdoor Permutations

Recall that a collection of trapdoor permutations, as defined in Definition 2.4.5, is a

collection of permutations, { f α }α , accompanied by four probabilistic polynomial-time
algorithms, denoted I, S, F, and B (for index, sample, forward, and backward), such
that the following (syntactic) conditions hold:

1. On input 1n , algorithm I selects a random n-bit long index α of a permutation f α ,

along with a corresponding trapdoor τ ;
2. On input α, algorithm S samples the domain of f α , returning an almost uniformly
distributed element in it;
3. For x in the domain of f α , given α and x, algorithm F returns f α (x) (i.e., F(α, x) =
f α (x));
4. For y in the range of f α if (α, τ ) is a possible output of I (1n ), then, given τ and y,
algorithm B returns f α−1 (y) (i.e., B(τ, y) = f α−1 (y)).

The hardness condition in Definition 2.4.5 refers to the difficulty of inverting f α on a

uniformly distributed element of its range, when given only the range element and α.
That is, let I1 (1n ) denote the first element in the output of I (1n ) (i.e., the index); then for
every probabilistic polynomial-time algorithm A (resp., every non-uniform family of

765
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

polynomial-size circuit A = {An }n ), every positive polynomial p, and all sufficiently

large n’s,
1
Pr[A(I1 (1n ), f I1 (1n ) (S(I1 (1n ))) = S(I1 (1n ))] < (C.1)
p(n)
Namely, A (resp., An ) fails to invert f α on f α (x), where α and x are selected by I and
S as here. An equivalent way of writing Eq. (C.1) is
1
Pr[A(I1 (1n ), S
(I1 (1n ), Rn )) = f I−1
1 (1 )


n (S (I 1 (1 ), Rn ))] <
n
(C.2)
p(n)
where S
is the residual two-input (deterministic) algorithm obtained from S when
treating the coins of the latter as an auxiliary input, and Rn denotes the distribution of
the coins of S on n-bit long inputs. That is, A fails to invert f α on x, where α and x are
selected as earlier.
Although this definition suffices for many applications, in some cases we will need
an enhanced hardness condition. Specifically, we will require it to be hard to invert f α
on a random input x (in the domain of f α ), even when given the coins used by S in the
generation of x. (Note that given these coins (and the index α), the resulting domain
element x is easily determined.)

Definition C.1.1 (enhanced trapdoor permutations): Let { f α : Dα → Dα } be a col-

lection of trapdoor permutations as in Definition 2.4.5. We say that this collection is
enhanced (and call it an enhanced collection of trapdoor permutations) if for ev-
ery probabilistic polynomial-time algorithm A, every positive polynomial p, and all
sufficiently large n’s,
1
Pr[A(I1 (1n ), Rn ) = f I−1
1 (1 )


n (S (I1 (1 ), Rn ))] <
n
(C.3)
p(n)
where S
is as in the foregoing discussion. The non-uniform version is defined
analogously.

We comment that the RSA collection (presented in Section 2.4.3.1 and further discussed
in Section 2.4.4.2) is, in fact, an enhanced collection of trapdoor permutations,1 pro-
vided that RSA is hard to invert in the same sense as assumed in Section 2.4.3.1.
In contrast, the Rabin Collection (as defined in Section 2.4.3) does not satisfy
Definition C.1.1 (because the coins of the sampling algorithm give away a modular
square root of the domain element). Still, the Rabin Collection can be easily mod-
ify to yield an enhanced collection of trapdoor permutations, provided that factor-
ing is hard (in the same sense as assumed in Section 2.4.3). Actually, we present

1 Here and in the following, we assume that sampling Z ∗N , for a composite N , is trivial. However, sampling
Z ∗N (or even Z N ) by using a sequence of unbiased coins is not that trivial. The straightforward sampler may
def
take  = 2log2 N  random bits, view them as an integer in i ∈ {0, 1, ..., 2 − 1}, and output i mod N . This
yields an almost uniform sample in Z N . Also note that given an element e ∈ Z N , one can uniformly sample an
i ∈ {0, 1, ..., 2 − 1} such that i ≡ e (mod N ). Thus, the actual sampler does not cause trouble with respect
to the enhanced hardness requirement.

766

www.Ebook777.com
 C.1 ENHANCED TRAPDOOR PERMUTATIONS

two such possible modifications:

1. Modifying the functions. Rather than squaring modulo the composite N , we consider
the function of raising to the power of 4 modulo N . It can be shown that the resulting
permutations over the quadratic residues modulo N satisfy Definition C.1.1, provided
that factoring is hard. Specifically, given N and a random r ∈ Z N , the ability to
extract the 4th root of r 2 mod N (modulo N ) yields the ability to factor N , where
the algorithm is similar to the one used in order to establish the intractability of
extracting square roots.
2. Changing the domains. Rather than considering the permutation induced (by the
modular squaring function) on the set Q N of the quadratic residues modulo N , we
consider the permutations induced on the set M N , where M N contains all integers in
{1, ..., N /2} that have Jacobi symbol modulo N that equals 1. Note that as in the case
of Q N , each quadratic residue has a unique square root in M N (because exactly two
square roots have a Jacobi symbol that equals 1 and their sum equals N ).2 However,
unlike Q N , membership in M N can be determined in polynomial-time (when given
N without its factorization). Thus, sampling M N can be done in a straightforward
way, which satisfies Definition C.1.1.
Actually, squaring modulo N is a 1-1 mapping of M N to Q N . In order to obtain
a permutation over M N , we modify the function a little, such that if the result of
modular squaring is bigger than N /2, then we use its additive inverse (i.e., rather
than outputting y > N /2, we output N − y).

We comment that the special case of Definition 2.4.5 in which the domain of f α equals
{0, 1}|α| is a special case of Definition C.1.1 (because, without loss of generality, the
sampling algorithm may satisfy S
(α, r ) = r ). Clearly, the RSA and the Rabin collec-
tions can be slightly modified to fit the former special case.

Correction to Volume 1. Theorems 4.10.10, 4.10.14, and 4.10.16 (which in turn

are based on Remark 4.10.6) refer to the existence of certain non-interactive zero-
knowledge proofs. The claimed non-interactive zero-knowledge proof systems can be
constructed by assuming the existence of an enhanced collection of trapdoor permuta-
tions. However, in contrast to the original text, it is not known how to derive these proof
systems based on the existence of a (regular) collection of trapdoor permutations. See
further discussion in Section C.4.1.

Open Problem. Is it possible to convert any collection of trapdoor permutations into

an enhanced one? An affirmative answer will resolve open problems stated in Sec-
tions 7.7.6 and C.4.1, which refer to the assumptions required for General Secure
Multi-Party Computation and various types of Non-Interactive Zero-Knowledge proofs,
respectively.

2 As in the case of Q N , we use the fact that −1 has Jacobi symbol 1.

767
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

C.2. On Variants of Pseudorandom Functions

The focus of Section 3.6 was on a special case of pseudorandom functions, hereafter
referred to as the fixed-length variant. For some function  : N → N (e.g., (n) = n),
these functions map (n)-bit long strings to (n)-bit long strings, where n denotes the
lengths of the function’s seed. More general definitions were presented in Section 3.6.4.
In particular, functions mapping strings of arbitrary length to (n)-bit long strings were
considered. Here, we refer to the latter as the variable-length variant.
A natural question regarding these variants is how to directly (or efficiently) trans-
form functions of the fixed-length variant into functions of the variable-length variant.3
Exercises 30 and 31 in Chapter 3 implicitly suggest such a transformation, and so does
Proposition 6.3.7. Because of the interest in this natural question, we next state the
actual result explicitly.

Proposition C.2.1: Let { f s : {0, 1}(|s|) → {0, 1}(|s|) }s be a (fixed-length) pseudoran-

dom function ensemble, and {h r : {0, 1}∗ → {0, 1}(|r |) }r be a generalized hashing
ensemble with a (t, 1/t)-collision property,4 for some super-polynomial function
t : N → N. Then {gs,r = f s ◦ h r }s,r :|s|=|r | is a (variable-length) pseudorandom func-
tion ensemble.

Proof Idea: The proofs of Propositions 6.3.6 and 6.3.7 actually establish
Proposition C.2.1.

Comment. Alternative constructions of variable-length pseudorandom functions

based on fixed-length pseudorandom functions are presented in [25, 22, 13]. In these
works, the fixed-length pseudorandom functions are applied to each block of the input,
and so the number of applications is linearly related to the input length (rather than
being a single one). On the other hand, these works do not use variable-length hashing.
Indeed, these works presuppose that a fixed-length pseudorandom function (rather than
a variable-length one) is non-expensive (and, in practice, is available as an off-the-shelf
product).

C.3. On Strong Witness Indistinguishability

Unfortunately, we have to withdraw two claims regarding strong witness indistinguish-

able proofs as defined in Definition 4.6.2.5 Specifically, in general, strong witness

3 An indirect construction may use the fixed-length variant in order to obtain a one-way function, and then
construct the variable-length variant using this one-way function. Needless to say, this indirect construction is
very wasteful.
4 Recall that the (t, 1/t)-collision property means that for every n ∈ N and every x = y such that |x|, |y| ≤ t(n),
the probability that h r (x) = h r (y) is at most 1/t(n), where the probability is taken over all possible choices of
r ∈ {0, 1}n with uniform probability distribution.
5 We comment that the notion of strong witness indistinguishability was introduced by the author at a late stage
of writing [108].

768

www.Ebook777.com
 C.3 ON STRONG WITNESS INDISTINGUISHABILITY

indistinguishability is not closed under parallel composition (and so Lemma 4.6.7 is

wrong). Consequently, contrary to what is stated in Theorem 4.6.8, we do not know
whether there exist constant-round public-coin proofs with negligible error that are
strong witness indistinguishable for languages out of BPP.6 Before discussing the
reasons for withdrawing these claims and the consequences of doing so, we stress that
the flaws pointed out here only refer to strong witness indistinguishability and not
to (regular) witness indistinguishability. That is, as stated in Lemma 4.6.6, (regular)
witness indistinguishability is closed under parallel composition, and thus the part of
Theorem 4.6.8 that refers to regular witness indistinguishability is valid (i.e., providing
constant-round public-coin proofs with negligible error that are witness indistinguish-
able for N P).

Notation. To facilitate the rest of the discussion, we let WI stand for “(regular) witness
indistinguishability” and strong-WI stand for “strong witness indistinguishability.”

C.3.1. On Parallel Composition

A counter-example to Lemma 4.6.7 can be derived by using the protocol presented at the
end of Section 4.5.4.1 (and assuming the existence of one-way functions); that is, this
protocol is (zero-knowledge and hence) strong-WI, but executing it twice in parallel (on
the same common input) is not strong-WI. Tracing the error in the reasoning outlined in
Section 4.6.2, we stress a fundamental difference between WI and strong-WI. Under the
former (i.e., under the definition of WI), the indistinguishability of executions, in which
the prover uses one out of two possible NP-witnesses (for the same common input),
holds even when the (adversary) verifier is given these two NP-witnesses. The analogous
claim does not necessarily hold for strong-WI, because these two NP-witnesses (even
presented in random order) may allow for distinguishing one possible common input
from the other (provided that these two possibilities are not identical, unlike in the
case of WI). Now, observe that the single-session adversary constructed in the proof
of Lemma 4.6.6 needs to get the NP-witnesses that correspond to the other sessions
in order to emulate these sessions. However, these other NP-witnesses may determine
the two possible NP-witnesses for the current session, and so the indistinguishability
of the executions of the current session is no longer guaranteed. Furthermore, the other
NP-witnesses may even uniquely determine the NP-witness (or the input triple) used
in the current session. Indeed, the source of trouble is in the possible dependence
among the NP-witnesses used in the various sessions. Consequently, we can resurrect
parallel compositions (of strong-WI) for the special case in which the NP-witnesses
used in the various sessions are independently distributed. Actually, we need statistical
independence among the (entire) input triples used in the various sessions.

Lemma C.3.1 (Parallel Composition for Strong Witness Indistinguishability, Revis-

ited): Let L ∈ N P, R L , (P, V ), Q, R LQ , and PQ be as in Lemma 4.6.6, and suppose

6 Theorem 4.6.8 does not mention the public-coin condition, but the construction that is supposed to support it
is of the public-coin type. Note that constant-round zero-knowledge protocols are presented in Section 4.9, but
these are in relaxed models and are not of the public-coin type.

769
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

that (P, V ) is strong witness indistinguishable. Then for every two probability en-
1 1 1 2 2 2 j j j
sembles {(X n , Y n , Z n )}n∈N and {(X n , Y n , Z n )}n∈N such that X n = (X n,1 , ..., X n, Q(n) ),
j j j j j j j j j
Y n = (Yn,1 , ..., Yn, Q(n) ), and Z n = (Z n,1 , ..., Z n, Q(n) ), where (X n,i , Yn,i , Z n,i ) is inde-
  
pendent of (X n,k , Yn,k , Z n,k )k =i,∈{1,2} , the following holds:

1 1 2 2
If {(X n , Z n )}n∈N and {(X n , Z n )}n∈N are computationally indistinguishable,
1 1 1 2 2 2
then so are { PQ (Y n ), VQ∗ (Z n ) (X n )}n∈N and { PQ (Y n ), VQ∗ (Z n ) (X n )}n∈N ,
for every probabilistic polynomial-time machine VQ∗ .

j j
We stress that the components of Y n (resp., Z n ) may depend on the corresponding
j j
components of X n , but they are independent of the other components of Y n (resp.,
j j
Z n ), as well as of the other components of X n . Note that statistical independence of
this form holds vacuously in Lemma 4.6.6, which refers to fixed sequences of strings.
Lemma C.3.1 is proved by extending the proof of Lemma 4.6.6. Specifically, we consider
hybrids as in the original proof, and construct a verifier V ∗ that interacts with P on the
i-th session (or copy), while emulating all the other sessions (resp., copies). Toward this
j
emulation, we provide V ∗ with the corresponding Q(n) − 1 components of both Y n ’s
j j
(as well as of both X n ’s and Z n ’s). Fixing the best possible choice for these Q(n) − 1
components, we derive a verifier that interacts with P and contradicts the hypothesis
that (P, V ) is strong witness indistinguishable. The key point is that revealing (or fixing)
j
the other Q(n) − 1 components of both Y n ’s does not allow for distinguishing the i-th
1 1 2 2
component of X n and Z n from the i-th component of X n and Z n .

C.3.2. On Theorem 4.6.8 and an Afterthought

Unfortunately, Theorem 4.6.8 is proved by a parallel composition that refers to the same
common input (and the same NP-witness). Thus, Lemma C.3.1 is not applicable, and
consequently we do not know whether the part of Theorem 4.6.8 that refers to strong
witness indistinguishable proofs is valid (when referring to public-coin proofs). This is
indeed an interesting open problem.
We comment that one can reduce the construction of constant-round (public-coin)
strong witness indistinguishable proofs with negligible error for N P to the construction
j j
of such proofs for the special case in which the two X n ’s (and Yn ’s) are identically
j
distributed (and the Z n ’s are only computationally indistinguishable). Consider, for
example, the following protocol:
1. The prover sends a commitment to the value 0.
2. Using a (regular) witness indistinguishable proof (as provided by Theorem 4.6.8),
the prover proves that either the common input is in the language or the string sent
at Step 1 is a commitment to 1.
j
Let us denote by Tn the transcript of the execution of this step, when the common
j j j
input is X n (and the parties use auxiliary inputs Yn and Z n , respectively). It can
770

www.Ebook777.com
 C.3 ON STRONG WITNESS INDISTINGUISHABILITY

j
be proven that the Tn ’s are computationally indistinguishable (by considering what
happens if at Step 1 the prover sends a commitment to 1).

3. Using a strong witness indistinguishable proof (which is indeed the missing compo-
nent or the sub-protocol to which the current protocol is reduced), the prover proves
that the string sent at Step 1 is a commitment to 0.
Note that it suffices to show that the verifier cannot distinguish the two possible
transcript distributions of the current step, where both possible distributions refer
to executions with the same common input (i.e., the commitment) and the same
prover’s auxiliary input (i.e., the decommitment information). In contrast, these
two distributions (of executions) refer to two different distributions of the verifier’s
auxiliary input (i.e., either Tn1 or Tn2 ), which are indistinguishable.
The foregoing reduction demonstrates that the notion of strong witness indistinguisha-
bility actually refers to issues that are fundamentally different from witness indistin-
guishability. Specifically, the issue is whether or not the interaction with the prover helps
to distinguish between two possible distributions of some auxiliary information (which
are indistinguishable without such an interaction). Furthermore, this issue arises also
in case the prover’s auxiliary inputs (i.e., the “witnesses”) are identically distributed.

C.3.3. Consequences
In view of the fact that we do not have constant-round public-coin strong witness
indistinguishable proofs with negligible error for N P, we suggest replacing the use of
such proofs with some cumbersome patches. A typical example is the construction of
non-oblivious commitment schemes (i.e., Theorem 4.9.4).

Non-Oblivious Commitment Schemes. We begin the discussion by noting that the

specific formulation appearing in Definition 4.9.3 is wrong. One should partition the
commit phase into two sub-phases, such that the second sub-phase is a proof-of-
knowledge of the input and coins used by the sender at the first sub-phase, which
in turn should constitute (by itself ) a commitment scheme. That is, the view in the rela-
tion displayed in Definition 4.9.3 should be the view of the first sub-phase (rather than
the view of the entire commit phase). In fact, for the current implementation, we need a
relaxed definition in which one only proves knowledge of the input (but not of the coins)
used by the sender at the first sub-phase. We stress that the input value proved to be
known must be such that it is impossible for the sender to later decommit to a different
value. Indeed, in the relaxed form, we do not require a later decommitment to be at all
possible; we only require that if decommitment takes place, then the outcome should
match the said input value. Note that this relaxed form suffices for the proof presented in
Section 4.9.2.2.
Next, we modify the construction used in the proof of Theorem 4.9.4 as follows.
First, rather than sending one ordinary commitment to the input, we send many such
(independent) commitments. Secondly, rather than using a (constant-round) proof-of-
knowledge with negligible error, we use one that has constant error. The point is that
771
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

such a (constant-round) proof-of-knowledge that is zero-knowledge (and, hence, strong

witness indistinguishable) is known. We invoke this proof system many times, in par-
allel, where each invocation is applied to a different commitment. Thus, we can apply
Lemma C.3.1 and conclude that these executions are strong witness indistinguishable
(where the witnesses are the coins used in the ordinary commitments), and therefore,
the entire protocol constitutes a (complicated) commitment scheme. Finally, one can
establish the non-oblivious property by using the knowledge-extractor associated with
the proof system. Note that we can only extract the committed input and part of the
coins used at the first stage (i.e., the coins used in some of the ordinary commitments
but not necessarily the coins used in all of them). Furthermore, it may be that we also
accept in case the sequence of strings sent at the first stage does not correspond to any
legitimate sequence (i.e., of commitments to the same value). However, if we extract
one value, then it is impossible for the sender to later decommit to a different value,
because the extracted value always fits at least one of the individual commitments.

Other Applications. Fortunately, Theorem 4.9.4 is the only place where strong witness
indistinguishable proofs are used in this work. We believe that in many other applications
of strong witness indistinguishable proofs, an analogous modification can be carried
out (in order to salvage the application). A typical example appears in [7]. Indeed, the
current situation is very unfortunate, and we hope that it will be redeemed in the future.
Specifically, we propose the following open problem:

Open Problem. Construct constant-round public-coin strong witness indistinguish-

able proofs (and proofs-of-knowledge) with negligible error for N P, or prove that this
cannot be done. Recall that zero-knowledge arguments of this nature are known [5].
The challenge is in providing such proofs.

C.4. On Non-Interactive Zero-Knowledge

In retrospect, it appears that Section 4.10 is too laconic. As is usually the case, laconic
style gives rise to inaccuracies and gaps, which we wish to address here. (See also
Section C.6.)

C.4.1. On NIZKs with Efficient Prover Strategies

In continuation of Remark 4.10.6 and following [32], we briefly discuss the issues that
arise when we wish to implement Construction 4.10.4 by an efficient prover. Recall that
Remark 4.10.6 outlines such an implementation, while using a family of trapdoor per-
mutations of the form { f α : {0, 1}|α| → {0, 1}|α| }α∈I , where the index-set I is efficiently
recognizable. Unfortunately, no family of trapdoor permutations of this particular form
(and, in particular, with an efficiently recognizable I ) is known. Thus, we first extend
the treatment to the case in which I is not necessarily efficiently recognizable. The
problem we encounter is that the prover may select (and send) a function that is not in

772

www.Ebook777.com
 C.4 ON NON-INTERACTIVE ZERO-KNOWLEDGE

the family (i.e., an α not in I ). In such a case, the function is not necessarily 1-1, and,
consequently, the soundness property may be violated. This concern can be addressed
by using a (simple) non-interactive (zero-knowledge) proof for establishing that the
function is “typically 1-1” (or, equivalently, is “almost onto the designated range”).
The proof proceeds by presenting pre-images (under the function) of random elements
specified in the reference string. Note that for any fixed polynomial p, we can only
prove that the function is 1-1 on at least a 1 − (1/ p(n)) fraction of the designated range
(i.e., {0, 1}n ), yet this suffices for moderate soundness of the entire proof system (which
in turn can be amplified by repetitions). For further details, consult [32].
Although the known candidate trapdoor permutations can be modified to fit this form,
we wish to further generalize the result such that any enhanced trapdoor permutation
(as in Definition C.1.1) can be used. This can be done by letting the reference string
consist of the coin sequences used by the domain-sampling algorithm (rather than of
elements of the function’s domain). By virtue of the enhanced hardness condition (i.e.,
Eq. (C.3)), the security of the hard-core is preserved, and so is the zero-knowledge
property.
As stated at the end of Section C.1, in contrast to what was claimed in Remark 4.10.6,
we do not known how to extend the construction to arbitrary (rather than enhanced)
trapdoor permutations. This leads to the following open problem.

Open Problem. Under what intractability assumptions is it possible to construct non-

interactive zero-knowledge proofs with efficient prover strategies for any set in N P? In
particular, does the existence of arbitrary collections of trapdoor permutations suffice?
We comment that the assumption used here affects the assumption used in (general)
constructions of public-key encryption schemes that are secure under chosen ciphertext
attacks (see, e.g., Theorem 5.4.31).

C.4.2. On Unbounded NIZKs

The preliminary discussion is Section 4.10.3.1 reduces the general treatment to a treat-
ment of assertions of a priori bounded length, but the former is not defined formally.
To close this gap, we note that a definition that covers assertions of a priori unbounded
poly(n)
length can be derived from Definition 4.10.11 by considering inputs in ∪i=1 L i , rather
than in L n ε . In view of the key role of efficient provers in this setting, it is also ade-
quate to present a definition that covers this aspect. This can be done analogously to
the formulations used in the following Proposition C.4.1.
The proof of Proposition 4.10.13 relies on the fact that witness indistinguishability
of non-interactive protocols is preserved under parallel composition even if the same
reference string is used in all copies. That is, we claim and use the following result
(where R is typically an NP-relation):

Proposition C.4.1: Let P be a probabilistic polynomial-time algorithm such that

def
for every infinite sequence of triples of the form t = (x, u, v), where (x, u), (x, v) ∈
R, it holds that {(Upoly(|x|) , P(x, u, Upoly(|x|) ))}t and {(Upoly(|x|) , P(x, v, Upoly(|x|) ))}t

773
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

are computationally indistinguishable.7 Then for every polynomial p and ev-

def
ery infinite sequence of sequences of the form s = (x1 , ..., x t , u 1 , ..., u t , v1 , ..., vt ),
def def
where n = |x 1 | = · · · = |x t |, t = p(n) and (x j , u j ), (x j , v j ) ∈ R for j = 1, . . , t,
it holds that the ensembles {(Upoly(n) , P(x1 , u 1 , Upoly(n) ), ..., P(xt , u t , Upoly(n) ))}s
and {(Upoly(n) , P(x1 , v1 , Upoly(n) ), ..., P(xt , vt , Upoly(n) ))}s are computationally indistin-
guishable.

We stress that the same reference string (i.e., Upoly(n) ) is used in all invocations of the
prover P. Thus, Proposition C.4.1 does not refer to multiple samples of computationally
indistinguishable ensembles (nor even to independent samples from a sequence of
computationally indistinguishable pairs of ensembles, as would have been the case if
the various invocations were to use independently distributed reference strings). Still,
Proposition C.4.1 can be established by using the hybrid technique. The key observation
is that, given a single proof with respect to some reference string along with the reference
string (as well as the relevant sequence s), one can efficiently generate all the other proofs
(with respect to the same reference string). Indeed, the internal coins used by P in each
of these proofs are independent.

C.4.3. On Adaptive NIZKs

In Definition 4.10.15, the adaptive zero-knowledge condition should be quantified only
over efficiently computable input-selection strategies. Furthermore, it seems that also
the witness-selection strategies should be restricted to ones implemented by polynomial-
size circuits. The revised form is presented in Definition 5.4.22.
A few words regarding the proof of Theorem 4.10.16 seem appropriate. The (two-
stage) simulation procedure itself is sketched in footnote 29 (of Chapter 4). Recall that
at the first stage, we generate matrices at random, and replace the useful matrices with
all-zero matrices (i.e., matrices of f -images that have pre-images with hard-core value
equal to zero). In the second stage, when given an adaptively chosen graph, we reveal
all elements of all non-useful matrices and the required elements of the useful matrices
(i.e., the non-edges), where revealing an element means revealing the corresponding
f -pre-image. In establishing the quality of this simulation procedure, we rely on the
hypothesis that the input graph, as well as a Hamiltonian cycle in it, are determined
by a polynomial-size circuit.8 Loosely speaking, assuming toward the contradiction
that the simulation can be distinguished from the real proof, we construct a circuit that
distinguishes a sequence of random f (x)’s with b(x) = 0 from a sequence of random
f (x)’s with b(x) = 1. This “b-value distinguisher” places the tested f -images in the
suitable entries (i.e., those corresponding to the predetermined Hamiltonian cycles) of
useful matrices, fills up the rest of the entries of the useful matrices with elements it
generates in { f (x) : b(x) = 0}, and fills the entries of non-useful matrices with random
f -images that it generates (conditioned on their yielding non-useful matrices). We stress

7 Recall that the distinguisher is also given the index of the distribution, which in this case is the triple t.
8 Indeed, here is where we use the fact that the corrected definition (see Definition 5.4.22) refers only to input-
selection and witness-selection strategies that can be implemented by polynomial-size circuits.

774

www.Ebook777.com
 C.5 SOME DEVELOPMENTS REGARDING ZERO-KNOWLEDGE

that the simulator generates f -images by selecting random pre-images and applying
f to each of them, and so it knows the pre-images and can reveal them later. Next,
the simulator determines the input graph and the corresponding Hamiltonian cycle (by
using the abovementioned polynomial-size circuit) and acts as the real prover. Finally,
it feeds the original distinguisher with the corresponding output. Observe that in case
the given sequence of f (x)’s satisfies b(x) = 0 (resp., b(x) = 1) for each f (x), the
“b-value distinguisher” produces outputs distributed exactly as in the simulation (resp.,
the real proof ).

C.5. Some Developments Regarding Zero-Knowledge

A recent result by Barak [5] calls for reevaluation of the significance of all nega-
tive results regarding black-box zero-knowledge9 (as defined in Definition 4.5.10).
In particular, relying on standard intractability assumptions, Barak presents round-
efficient public-coin zero-knowledge arguments for N P (using non-black-box simu-
lators), whereas only BPP can have such black-box zero-knowledge arguments (see
comment following Theorem 4.5.11). It is interesting to note that Barak’s simulator
works in strict (rather than expected) probabilistic polynomial-time, addressing an open
problem mentioned in Section 4.12.3. Barak’s result is further described in Section C.5.2
In Section C.5.1, we review some recent progress in the study of the preservation of
zero-knowledge under concurrent composition. We seize the opportunity to provide a
wider perspective on the question of the preservation of zero-knowledge under various
forms of protocol composition operations.
We mention that the two problems discussed in this section (i.e., the “preservation of
security under various forms of protocol composition” and the “use of the adversary’s
program within the proof of security”) arise also with respect to the security of other
cryptographic primitives. Thus, the study of zero-knowledge protocols serves as a good
benchmark for the study of various problems regarding cryptographic protocols.

C.5.1. Composing Zero-Knowledge Protocols

A natural question regarding zero-knowledge proofs (and arguments) is whether or not
the zero-knowledge condition is preserved under a variety of composition operations.
Three types of composition operation were considered in the literature: sequential com-
position, parallel composition, and concurrent composition. We note that the preserva-
tion of zero-knowledge under these forms of composition not only is interesting for its
own sake but also sheds light on the preservation of the security of general protocols
under these forms of composition.
We stress that when we talk of the composition of protocols (or proof systems), we
mean that the honest users are supposed to follow the prescribed program (specified
in the protocol description) that refers to a single execution. That is, the actions of

9 Specifically, one should reject the interpretation, offered in Section 4.5 (see Sections 4.5.0, 4.5.4.0, and 4.5.4.2),
by which negative results regarding black-box zero-knowledge indicate the inherent limitations of zero-
knowledge.

775
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

honest parties in each execution are independent of the messages they received in other
executions. The adversary, however, may coordinate the actions it takes in the various
executions, and in particular, its actions in one execution may also depend on messages
it received in other executions.
Let us motivate the asymmetry between the postulate that honest parties act inde-
pendently in different executions and the absence of such an assumption with respect
to the adversary’s actions. Typically, coordinating actions in different executions is dif-
ficult but not impossible. Thus, it is desirable to use stand-alone protocols that preserve
security under “composition” (as defined earlier), rather than to use protocols that in-
clude inter-execution coordination actions. Note that at the very least, inter-execution
coordination requires users to keep track of all executions that they perform. Actually,
trying to coordinate honest executions is even more problematic than it seems, because
one may need to coordinate executions of different honest parties (e.g., all employees of
a big corporation or an agency under attack), which in many cases is highly unrealistic.
On the other hand, the adversary attacking the system may be willing to go to the extra
trouble of coordinating its attack in the various executions of the protocol.
For T ∈ {sequential, parallel, concurrent}, we say that a protocol is T -zero-
knowledge if it is zero-knowledge under a composition of type T . The definitions of
T -zero-knowledge are derived from the standard definition by considering appropriate
adversaries (i.e., adversarial verifiers), that is, adversaries that can initiate a polynomial
number of interactions with the prover, where these interactions are scheduled according
to the type T .10 The corresponding simulator (which, as usual, interacts with nobody) is
required to produce an output that is computationally indistinguishable from the output
of such a type T adversary.

C.5.1.1. Sequential Composition

Sequential composition refers to a situation in which the protocol is invoked (polyno-
mially) many times, where each invocation follows the termination of the previous one.
At the very least, security (e.g., zero-knowledge) should be preserved under sequential
composition, or else the applicability of the protocol is highly limited (because one
cannot safely use it more than once).
We mention that whereas the “simplified” version of zero-knowledge (i.e., without
auxiliary inputs, as in Definition 4.3.2) is not closed under sequential composition
(see [113]), the actual version (i.e., with auxiliary inputs, as in Definition 4.3.10) is
closed under sequential composition (see Section 4.3.4). We comment that the same
phenomenon arises when trying to use a zero-knowledge proof as a sub-protocol inside
larger protocols. Indeed, it is for these reasons that the augmentation of the “basic”
definition by auxiliary inputs was adopted in all subsequent works.11

10 Without loss of generality, we may assume that the adversary never violates the scheduling condition; it may
instead send an illegal message at the latest possible adequate time. Furthermore, without loss of generality, we
may assume that all the adversary’s messages are delivered at the latest possible adequate time.
11 The preliminary version of Goldwasser, Micali, and Rackoff ’s work [124] uses the “basic” definition (i.e.,

Definition 4.3.2), whereas the final version of that work as well as most subsequent works use the augmented

776

www.Ebook777.com
 C.5 SOME DEVELOPMENTS REGARDING ZERO-KNOWLEDGE

C.5.1.2. Parallel Composition

Parallel composition refers to a situation in which (polynomially) many instances of
the protocol are invoked at the same time and proceed at the same pace. That is, we
assume a synchronous model of communication, and consider (polynomially) many
executions that are totally synchronized, such that the i-th message in all instances is
sent exactly (or approximately) at the same time. (Natural extensions of this model are
discussed here as well as at the end of Section C.5.1.3.)
It turns out that, in general, zero-knowledge is not closed under parallel composition.
A simple counter-example (to the “parallel composition conjecture”) is outlined in
Section 4.5.4.1 (following [113]). This counter-example consists of a simple protocol
that is zero-knowledge (in a strong sense) but is not closed under parallel composition
(not even in a very weak sense).12
We comment that in the 1980s, parallel composition was studied mainly in the
context of round-efficient error reduction (cf. [91, 113]); that is, the aim was to construct
full-fledged zero-knowledge proofs (with negligible soundness error) by composing (in
parallel) a basic zero-knowledge protocol of high (but bounded away from 1) soundness
error. Since alternative ways of constructing constant-round zero-knowledge proofs
(and arguments) were found (cf. [112, 90, 47]), interest in parallel composition (of zero-
knowledge protocols) has died. In retrospect, this was a conceptual mistake, because
parallel composition (and mild extensions of this notion) capture the preservation of
security in a fully synchronous (or almost fully synchronous) communication network.
We note that the almost fully synchronous communication model is quite realistic in
many settings, although it is certainly preferable not to assume even weak synchronism.
Although, in general, zero-knowledge is not closed under parallel composition, under
standard intractability assumptions (e.g., the intractability of factoring), there exist zero-
knowledge protocols for N P that are closed under parallel composition. Furthermore,
these protocols have a constant number of rounds (cf. [109] for proofs and [82] for
arguments).13 Both results also extend to concurrent composition in a synchronous
communication model, where the extension is in allowing protocol invocations to start at
different times (and, in particular, executions may overlap but not run simultaneously).
We comment that parallel composition is also problematic in the context of reducing
the soundness error of arguments (cf. [24]), but our focus here is on the zero-knowledge
aspect of protocols, regardless of whether they are proofs, arguments, or neither.

C.5.1.3. Concurrent Composition (with and without Timing)

Concurrent composition generalizes both sequential and parallel composition. Here
(polynomially) many instances of the protocol are invoked at arbitrary times and proceed

definition (i.e., Definition 4.3.10). In some works, the “basic” definition is used for simplicity, but typically one
actually needs and means the augmented definition.
12 The presentation in Section 4.5.4.1 is in terms of two protocols, each being zero-knowledge, such that executing

them in parallel is not zero-knowledge. These two protocols can be easily combined into one protocol (e.g., by
letting the second party determine, in its first message, which of the two protocols to execute).
13 In the case of parallel zero-knowledge proofs, there is no need to specify the soundness error because it can

always be reduced via parallel composition. As mentioned later, this is not the case with respect to arguments.

777
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

at an arbitrary pace. That is, we assume an asynchronous (rather than synchronous)

model of communication.
In the 1990s, when extensive two-party (and multi-party) computations became a
reality (rather than a vision), it became clear that it is (at least) desirable that crypto-
graphic protocols maintain their security under concurrent composition (cf. [77]). In
the context of zero-knowledge, concurrent composition was first considered by Dwork,
Naor, and Sahai [82]. Actually, two models of concurrent composition were considered
in the literature, depending on the underlying model of communication (i.e., a purely
asynchronous model and an asynchronous model with timing).

Concurrent Composition in the Pure Asynchronous Model. Here we refer to the

standard model of asynchronous communication. In comparison to the timing model,
the pure asynchronous model is a simpler model, and using it requires no assumptions
about the underlying communication channels. However, it seems harder to construct
concurrent zero-knowledge protocols for this model. In particular, for a while it was not
known whether concurrent zero-knowledge proofs for N P exist at all (in this model).
Under standard intractability assumptions (e.g., the intractability of factoring), this
question was affirmatively resolved by Richardson and Kilian [175]. Following their
work, research has focused on determining the round-complexity of concurrent zero-
knowledge proofs for N P. Currently, this question is still open, and the state of the art
regarding it is as follows:

r Under standard intractability assumptions, every language in N P has a concurrent

zero-knowledge proof with almost logarithmically many rounds (cf. [169], building
upon [138], which in turn builds over [175]). Furthermore, the zero-knowledge
property can be demonstrated by using a black-box simulator (see the definition in
Section 4.5.4.2 and the discussion in Section C.5.2).
r Black-box simulators cannot demonstrate the concurrent zero-knowledge property
of non-trivial proofs (or arguments) having significantly less than logarithmically
many rounds (cf. Canetti et al. [58]).14
r Recently, Barak [5] demonstrated that the “black-box simulation barrier” can be
bypassed. With respect to concurrent zero-knowledge, he obtained only the following
partial result: Under standard intractability assumptions, every language in N P
has a constant-round zero-knowledge argument (rather than proof) that maintains
security as long as an a priori bounded (polynomial) number of executions take place
concurrently. (The length of the messages in his protocol grows linearly with this a
priori bound.)

Thus, it is currently unknown whether or not constant-round arguments for N P may

be concurrent zero-knowledge (in the pure asynchronous model).

14 By non-trivial proof systems we mean ones for languages outside BPP, whereas by significantly less than loga-
rithmic we mean any function f : N → N satisfying f (n) = o(log n/ log log n). In contrast, by almost logarithmic
we mean any function f satisfying f (n) = ω(log n).

778

www.Ebook777.com
 C.5 SOME DEVELOPMENTS REGARDING ZERO-KNOWLEDGE

Concurrent Composition under the Timing Model. A model of naturally limited

asynchronousness (which certainly covers the case of parallel composition) was intro-
duced by Dwork, Naor, and Sahai [82]. Essentially, they assume that each party holds a
local clock such that the relative clock rates are bounded by an a priori known constant,
and they consider protocols that employ time-driven operations (i.e., time-out incom-
ing messages and delay outgoing messages). The benefit of the timing model is that
it seems easier to construct concurrent zero-knowledge protocols for it. Specifically,
using standard intractability assumptions, constant-round arguments and proofs that
are concurrent zero-knowledge under the timing model do exist (cf. [82] and [109],
respectively). The disadvantages of the timing model are discussed next.
The timing model consists of the assumption that talking about the actual timing of
events is meaningful (at least in a weak sense) and of the introduction of time-driven
operations. The timing assumption amounts to postulating that each party holds a local
clock and knows a global bound, denoted ρ ≥ 1, on the relative rates of the local
clocks.15 Furthermore, it is postulated that the parties know a (pessimistic) bound,
denoted , on the message-delivery time (which also includes the local computation
and handling times). In our opinion, these timing assumptions are most reasonable, and
are unlikely to restrict the scope of applications for which concurrent zero-knowledge
is relevant. We are more concerned about the effect of the time-driven operations
introduced in the timing model. Recall that these operations are the time-out of
incoming messages and the delay of outgoing messages. Furthermore, typically the
delay period is at least as long as the time-out period, which in turn is at least  (i.e.,
the time-out period must be at least as long as the pessimistic bound on message-
delivery time so as not to disrupt the proper operation of the protocol). This means
that the use of these time-driven operations yields a slowing down of the execution of
the protocol (i.e., running it at the rate of the pessimistic message-delivery time, rather
than at the rate of the actual message-delivery time, which is typically much faster).
Still, in the absence of more appealing alternatives (i.e., a constant-round concurrent
zero-knowledge protocol for the pure asynchronous model), the use of the timing model
may be considered reasonable. (We comment than other alternatives to the timing model
include various set-up assumptions; cf. [55, 72].)

Back to Parallel Composition. Given our opinion about the timing model, it is not
surprising that we consider the problem of parallel composition almost as important as
the problem of concurrent composition in the timing model. Firstly, it is quite reasonable
to assume that the parties’ local clocks have approximately the same rate, and that
drifting is corrected by occasional clock synchronization. Thus, it is reasonable to
assume that the parties have an approximately good estimate of some global time.
Furthermore, the global time may be partitioned into phases, each consisting of a
constant number of rounds, so that each party wishing to execute the protocol just
delays its invocation to the beginning of the next phase. Thus, concurrent execution

15 The rate should be computed with respect to reasonable intervals of time; for example, for  as defined next,
one may assume that a time period of  units is measured as 
units of time on the local clock, where
/ρ ≤ 
≤ ρ.

779
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

of (constant-round) protocols in this setting amounts to a sequence of (time-disjoint)

almost parallel executions of the protocol. Consequently, proving that the protocol is
parallel zero-knowledge suffices for concurrent composition in this setting.

Relation to Resettable Zero-Knowledge. Going to the other extreme, we mention that

there exists a natural model of zero-knowledge that is even stronger than concurrent
zero-knowledge (even in the pure asynchronous model). Specifically, “resettable zero-
knowledge” as defined in [55] implies concurrent zero-knowledge.

C.5.2. Using the Adversary’s Program in the Proof of Security

Recall that the definition of zero-knowledge proofs states that whatever an efficient
adversary can compute after interacting with the prover can be efficiently computed
from scratch by a so-called simulator (which works without interacting with the prover).
Although the simulator may depend arbitrarily on the adversary, the need to present a
simulator for each feasible adversary seems to require the presentation of a universal
simulator that is given the adversary’s strategy (or program) as another auxiliary in-
put. The question addressed in this section is how the universal simulator can use the
adversary’s program.
The adversary’s program (or strategy) is actually a function that determines for each
possible view of the adversary (i.e., its input, random choices, and the message it has
received so far) which message will be sent next. Thus, we identify the adversary’s
program with this next-message function. As stated previously, until very recently,
all universal simulators (constructed toward demonstrating zero-knowledge properties)
have used the adversary’s program (or rather its next-message function) as a black-box
(i.e., the simulator invoked the next-message function on a sequence of arguments of
its choice). Furthermore, in view of the presumed difficulty of “reverse-engineering”
programs, it was commonly believed that nothing is lost by restricting attention to simu-
lators, called black-box simulators, that only make black-box usage of the adversary’s
program. Consequently, Goldreich and Krawczyk conjectured that impossibility results
regarding black-box simulation represent inherent limitations of zero-knowledge itself,
and studied the limitations of the former [113].

In particular, they showed that parallel composition of the protocol of

Construction 4.4.7 (as well as of any constant-round public-coin protocol) cannot
be proven to be zero-knowledge using a black-box simulator, unless the language
(i.e., 3-Colorability) is in BPP. In fact, their result refers to any constant-round
public-coin protocol with negligible soundness error, regardless of how such a pro-
tocol is obtained. This result was taken as strong evidence toward the conjecture
that a constant-round public-coin protocol with negligible soundness error cannot
be zero-knowledge (unless the language is in BPP).

Similarly, as mentioned in Section C.5.1.3, it was shown that protocols of a sub-

logarithmic number of rounds cannot be proven to be concurrent zero-knowledge via
a black-box simulator [58]. Again, this was taken as evidence toward the conjecture
that such protocols cannot be concurrent zero-knowledge.

780

www.Ebook777.com
 C.5 SOME DEVELOPMENTS REGARDING ZERO-KNOWLEDGE

In contrast to these conjectures (and to the reasoning underlying them), Barak showed
how to construct non-black-box simulators and obtained several results that were known
to be unachievable via black-box simulators [5]. In particular, under standard intractabil-
ity assumptions (see also [7]), he presented constant-round public-coin zero-knowledge
arguments with negligible soundness error for any language in N P. (Moreover, the
simulator runs in strict polynomial-time, which is impossible for black-box simula-
tors of non-trivial constant-round protocols [9].) Furthermore, these protocols pre-
serve zero-knowledge under a fixed16 polynomial number of concurrent executions,
in contrast to the result of [58] (regarding black-box simulators) that also holds in
that restricted case. Thus, Barak’s result calls for the reevaluation of many common
beliefs. Most concretely, it says that results regarding black-box simulators do not
reflect inherent limitations of zero-knowledge (but rather an inherent limitation of a
natural way of demonstrating the zero-knowledge property). Most abstractly, it says
that there are meaningful ways of using a program other than merely invoking it as a
black-box.
Does this means that a method was found to “reverse-engineer” programs or to
“understand” them? We believe that the answer is negative. Barak [5] is using the
adversary’s program in a significant way (i.e., more significant than just invoking it),
without “understanding” it. So, how does he use the program?
The key idea underlying Barak’s protocol [5] is to have the prover prove that either the
original NP-assertion is valid or that he (i.e., the prover) “knows the verifier’s residual
strategy” (in the sense that it can predict the next verifier message). Indeed, in a real
interaction (with the honest verifier), it is infeasible for the prover to predict the next
verifier message, and so computational soundness of the protocol follows. However,
a simulator that is given the code of the verifier’s strategy (and not merely oracle
access to that code) can produce a valid proof of the disjunction by properly executing
the sub-protocol using its knowledge of an NP-witness for the second disjunctive. The
simulation is computational indistinguishable from the real execution, provided that one
cannot distinguish an execution of the sub-protocol in which one NP-witness (i.e., an
NP-witness for the original assertion) is used from an execution in which the second NP-
witness (i.e., an NP-witness for the auxiliary assertion) is used. That is, the sub-protocol
should be a witness indistinguishable argument system (see Sections 4.6 and 4.8). We
warn the reader that the actual implementation of this idea requires overcoming several
technical difficulties (cf. [5, 7]).

Perspective. In retrospect, taking a wide perspective, it should not come as a surprise

that the program’s code yields extra power beyond black-box access to it. Feeding a
program with its own code (or part of it) is the essence of the diagonalization technique,
and this, too, is done without reverse engineering. Furthermore, various non-black-box
techniques have appeared before in the cryptographic setting, but they were used in the
more natural context of devising an attack on an (artificial) insecure scheme (e.g., toward

16 The protocol depends on the polynomial that bounds the number of executions, and thus is not known to be
concurrent zero-knowledge (because the latter requires fixing the protocol and then considering any polynomial
number of concurrent executions).

781
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

proving the failure of the “Random Oracle Methodology” [54] and the impossibility of
software obfuscation [8]). In contrast, in [5] (and [6]), the code of the adversary is being
used within a sophisticated proof of security. What we wish to highlight here is that
non-black-box usage of programs is also relevant to proving (rather than to disproving)
the security of systems.

Digest: Witness Indistinguishability and the FLS-Technique

The foregoing description (of [5]), as well as several other sophisticated constructions of
zero-knowledge protocols (e.g., [89, 175]), make crucial use of a technique introduced
by Feige, Lapidot, and Shamir [89], which in turn is based on the notion of witness
indistinguishability (introduced by Feige and Shamir [91]). This technique, hereafter
referred to as the FLS-technique, was used in Construction 4.10.12, but we wish to
further discuss it next.
Following is a sketchy description of a special case of the FLS-technique, whereas
the abovementioned application uses a more general version (which refers to proofs-of-
knowledge, as defined in Section 4.7).17 In this special case, the technique consists of
the following construction schema, which uses witness indistinguishable protocols for
N P in order to obtain zero-knowledge protocols for N P. On common input x ∈ L,
where L is the NP-set defined by the witness relation R, the following two steps are
performed:

1. The parties generate an instance x
for an auxiliary NP-set L
, where L
is defined

by a witness relation R
. The generation protocol in use must satisfy the following
two conditions:
(a) If the verifier follows its prescribed strategy, then no matter which feasible
strategy is used by the prover, with high probability, the protocol’s outcome is a
no-instance of L
.
(b) There exists an efficient (non-interactive) procedure for producing a (random)
transcript of the generation protocol along with an NP-witness for the corre-
sponding outcome (which is a yes-instance of L
), such that the produced tran-
script is computationally indistinguishable from the transcript of a real execution
of the protocol.
2. The parties execute a witness indistinguishable protocol for the set L

defined
by the witness relation R

= {((u, u
), (v, v
)) : (u, v) ∈ R ∨ (u
, v
) ∈ R
}. The sub-
protocol is such that the corresponding prover can be implemented in probabilistic
polynomial-time, given an NP-witness for (u, u
) ∈ L

. The sub-protocol is invoked
on common input (x, x
), where x
is the outcome of Step 1, and the sub-prover

17 In the general case, the generation protocol may generate an instance x
in L
, but it is infeasible for
the prover to obtain a corresponding witness (i.e., a w
such that (x
, w
) ∈ R
). In the second step,
the sub-protocol in use ought to be a proof-of-knowledge, and computational soundness of the main
protocol will follow (because otherwise, the prover, using a knowledge-extractor, can obtain a witness
for x
∈ L
).

782

www.Ebook777.com
 C.6 ADDITIONAL CORRECTIONS AND COMMENTS

is invoked with the corresponding NP-witness as auxiliary input (i.e., with (w, λ),
where w is the NP-witness for x given to the main prover).
The computational soundness of this protocol follows by Property (a) of the gen-
eration protocol (i.e., with high probability x
∈ L
, and so x ∈ L follows by the
soundness of the protocol used in Step 2). To demonstrate the zero-knowledge prop-
erty, we first generate a simulated transcript of Step 1 (with outcome x
∈ L
), along
with an adequate NP-witness (i.e., w
such that (x
, w
) ∈ R
), and then emulate
Step 2 by feeding the sub-prover strategy with the NP-witness (λ, w
). Combining
Property (b) of the generation protocol and the witness indistinguishability prop-
erty of the protocol used in Step 2, the simulation is indistinguishable from the real
execution.

C.6. Additional Corrections and Comments

Regarding Constriction 4.10.7 and the Proof of Proposition 4.10.9. The current
description of the setting of the mapping of the input graph G to the Hamiltonian matrix
H (via the two mappings π1 and π2 ) is confusing and even inaccurate. Instead, one may
identify the rows (resp., columns) of H with [n] and use a single permutation π over
[n] (which supposedly maps the vertices of G to those of H ).18 Alternatively, one may
compose this permutation π with the two (1-1) mappings φi ’s (where φi : [n] → [n 3 ]
is as in the original text), and obtain related πi ’s (i.e., πi (v) = φi (π(v))), which should
be used as in the original text. We stress that the real prover determines π to be an
isomorphism between the Hamiltonian cycle of G and the Hamiltonian cycle of H ,
whereas the simulator selects π at random.

Arguments-of-Knowledge. In continuation of Sections 4.7 and 4.9.2, we mention that

the round-efficient argument system of [90] is actually an “argument-of-knowledge”
(with negligible error). The interested reader is referred to [9] for further improvements
regarding such proof systems. Essentially, using a relaxed (yet satisfactory) definition of
an argument-of-knowledge, the latter work presents a constant-round zero-knowledge
argument-of-knowledge with strict (rather than expected) probabilistic polynomial-
time simulator and knowledge-extractor.

Some Missing Credits. The sequential composition lemma for zero-knowledge pro-
tocols (i.e., Lemma 4.3.11) is due to [119]. The notions of strong witness indistin-
guishability (Definition 4.6.2) and strong proofs-of-knowledge (Section 4.7.6), and
the Hidden Bit Model (Section 4.10.2) have first appeared in early versions of this
work.

18 The identification is via the two mappings φ1 and φ2 mentioned in the original text. We stress that these mappings
only depend on the matrix M that contains H .

783
 Free ebooks ==> www.Ebook777.com
CORRECTIONS AND ADDITIONS TO VOLUME 1

C.7. Additional Mottoes

Motto for Section 3.2

Indistinguishable things are identical

(or should be considered as identical).
The Principle of Identity of Indiscernibles
G. W. Leibniz (1646–1714)

(Leibniz admits that counter-examples to this principle are conceivable but will not
occur in real life because God is much too benevolent.)

Motto for Chapter 4

A: Please.
B: Please.
A: I insist.
B: So do I.
A: OK then, thank you.
B: You are most welcome.
A protocol for two Italians to pass through a door.
Source: Silvio Micali, 1985.

(The protocol is zero-knowledge because it can be simulated without knowing any of

the secrets of these Italians; in fact, the execution is independent of their secrets as well
as of anything else.)

784

www.Ebook777.com
 Bibliography

[1] W. Alexi, B. Chor, O. Goldreich, and C. P. Schnorr. RSA/Rabin Functions: Certain

Parts Are as Hard as the Whole. SIAM Journal on Computing, Vol. 17, April 1988,
pages 194–209.
[2] J. H. An and M. Bellare. Constructing VIL-MACs from FIL-MACs: Message Authenti-
cation under Weakened Assumptions. In Crypto99, Springer Lecture Notes in Computer
Science (Vol. 1666), 1999, pages 252–269.
[3] H. Attiya and J. Welch. Distributed Computing: Fundamentals, Simulations and Advanced
Topics. London: McGraw-Hill, 1998.
[4] E. Bach and J. Shallit. Algorithmic Number Theory (Volume I: Efficient Algorithms).
Cambridge, MA: MIT Press, 1996.
[5] B. Barak. How to Go Beyond the Black-Box Simulation Barrier. In 42nd IEEE Symposium
on Foundations of Computer Science, 2001, pages 106–115.
[6] B. Barak. Constant-Round Coin-Tossing with a Man in the Middle or Realizing the Shared
Random-String Model. In 43th IEEE Symposium on Foundations of Computer Science,
2002, pages 345–355.
[7] B. Barak and O. Goldreich. Universal Arguments and Their Applications. In the 17th
IEEE Conference on Computational Complexity, 2002, pages 194–203.
[8] B. Barak, O. Goldreich, R. Impagliazzo, S. Rudich, A. Sahai, S. Vadhan, and K. Yang. On
the (Im)possibility of Software Obfuscation. In Crypto01, Springer-Verlag Lecture Notes
in Computer Science (Vol. 2139), 2001, pages 1–18.
[9] B. Barak and Y. Lindell. Strict Polynomial-Time in Simulation and Extraction. In 34th
ACM Symposium on the Theory of Computing, 2002, pages 484–493.
[10] D. Beaver. Foundations of Secure Interactive Computing. In Crypto91, Springer-Verlag
Lecture Notes in Computer Science (Vol. 576), 1992, pages 377–391.
[11] D. Beaver. Secure Multi-Party Protocols and Zero-Knowledge Proof Systems Tolerating
a Faulty Minority. Journal of Cryptology, Vol. 4, 1991, pages 75–122.
[12] M. Bellare. A Note on Negligible Functions. Journal of Cryptology, Vol. 15, 2002,
pages 271–284.
[13] M. Bellare, R. Canetti, and H. Krawczyk. Pseudorandom Functions Revisited: The Cas-
cade Construction and Its Concrete Security. In 37th IEEE Symposium on Foundations of
Computer Science, 1996, pages 514–523.

785
 Free ebooks ==> www.Ebook777.com
BIBLIOGRAPHY

[14] M. Bellare, R. Canetti, and H. Krawczyk. Keying Hash Functions for Message Authen-
tication. In Crypto96, Springer Lecture Notes in Computer Science (Vol. 1109), 1996,
pages 1–15.
[15] M. Bellare, R. Canetti, and H. Krawczyk. Modular Approach to the Design and Analysis
of Authentication and Key Exchange Protocols. In 30th ACM Symposium on the Theory
of Computing, 1998, pages 419–428.
[16] M. Bellare, A. Desai, D. Pointcheval, and P. Rogaway. Relations among Notions of Security
for Public-Key Encryption Schemes. In Crypto98, Springer Lecture Notes in Computer
Science (Vol. 1462), 1998, pages 26–45.
[17] M. Bellare and O. Goldreich. On Defining Proofs of Knowledge. In Crypto92, Springer-
Verlag Lecture Notes in Computer Science (Vol. 740), 1992, pages 390–420.
[18] M. Bellare, O. Goldreich, and S. Goldwasser. Incremental Cryptography: The Case of
Hashing and Signing. In Crypto94, Springer-Verlag Lecture Notes in Computer Science
(Vol. 839), 1994, pages 216–233.
[19] M. Bellare, O. Goldreich, and S. Goldwasser. Incremental Cryptography and Applica-
tion to Virus Protection. In 27th ACM Symposium on the Theory of Computing, 1995,
pages 45–56.
[20] M. Bellare, O. Goldreich, and H. Krawczyk. Stateless Evaluation of Pseudorandom Func-
tions: Security Beyond the Birthday Barrier. In Crypto99, Springer Lecture Notes in
Computer Science (Vol. 1666), 1999, pages 270–287.
[21] M. Bellare and S. Goldwasser. New Paradigms for Digital Signatures and Message Authen-
tication Based on Non-Interative Zero-Knowledge Proofs. In Crypto89, Springer-Verlag
Lecture Notes in Computer Science (Vol. 435), 1990, pages 194–211.
[22] M. Bellare, R. Guerin, and P. Rogaway. XOR MACs: New Methods for Message Authenti-
cation Using Finite Pseudorandom Functions. In Crypto95, Springer-Verlag Lecture Notes
in Computer Science (Vol. 963), 1995, pages 15–28.
[23] M. Bellare, S. Halevi, A. Sahai, and S. Vadhan. Trapdoor Functions and Public-Key
Cryptosystems. In Crypto98, Springer Lecture Notes in Computer Science (Vol. 1462),
1998, pages 283–298.
[24] M. Bellare, R. Impagliazzo, and M. Naor. Does Parallel Repetition Lower the Error in
Computationally Sound Protocols? In 38th IEEE Symposium on Foundations of Computer
Science, 1997, pages 374–383.
[25] M. Bellare, J. Kilian, and P. Rogaway. The Security of Cipher Block Chaining. In Crypto94,
Springer-Verlag Lecture Notes in Computer Science (Vol. 839), 1994, pages 341–358.
[26] M. Bellare and S. Micali. How to Sign Given Any Trapdoor Function. Journal of the ACM,
Vol. 39, 1992, pages 214–233.
[27] D. Beaver, S. Micali, and P. Rogaway. The Round Complexity of Secure Protocols. In
22nd ACM Symposium on the Theory of Computing, 1990, pages 503–513.
[28] M. Bellare and P. Rogaway. Random Oracles Are Practical: A Paradigm for Designing
Efficient Protocols. In 1st Conf. on Computer and Communications Security, ACM, 1993,
pages 62–73.
[29] M. Bellare and P. Rogaway. Entity Authentication and Key Distribution. In Crypto93,
Springer-Verlag Lecture Notes in Computer Science (Vol. 773), 1994, pages 232–249.
[30] M. Bellare and P. Rogaway. Provably Secure Session Key Distribution: The Three Party
Case. In 27th ACM Symposium on the Theory of Computing, 1995, pages 57–66.
[31] M. Bellare and P. Rogaway. The Exact Security of Digital Signatures: How to Sign with
RSA and Rabin. In EuroCrypt96, Springer Lecture Notes in Computer Science (Vol. 1070),
1996, pages 399–416.
[32] M. Bellare and M. Yung. Certifying Permutations: Noninteractive Zero-Knowledge Based
on Any Trapdoor Permutation. Journal of Cryptology, Vol. 9, 1996, pages 149–166.

786

www.Ebook777.com
 BIBLIOGRAPHY

[33] M. Ben-Or, R. Canetti, and O. Goldreich. Asynchronous Secure Computation. In

25th ACM Symposium on the Theory of Computing, 1993, pages 52–61. See details
in [49].
[34] M. Ben-Or, S. Goldwasser, and A. Wigderson. Completeness Theorems for Non-
Cryptographic Fault-Tolerant Distributed Computation. In 20th ACM Symposium on the
Theory of Computing, 1988, pages 1–10.
[35] J. Black, S. Halevi, H. Krawczyk, T. Krovetz, and P. Rogaway. UMAC: Fast and Se-
cure Message Authentication. In Crypto99, Springer Lecture Notes in Computer Science
(Vol. 1666), 1999, pages 216–233.
[36] M. Blum. How to Exchange Secret Keys. ACM Trans. Comput. Sys., Vol. 1, 1983, pages
175–193.
[37] M. Blum. Coin Flipping by Phone. In the 24th IEEE Computer Conference (CompCon),
February 1982, pages 133–137. See also SIGACT News, Vol. 15, No. 1, 1983.
[38] L. Blum, M. Blum, and M. Shub. A Simple Secure Unpredictable Pseudo-Random Number
Generator. SIAM Journal on Computing, Vol. 15, 1986, pages 364–383.
[39] M. Blum, A. De Santis, S. Micali, and G. Persiano. Non-Interactive Zero-Knowledge
Proof Systems. SIAM Journal on Computing, Vol. 20, No. 6, 1991, pages 1084–1118.
(Considered the journal version of [40].)
[40] M. Blum, P. Feldman, and S. Micali. Non-Interactive Zero-Knowledge and Its Applica-
tions. In 20th ACM Symposium on the Theory of Computing, 1988, pages 103–112. See
[39].
[41] M. Blum and S. Goldwasser. An Efficient Probabilistic Public-Key Encryption Scheme
Which Hides All Partial Information. In Crypto84, Springer-Verlag Lecture Notes in
Computer Science (Vol. 196), 1985, pages 289–302.
[42] M. Blum and S. Micali. How to Generate Cryptographically Strong Sequences of Pseudo-
Random Bits. SIAM Journal on Computing, Vol. 13, 1984, pages 850–864. Preliminary
version in 23rd IEEE Symposium on Foundations of Computer Science, 1982.
[43] J. B. Boyar. Inferring Sequences Produced by Pseudo-Random Number Generators. Jour-
nal of the ACM, Vol. 36, 1989, pages 129–141.
[44] G. Brassard. A Note on the Complexity of Cryptography. IEEE Trans. on Inform. Th.,
Vol. 25, 1979, pages 232–233.
[45] G. Brassard. Quantum Information Processing: The Good, the Bad and the Ugly. In
Crypto97, Springer Lecture Notes in Computer Science (Vol. 1294), 1997 pages 337–
341.
[46] G. Brassard, D. Chaum, and C. Crépeau. Minimum Disclosure Proofs of Knowledge. Jour-
nal of Computer and System Science, Vol. 37, No. 2, 1988, pages 156–189. Preliminary
version by Brassard and Crépeau in 27th IEEE Symposium on Foundations of Computer
Science, 1986.
[47] G. Brassard, C. Crépeau, and M. Yung. Constant-Round Perfect Zero-Knowledge Com-
putationally Convincing Protocols. Theoretical Computer Science, Vol. 84, 1991, pages
23–52.
[48] C. Cachin and U. Maurer. Unconditional Security Against Memory-Bounded Adversaries.
In Crypto97, Springer Lecture Notes in Computer Science (Vol. 1294), 1997, pages 292–
306.
[49] R. Canetti. Studies in Secure Multi-Party Computation and Applications. Ph.D. the-
sis, Department of Computer Science and Applied Mathematics, Weizmann Institute of
Science, Rehovot, Israel, June 1995. Available from http://theory.lcs.mit.edu/
∼tcryptol/BOOKS/ran-phd.html.
[50] R. Canetti. Security and Composition of Multi-party Cryptographic Protocols. Journal of
Cryptology, Vol. 13, No. 1, 2000, pages 143–202.

787
 Free ebooks ==> www.Ebook777.com
BIBLIOGRAPHY

[51] R. Canetti. Universally Composable Security: A New Paradigm for Cryptographic Proto-
cols. In 42nd IEEE Symposium on Foundations of Computer Science, 2001, pages 136–
145. Full version (with different title) is available from Cryptology ePrint Archive, Report
2000/067.
[52] R. Canetti, I. Damgard, S. Dziembowski, Y. Ishai, and T. Malkin. On Adaptive Versus
Non-Adaptive Security of Multiparty Protocols. Journal of Cryptology, forthcoming.
[53] R. Canetti, U. Feige, O. Goldreich, and M. Naor. Adaptively Secure Multiparty Compu-
tation. In 28th ACM Symposium on the Theory of Computing, 1996, pages 639–648.
[54] R. Canetti, O. Goldreich, and S. Halevi. The Random Oracle Methodology, Revisited. In
30th ACM Symposium on the Theory of Computing, 1998, pages 209–218.
[55] R. Canetti, O. Goldreich, S. Goldwasser, and S. Micali. Resettable Zero-Knowledge. In
32nd ACM Symposium on the Theory of Computing, 2000, pages 235–244.
[56] R. Canetti, S. Halevi, and A. Herzberg. How to Maintain Authenticated Communication
in the Presence of Break-Ins. Journal of Cryptology, Vol. 13, No. 1, 2000, pages 61–106.
[57] R. Canetti and A. Herzberg. Maintaining Security in the Presence of Transient Faults. In
Crypto94, Springer-Verlag Lecture Notes in Computer Science (Vol. 839), 1994, pages
425–439.
[58] R. Canetti, J. Kilian, E. Petrank, and A. Rosen. Black-Box Concurrent Zero-Knowledge
Requires
(log
˜ n) Rounds. In 33rd ACM Symposium on the Theory of Computing, 2001,
pages 570–579.
[59] R. Canetti, Y. Lindell, R. Ostrovsky, and A. Sahai. Universally Composable Two-Party and
Multi-Party Secure Computation. In 34th ACM Symposium on the Theory of Computing,
2002, pages 494–503.
[60] L. Carter and M. Wegman. Universal Hash Functions. Journal of Computer and System
Science, Vol. 18, 1979, pages 143–154.
[61] D. Chaum. Blind Signatures for Untraceable Payments. In Crypto82. New York: Plenum
Press, 1983, pages 199–203.
[62] D. Chaum, C. Crépeau, and I. Damgård. Multi-party Unconditionally Secure Protocols.
In 20th ACM Symposium on the Theory of Computing, 1988, pages 11–19.
[63] B. Chor, S. Goldwasser, S. Micali, and B. Awerbuch. Verifiable Secret Sharing and Achiev-
ing Simultaneity in the Presence of Faults. In 26th IEEE Symposium on Foundations of
Computer Science, 1985, pages 383–395.
[64] B. Chor and E. Kushilevitz. A Zero-One Law for Boolean Privacy. SIAM J. on Disc. Math.,
Vol. 4, 1991, pages 36–47.
[65] R. Cleve. Limits on the Security of Coin Flips When Half the Processors Are Faulty. In
18th ACM Symposium on the Theory of Computing, 1986, pages 364–369.
[66] J. D. Cohen and M. J. Fischer. A Robust and Verifiable Cryptographically Secure Election
Scheme. In 26th IEEE Symposium on Foundations of Computer Science, 1985, pages
372–382.
[67] R. Cramer and I. Damgård. New Generation of Secure and Practical RSA-Based Signa-
tures. In Crypto96, Springer Lecture Notes in Computer Science (Vol. 1109), 1996, pages
173–185.
[68] R. Cramer and V. Shoup. A Practical Public-Key Cryptosystem Provably Secure Against
Adaptive Chosen Ciphertext Attacks. In Crypto98, Springer-Verlag Lecture Notes in Com-
puter Science (Vol. 1462), 1998, pages 13–25.
[69] C. Crépeau. Efficient Cryptographic Protocols Based on Noisy Channels. In EuroCrypt97,
Springer, Lecture Notes in Computer Science (Vol. 1233), 1997, pages 306–317.
[70] I. Damgård. Collision Free Hash Functions and Public Key Signature Schemes. In
EuroCrypt87, Springer-Verlag Lecture Notes in Computer Science (Vol. 304), 1988,
pages 203–216.

788

www.Ebook777.com
 BIBLIOGRAPHY

[71] I. Damgård. A Design Principle for Hash Functions. In Crypto89, Springer-Verlag Lecture
Notes in Computer Science (Vol. 435), 1990, pages 416–427.
[72] I. Damgård. Concurrent Zero-Knowledge in Easy in Practice: Theory of Cryptography
Library, 99-14, June 1999. http://philby.ucsd.edu/cryptolib. See also “Efficient
Concurrent Zero-Knowledge in the Auxiliary String Model” (in Eurocrypt’00, 2000).
[73] A. De Santis, G. Di Crescenzo, R. Ostrovsky, G. Persiano, and A. Sahai. Robust Non-
interactive Zero-Knowledge. In Crypto01, Springer Lecture Notes in Computer Science
(Vol. 2139), 2001, pages 566–598.
[74] Y. Desmedt and Y. Frankel. Threshold Cryptosystems. In Crypto89, Springer-Verlag Lec-
ture Notes in Computer Science (Vol. 435), 1990, pages 307–315.
[75] W. Diffie and M. E. Hellman. New Directions in Cryptography. IEEE Trans. on Info.
Theory, IT-22, Nov. 1976, pages 644–654.
[76] H. Dobbertin. The Status of MD5 after a Recent Attack. In CryptoBytes, RSA Lab., Vol. 2,
No. 2, 1996, pages 1–6.
[77] D. Dolev, C. Dwork, and M. Naor. Non-Malleable Cryptography. In 23rd ACM Symposium
on the Theory of Computing, 1991, pages 542–552. Full version available from authors.
[78] D. Dolev, C. Dwork, O. Waarts, and M. Yung. Perfectly Secure Message Transmission.
Journal of the ACM, Vol. 40 (1), 1993, pages 17–47.
[79] D. Dolev and A. C. Yao. On the Security of Public-Key Protocols. IEEE Trans. on Inform.
Theory, Vol. 30, No. 2, 1983, pages 198–208.
[80] D. Dolev and H. R. Strong. Authenticated Algorithms for Byzantine Agreement. SIAM
Journal on Computing, Vol. 12, 1983, pages 656–666.
[81] C. Dwork and M. Naor. An Efficient Existentially Unforgeable Signature Scheme and Its
Application. Journal of Cryptology, Vol. 11 (3), 1998, pages 187–208
[82] C. Dwork, M. Naor, and A. Sahai. Concurrent Zero-Knowledge. In 30th ACM Symposium
on the Theory of Computing, 1998, pages 409–418.
[83] S. Even and O. Goldreich. On the Security of Multi-party Ping-Pong Protocols. In 24th
IEEE Symposium on Foundations of Computer Science, 1983, pages 34–39.
[84] S. Even, O. Goldreich, and A. Lempel. A Randomized Protocol for Signing Contracts.
CACM, Vol. 28, No. 6, 1985, pages 637–647.
[85] S. Even, O. Goldreich, and S. Micali. On-line/Off-line Digital Signatures. Journal of
Cryptology, Vol. 9, 1996, pages 35–67.
[86] S. Even, A.L. Selman, and Y. Yacobi. The Complexity of Promise Problems with
Applications to Public-Key Cryptography. Information and Control, Vol. 61, 1984,
pages 159–173.
[87] S. Even and Y. Yacobi. Cryptography and NP-Completeness. In Proceedings of 7th ICALP,
Springer-Verlag Lecture Notes in Computer Science (Vol. 85), 1980, pages 195–207. See
[86].
[88] U. Feige, A. Fiat, and A. Shamir. Zero-Knowledge Proofs of Identity. Journal of Cryptol-
ogy, Vol. 1, 1988, pages 77–94.
[89] U. Feige, D. Lapidot, and A. Shamir. Multiple Non-Interactive Zero-Knowledge Proofs
under General Assumptions. SIAM Journal on Computing, Vol. 29 (1), 1999, pages 1–28.
[90] U. Feige and A. Shamir. Zero-Knowledge Proofs of Knowledge in Two Rounds.
In Crypto89, Springer-Verlag Lecture Notes in Computer Science (Vol. 435), 1990,
pages 526–544.
[91] U. Feige and A. Shamir. Witness Indistinguishability and Witness Hiding Protocols. In
22nd ACM Symposium on the Theory of Computing, 1990, pages 416–426.
[92] A. Fiat and A. Shamir. How to Prove Yourself: Practical Solution to Identification and
Signature Problems. In Crypto86, Springer-Verlag Lecture Notes in Computer Science
(Vol. 263), 1987, pages 186–189.

789
 Free ebooks ==> www.Ebook777.com
BIBLIOGRAPHY

[93] M. Fischer, S. Micali, C. Rackoff, and D. K. Wittenberg. An Oblivious Transfer Protocol

Equivalent to Factoring. Unpublished manuscript, 1986. Preliminary versions were pre-
sented in EuroCrypt84 and in the NSF Workshop on Mathematical Theory of Security,
Endicott House, 1985.
[94] A. M. Frieze, J. Håstad, R. Kannan, J. C. Lagarias, and A. Shamir. Reconstructing Trun-
cated Integer Variables Satisfying Linear Congruences. SIAM Journal on Computing,
Vol. 17, 1988, pages 262–280.
[95] M. R. Garey and D. S. Johnson. Computers and Intractability: A Guide to the Theory of
NP-Completeness. New York: W. H. Freeman and Company, 1979.
[96] P. S. Gemmell. An Introduction to Threshold Cryptography. In CryptoBytes, RSA Lab.,
Vol. 2, No. 3, 1997, pages 7–12.
[97] R. Gennaro, M. Rabin, and T. Rabin. Simplified VSS and Fast-Track Multiparty Com-
putations with Applications to Threshold Cryptography. In 17th ACM Symposium on
Principles of Distributed Computing, 1998, pages 101–112.
[98] R. Gennaro and L. Trevisan. Lower Bounds on the Efficiency of Generic Cryptographic
Constructions. In 41st Symposium on Foundations of Computer Science, 2000, pages
305–313.
[99] E. N. Gilbert, F. J. MacWilliams, and N. J. A. Sloane. Codes Which Detect Deception.
Bell Syst. Tech. J., Vol. 53, 1974, pages 405–424.
[100] O. Goldreich. Two Remarks Concerning the GMR Signature Scheme. In Crypto86,
Springer-Verlag Lecture Notes in Computer Science (Vol. 263), 1987, pages 104–110.
[101] O. Goldreich. Foundation of Cryptography – Class Notes. Preprint, Spring 1989. See [102].
Superseded by the current work.
[102] O. Goldreich. Lecture Notes on Encryption, Signatures and Cryptographic Protocol.
Extracts from [101]. Available from http://www.wisdom.weizmann.ac.il/∼oded
/foc.html. Superseded by the current work.
[103] O. Goldreich. A Note on Computational Indistinguishability. Information Processing Let-
ters, Vol. 34, May 1990, pages 277–281.
[104] O. Goldreich. A Uniform Complexity Treatment of Encryption and Zero-Knowledge.
Journal of Cryptology, Vol. 6, No. 1, 1993, pages 21–53.
[105] O. Goldreich. Foundation of Cryptography – Fragments of a Book. February 1995. Avail-
able from http://www.wisdom.weizmann.ac.il/∼oded/foc.html. Superseded by
the current work.
[106] O. Goldreich. Modern Cryptography, Probabilistic Proofs and Pseudorandomness. Algo-
rithms and Combinatorics series, Vol. 17. Heidelberg: Springer, 1999.
[107] O. Goldreich. Secure Multi-Party Computation. Unpublished manuscript, 1998. Available
from http://www.wisdom.weizmann.ac.il/∼oded/foc.html. Superseded by the
current work.
[108] O. Goldreich. Foundation of Cryptography – Basic Tools. New York: Cambridge University
Press, 2001.
[109] O. Goldreich. Concurrent Zero-Knowledge With Timing, Revisited. In 34th ACM Sym-
posium on the Theory of Computing, 2002, pages 332–340.
[110] O. Goldreich, S. Goldwasser, and S. Micali. How to Construct Random Functions. Journal
of the ACM, Vol. 33, No. 4, 1986, pages 792–807.
[111] O. Goldreich, S. Goldwasser, and S. Micali. On the Cryptographic Applications of Random
Functions. In Crypto84, Springer-Verlag Lecture Notes in Computer Science (Vol. 263),
1985, pages 276–288.
[112] O. Goldreich and A. Kahan. How to Construct Constant-Round Zero-Knowledge Proof
Systems for NP. Journal of Cryptology, Vol. 9, No. 2, 1996, pages 167–189. Preliminary
versions date to 1988.

790

www.Ebook777.com
 BIBLIOGRAPHY

[113] O. Goldreich and H. Krawczyk. On the Composition of Zero-Knowledge Proof Systems.

SIAM Journal on Computing, Vol. 25, No. 1, February 1996, pages 169–192.
[114] O. Goldreich and L. A. Levin. Hard-Core Predicates for Any One-Way Function. In 21st
ACM Symposium on the Theory of Computing, 1989, pages 25–32.
[115] O. Goldreich and Y. Lindell. Session-Key Generation Using Human Passwords. In
Crypto01, Springer-Verlag Lecture Notes in Computer Science (Vol. 2139), 2001, pages
408–432.
[116] O. Goldreich, Y. Lustig, and M. Naor. On Chosen Ciphertext Security of Multiple En-
cryptions. Cryptology ePrint Archive, Report 2002/089, 2002.
[117] O. Goldreich, S. Micali, and A. Wigderson. Proofs That Yield Nothing but Their Validity
or All Languages in NP Have Zero-Knowledge Proof Systems. Journal of the ACM,
Vol. 38, No. 1, 1991, pages 691–729. Preliminary version in 27th IEEE Symposium on
Foundations of Computer Science, 1986.
[118] O. Goldreich, S. Micali, and A. Wigderson. How to Play Any Mental Game – A Com-
pleteness Theorem for Protocols with Honest Majority. In 19th ACM Symposium on the
Theory of Computing, 1987, pages 218–229.
[119] O. Goldreich and Y. Oren. Definitions and Properties of Zero-Knowledge Proof Systems.
Journal of Cryptology, Vol. 7, No. 1, 1994, pages 1–32.
[120] O. Goldreich and R. Vainish. How to Solve Any Protocol Problem – An Efficiency Im-
provement. In Crypto87, Springer Verlag Lecture Notes in Computer Science (Vol. 293),
1988, pages 73–86.
[121] S. Goldwasser and L. A. Levin. Fair Computation of General Functions in Presence
of Immoral Majority. In Crypto90, Springer-Verlag Lecture Notes in Computer Science
(Vol. 537), 1991, pages 77–93.
[122] S. Goldwasser and Y. Lindell. Secure Computation Without Agreement. In 16th Interna-
tional Symposium on Distributed Computing (DISC), Springer-Verlag Lecture Notes in
Computer Science (Vol. 2508), 2002, pages 17–32.
[123] S. Goldwasser and S. Micali. Probabilistic Encryption. Journal of Computer and System
Science, Vol. 28, No. 2, 1984, pages 270–299. Preliminary version in 14th ACM Symposium
on the Theory of Computing, 1982.
[124] S. Goldwasser, S. Micali, and C. Rackoff. The Knowledge Complexity of Interactive Proof
Systems. SIAM Journal on Computing, Vol. 18, 1989, pages 186–208. Preliminary version
in 17th ACM Symposium on the Theory of Computing, 1985.
[125] S. Goldwasser, S. Micali, and R. L. Rivest. A Digital Signature Scheme Secure Against
Adaptive Chosen-Message Attacks. SIAM Journal on Computing, Vol. 17, No. 2, April
1988, pages 281–308.
[126] S. Goldwasser, S. Micali, and P. Tong. Why and How to Establish a Private Code in a
Public Network. In 23rd IEEE Symposium on Foundations of Computer Science, 1982,
pages 134–144.
[127] S. Goldwasser, S. Micali, and A. C. Yao. Strong Signature Schemes. In 15th ACM Sym-
posium on the Theory of Computing, 1983, pages 431–439.
[128] S. Goldwasser and R. Ostrovsky. Invariant Signatures and Non-Interactive Zero-
Knowledge Proofs Are Equivalent. In Crypto92, Springer-Verlag Lecture Notes in Com-
puter Science (Vol. 740), 1992, pages 228–245.
[129] S. Haber and S. Micali. Private communication, 1986.
[130] J. Håstad, R. Impagliazzo, L. A. Levin, and M. Luby. A Pseudorandom Generator from
Any One-way Function. SIAM Journal on Computing, Vol. 28, No. 4, 1999, pages 1364–
1396. Preliminary versions by Impagliazzo et al. in 21st ACM Symposium on the Theory
of Computing (1989) and Håstad in 22nd ACM Symposium on the Theory of Computing
(1990).

791
 Free ebooks ==> www.Ebook777.com
BIBLIOGRAPHY

[131] M. Hirt and U. Maurer. Complete Characterization of Adversaries Tolerable in Secure

Multi-party Computation. Journal of Cryptology, Vol. 13, No. 1, 2000, pages 31–60.
[132] R. Impagliazzo and M. Luby. One-Way Functions Are Essential for Complexity Based
Cryptography. In 30th IEEE Symposium on Foundations of Computer Science, 1989, pages
230–235.
[133] R. Impagliazzo and S. Rudich. Limits on the Provable Consequences of One-Way Permu-
tations. In 21st ACM Symposium on the Theory of Computing, 1989, pages 44–61.
[134] A. Juels, M. Luby, and R. Ostrovsky. Security of Blind Digital Signatures. In Crypto97,
Springer-Verlag Lecture Notes in Computer Science (Vol. 1294), 1997, pages 150–164.
[135] J. Kahn, M. Saks, and C. Smyth. A Dual Version of Reimer’s Inequality and a Proof
of Rudich’s Conjecture. In 15th IEEE Conference on Computational Complexity, 2000,
pages 98–103.
[136] J. Katz and M. Yung. Complete Characterization of Security Notions for Probabilistic
Private-Key Encryption. In 32nd ACM Symposium on the Theory of Computing, 2000,
pages 245–254.
[137] J. Kilian. Basing Cryptography on Oblivious Transfer. In 20th ACM Symposium on the
Theory of Computing, 1988, pages 20–31.
[138] J. Kilian and E. Petrank. Concurrent and Resettable Zero-Knowledge in Poly-logarithmic
Rounds. In 33rd ACM Symposium on the Theory of Computing, 2001, pages 560–569.
[139] H. Krawczyk. LFSR-Based Hashing and Authentication. In Crypto94, Springer-Verlag
Lecture Notes in Computer Science (Vol. 839), 1994, pages 129–139.
[140] H. Krawczyk. New Hash Functions For Message Authentication. In EuroCrypt95,
Springer-Verlag Lecture Notes in Computer Science (Vol. 921), 1995, pages 301–310.
[141] A. Lempel. Cryptography in Transition. Computing Surveys, Vol. 11, No. 4, Dec. 1979,
pages 285–303.
[142] Y. Lindell. A Simpler Construction of CCA2-Secure Public-Key Encryption under General
Assumptions. In EuroCrypt03, Springer Lecture Notes in Computer Science (Vol. 2656),
2003, pages 241–254.
[143] Y. Lindell. Parallel Coin-Tossing and Constant-Round Secure Two-Party Computation. In
Crypto01, Springer Lecture Notes in Computer Science (Vol. 2139), 2001, pages 171–189.
[144] Y. Lindell, A. Lysyanskaya, and T. Rabin. On the Composition of Authenticated Byzantine
Agreement. In 34th ACM Symposium on the Theory of Computing, 2002, pages 514–523.
[145] M. Luby. Pseudorandomness and Cryptographic Applications. Princeton, NJ: Princeton
University Press, 1996.
[146] M. Luby and C. Rackoff. How to Construct Pseudorandom Permutations from Pseudo-
random Functions. SIAM Journal on Computing, Vol. 17, 1988, pages 373–386.
[147] N. Lynch. Distributed Algorithms. San Mateo, CA: Morgan Kaufmann Publishers, 1996.
[148] U. Maurer. Secret Key Agreement by Public Discussion from Common Information. IEEE
Trans. on Inform. Th., Vol. 39, No. 3, May 1993, pages 733–742.
[149] A. J. Menezes, P. C. van Oorschot, and S. A. Vanstone. Handbook of Applied Cryptography.
Boca Raton, FL: CRC Press, 1996.
[150] R. C. Merkle. Secure Communication over Insecure Channels. CACM, Vol. 21, No. 4,
1978, pages 294–299.
[151] R. C. Merkle. Protocols for Public Key Cryptosystems. In Proceedings of the 1980 Sym-
posium on Security and Privacy, 1980, pages 122–134.
[152] R. C. Merkle. A Digital Signature Based on a Conventional Encryption Function. In
Crypto87, Springer-Verlag Lecture Notes in Computer Science (Vol. 293), 1987, pages
369–378.
[153] R. C. Merkle. A Certified Digital Signature Scheme. In Crypto89, Springer-Verlag Lecture
Notes in Computer Science (Vol. 435), 1990, pages 218–238.

792

www.Ebook777.com
 BIBLIOGRAPHY

[154] R. C. Merkle and M. E. Hellman. Hiding Information and Signatures in Trapdoor Knap-
sacks. IEEE Trans. Inform. Theory, Vol. 24, 1978, pages 525–530.
[155] S. Micali, M. O. Rabin, and S. Vadhan. Verifiable Random Functions. In 40th IEEE
Symposium on Foundations of Computer Science, 1999, pages 120–130.
[156] S. Micali, C. Rackoff, and B. Sloan. The Notion of Security for Probabilistic Cryptosys-
tems. SIAM Journal on Computing, Vol. 17, 1988, pages 412–426.
[157] S. Micali and P. Rogaway. Secure Computation. In Crypto91, Springer-Verlag Lecture
Notes in Computer Science (Vol. 576), 1992, pages 392–404.
[158] D. Micciancio. Oblivious Data Structures: Applications to Cryptography. In 29th ACM
Symposium on the Theory of Computing, 1997, pages 456–464.
[159] National Bureau of Standards. Data Encryption Standard (DES). Federal Information
Processing Standards, Publ. 46, 1977.
[160] National Institute for Standards and Technology. Digital Signature Standard (DSS).
Federal Register, Vol. 56, No. 169, Aug. 1991.
[161] M. Naor. Bit Commitment Using Pseudorandom Generators. Journal of Cryptology,
Vol. 4, 1991, pages 151–158.
[162] M. Naor and O. Reingold. From Unpredictability to Indistinguishability: A Simple Con-
struction of Pseudorandom Functions from MACs. In Crypto98, Springer-Verlag Lecture
Notes in Computer Science (Vol. 1464), 1998, pages 267–282.
[163] M. Naor and M. Yung. Universal One-Way Hash Functions and their Crypto-
graphic Application. 21st ACM Symposium on the Theory of Computing, 1989,
pages 33–43.
[164] M. Naor and M. Yung. Public-Key Cryptosystems Provably Secure Against Chosen Ci-
phertext Attacks. In 22nd ACM Symposium on the Theory of Computing, 1990, pages
427–437.
[165] R. Ostrovsky, R. Venkatesan, and M. Yung. Secure Commitment Against Powerful Ad-
versary: A Security Primitive Based on Average Intractability. In Proceedings of the 9th
Symposium on Theoretical Aspects of Computer Science (STACS92), 1992, pages 439–
448.
[166] R. Ostrovsky and M. Yung. How to Withstand Mobile Virus Attacks. In 10th ACM Sym-
posium on Principles of Distributed Computing, 1991, pages 51–59.
[167] T. P. Pedersen and B. Pfitzmann. Fail-Stop Signatures. SIAM Journal on Computing,
Vol. 26, No. 2, 1997, pages 291–330. Based on several earlier works (see first footnote in
the paper).
[168] B. Pfitzmann. Digital Signature Schemes (General Framework and Fail-Stop Signatures).
Springer-Verlag Lecture Notes in Computer Science (Vol. 1100), 1996.
[169] M. Prabhakaran, A. Rosen, and A. Sahai. Concurrent Zero-Knowledge Proofs in Logarith-
mic Number of Rounds. In 43rd IEEE Symposium on Foundations of Computer Science,
2002, pages 366–375.
[170] M. O. Rabin. Digitalized Signatures. In Foundations of Secure Computation, R. A.
DeMillo et al., eds. New York: Academic Press, 1977, pages 155–168.
[171] M. O. Rabin. Digitalized Signatures and Public Key Functions as Intractable as Factoring.
TR-212, LCS, MIT, 1979.
[172] M. O. Rabin. How to Exchange Secrets by Oblivious Transfer. Tech. Memo TR-81, Aiken
Computation Laboratory, Harvard University, 1981.
[173] T. Rabin and M. Ben-Or. Verifiable Secret Sharing and Multi-party Protocols with Honest
Majority. In 21st ACM Symposium on the Theory of Computing, 1989, pages 73–85.
[174] C. Rackoff and D. R. Simon. Non-Interactive Zero-Knowledge Proof of Knowledge and
Chosen Ciphertext Attack. In Crypto91, Springer Verlag Lecture Notes in Computer
Science (Vol. 576), 1991, pages 433–444.

793
 Free ebooks ==> www.Ebook777.com
BIBLIOGRAPHY

[175] R. Richardson and J. Kilian. On the Concurrent Composition of Zero-Knowledge Proofs.

In EuroCrypt99, Springer-Verlag Lecture Notes in Computer Science (Vol. 1592), 1999,
pages 415–413.
[176] R. Rivest, A. Shamir, and L. Adleman. A Method for Obtaining Digital Signatures and
Public Key Cryptosystems. CACM, Vol. 21, Feb. 1978, pages 120–126.
[177] P. Rogaway. The Round Complexity of Secure Protocols. Ph.D. thesis, MIT June 1991.
Available from http://www.cs.ucdavis.edu/∼rogaway/papers.
[178] J. Rompel. One-Way Functions Are Necessary and Sufficient for Secure Signatures. In
22nd ACM Symposium on the Theory of Computing, 1990, pages 387–394.
[179] A. Sahai. Non-Malleable Non-Interactive Zero Knowledge and Achieving Chosen-
Ciphertext Security. In 40th IEEE Symposium on Foundations of Computer Science, 1999,
pages 543–553.
[180] A. Sahai. Improved Constructions Achieving Chosen-Ciphertext Security. Unpublished
manuscript, 2001. See [73].
[181] A. Shamir. On the Cryptocomplexity of Knapsack systems. In 11th ACM Symposium on
the Theory of Computing, 1979, pages 118–129.
[182] A. Shamir. How to Share a Secret. CACM, Vol. 22, Nov. 1979, pages 612–613.
[183] A. Shamir. A Polynomial-Time Algorithm for Breaking the Merkle-Hellman Cryptosys-
tem. In 23rd IEEE Symposium on Foundations of Computer Science, 1982, pages 145–152.
[184] A. Shamir, R. L. Rivest, and L. Adleman. Mental Poker. TM-125, LCS, MIT, 1979.
[185] C. E. Shannon. Communication Theory of Secrecy Systems. Bell System Technical Jour-
nal, Vol. 28, 1949, pages 656–715.
[186] D. Stinson. Universal Hashing and Authentication Codes. Designs, Codes and Cryptog-
raphy, Vol. 4, 1994, pages 369–380.
[187] S. Vadhan. Constructing Locally Computable Extractors and Cryptosystems in the
Bounded Storage Model. Journal of Cryptology, Vol. 17, No. 1, 2004, pages 43–77.
[188] M. Wegman and L. Carter. New Hash Functions and Their Use in Authentication and Set
Equality. Journal of Computer and System Science, Vol. 22, 1981, pages 265–279.
[189] A. D. Wyner. The Wire-Tap Channel. Bell System Technical Journal, Vol. 54, No. 8,
Oct. 1975, pages 1355–1387.
[190] A. C. Yao. Theory and Application of Trapdoor Functions. In 23rd IEEE Symposium on
Foundations of Computer Science, 1982, pages 80–91.
[191] A. C. Yao. How to Generate and Exchange Secrets. In 27th IEEE Symposium on Founda-
tions of Computer Science, 1986, pages 162–167.

794

www.Ebook777.com
 Index

Author Index Lapidot, D., 782

Adleman, L., 479, 587 Lempel, A., 757
Awerbuch, B., 757 Lipton, R., 480

Barak, B., 481, 775, 781 Merkle, R. C., 479

Beaver, D., 757 Micali, S., 379, 382, 479, 480, 587, 588, 589,
Ben-Or, M., 757 756, 757
Blum, M., 479, 480, 757
Naor, M., 480, 588, 778, 779
Canetti, R., 753, 757
Chaum, D., 757 Pfitzmann, B., 589
Chor, B., 757
Crépeau, C., 757 Rabin, M., 587, 757
Richardson, R., 778
Damgård, I., 757 Rivest, R. L., 478, 587, 588, 589
Diffie, W., 475, 478, 587 Rogaway, P., 757
Dolev, D., 480 Rompel, J., 588
Dwork, C., 480, 778, 779 Rudich, S., 481

Even, S., 757 Sahai, A., 480, 778, 779

Shamir, A., 478, 587, 782
Feige, U., 782 Shannon, C. E., 378, 476, 478
Feldman, P., 480
Wigderson, A., 756, 757
Goldreich, O., 479, 756, 757, 780
Goldwasser, S., 379, 382, 479, 480, 587, 588, Yao, A. C., 479, 587, 756, 757
589, 757 Yung, M., 480, 588

Hellman, M. E., 475, 478, 479, 587

Subject Index
Impagliazzo, R., 481 Averaging Argument. See Techniques

Kilian, J., 778 Byzantine Agreement, 711

Krawczyk, H., 780 Authenticated, 711–714, 717, 758

795
 Free ebooks ==> www.Ebook777.com
INDEX

Chinese Reminder Theorem, 421 The malicious model, 600, 603, 608,
Claw-free pairs. See One-way permutations 610–611 626, 634, 650–693,
Collision-free hashing. See Hashing 697–700, 708–741, 746–747
Collision-resistent hashing. See Hashing The semi-honest model, 600, 603, 608,
Commitment schemes 610–615, 619 626, 634–650, 696,
non-oblivious, 771 697, 700–708, 743–746
perfectly binding, 465–469 Two-party, 599, 600, 606–607, 608,
Computational indistinguishability, 382, 611–613, 615–693, 755
395–402, 446, 447–449, 457, 465, Universally Composable, 753
467–468, 479, 618, 770 Verifiable Secret Sharing. See Secret
by circuits, 382–393, 412, 417, 419, 431, Sharing
454, 618
Cryptographic protocols, 599–764 Discrete Logarithm Problem. See DLP
active adversary, 603 function
adaptive adversary, 603, 748–751 DLP function, 584
Authenticated Computation, 664–668,
671–674, 717–722 Encryption schemes, 373–496
Coin-tossing, 659–664, 674–677, 722–725 active attacks, 422–425, 431–474
Communication models, 602–603 asymmetric, 376
Computational limitations, 603 Basic Setting, 374–377
Concurrent executions, 752–755 Block-Ciphers, 408–418, 420
Definitional approach, 601–607 chosen ciphertext attacks, 423, 438–469,
Definitions, 615–634, 694–700, 742–743, 472–474
749, 752–754 chosen plaintext attacks, 423, 431–438
Environmentally-secure, 753–755 Definitions, 378–403
Fairness, 604, 747–748 indistinguishability of encryptions, 378,
functionality, 599 382–383, 403, 412, 415, 417, 419,
General circuit evaluation, 645–648, 424, 432, 459, 461, 479
705–707 multiple messages, 378, 389–393,
honest-but-curious adversary, 603 394–402, 429, 437–438, 443–449,
Image Transmission, 668–671, 672, 489
718–721 non-malleability, 422, 470–474
Input-Commitment, 677–680, 725–726 passive attacks, 422, 425–431
Multi-party, 599, 600, 604–606, 607–609, perfect privacy, 378, 476–477
610–611, 613–615, 693–747, 755 perfect security, 476–477
non-adaptive adversary, 603 Private-Key, 375–376, 377, 380, 381,
number of dishonest parties, 604 404–408, 410–413
Oblivious Transfer, 612, 614, 635, Probabilistic Encryption, 404, 410–422
640–645 Public-Key, 376, 377, 380, 381,
Oracle-aided, 636, 639, 644, 646, 652, 413–422
672, 674, 678, 681, 701, 704, 715, Randomized RSA, 416–417, 478
718, 721, 722, 726, 729, 737 Semantic Security, 378, 379–382, 478
Overview, 599–615 Stream-Ciphers, 404–408
passive adversary, 603 symmetric, 375
Privacy reductions, 635–640, 643, 644, The Blum-Goldwasser, 420–422, 478
647, 648, 701–703, 704 the mechanism, 376–377
Private Channels, 741–747 uniform-complexity treatment, 393–403
Pure oracle-aided, 721–722
Reactive, 609, 751–752 Factoring integers, 421, 584
Secret Broadcast, 716–717, 718, 722
Security reductions, 652–657, 673, 675, Hard-core predicates. See One-way
677, 678, 714–716, 719, 721, 723 permutations
Setup assumptions, 602, 608, 755 Hash and Sign. See Techniques

796

www.Ebook777.com
 INDEX

Hashing Proofs-of-Knowledge, 453, 669–671

collision-free, 512–523, 542–543, 558, for NP in zero-knowledge, 659, 669,
560–561, 562, 575 718–720
based on claw-free permutations, Protocols. See Cryptographic protocols
516–519 Pseudorandom functions, 410, 423, 424, 438,
via block-chaining, 519–521 450–452, 523–532, 556–558, 768
via tree-hashing, 521–523 generalized notion, 556, 768
collision-resistent. See collision-free, 513 non-uniform hardness, 411–412
Universal. See Hashing functions Verifiable, 590
Universal One-Way, 513, 543, 560–575, Pseudorandom generators, 404
588 Computational indistinguishability. See
Hashing functions, 527–537, 563–565, 596 Computational indistinguishability
AXU, 535–537, 589 non-uniform hardness, 392
collision probability, 528–531, 535 on-line, 407–408, 534–537
generalized, 530–531, 589
Hybrid Argument. See Techniques Quantum cryptography, 477

Interactive Proofs Rabin function, 766

perfect completeness, 658 hard-core, 422
Zero-Knowledge. See Zero-Knowledge Random Oracle Methodology, 478, 586–587
Random Oracle Model. See Random Oracle
Message authentication, 423, 497–537 Methodology
attacks and security, 502–507 Reducibility Argument. See Techniques
basic mechanism, 501–502 RSA function, 766
length-restricted, 507–516 hard-core function, 416
state-based, 531–537, 548, 585
Secret Sharing, 489, 730–731
NIZK. See Zero-Knowledge Verifiable, 729–735, 737–740, 752
Non-Interactive Zero-Knowledge. See Signature schemes, 497–523, 537–598
Zero-Knowledge attacks and security, 502–507
Non-uniform complexity, 378–393, 402, authentication-trees, 537, 545–560
618–619, 620, 622 basic mechanism, 501–502, 538
Fail-stop, 583–584
Oblivious Transfer. See Cryptographic incremental signing, 581–583
protocols length-restricted, 507–516
One-way functions, 423, 525, 538, 539–542, memory-dependent, 546–556, 559–560,
560–575 588
non-uniform hardness, 403, 411 off-line/on-line signing, 580–581
One-way permutations, 562, 563–565, one-time, 465–469, 538–575
570–571 super-security, 465–469, 576–580
claw-free collections, 516–519, 542, The refreshing paradigm, 537, 543–560
588 unique signature, 575–576
collection of, 765–768 Signatures. See Signature schemes
hard-core, 414–422, 431, 640–643 Simulation paradigm. See Techniques
modular squaring, 419–421 Synchronous communication, 603, 695, 777
RSA, 416, 766
with trapdoor, 403, 413–422, 423, Techniques
640–643, 648, 650, 765–768 Averaging Argument, 386
Hash and Sign, 513–516, 526–537,
Probabilistic encryption. see Encryption 542–543, 571–575, 576
schemes Hybrid Argument, 391, 402, 429, 448–449,
Probability ensembles, 379 457, 459–461, 467–468, 479, 593,
efficiently constructible, 394–403 637–638, 703, 754

797
 Free ebooks ==> www.Ebook777.com
INDEX

Techniques (cont.) Verifiable Secret Sharing. See Secret

Reducibility Argument, 385, 387, 402, Sharing
410, 510, 514, 518, 525, 540, 551,
564, 567, 569 Witness Indistinguishability, 782–783
the simulation paradigm, 379, 479, 601, Non-Interactive, 464–469
620 Strong, 768–772
Trapdoor permutations. See One-way
permutations Zero-Knowledge, 775–783
Composition of protocols, 775–780
Universal One-Way Hash Functions. See Concurrent composition, 777–780
Hashing for NP, 658, 664–671

798

www.Ebook777.com