Scribd
Upload
81 views

CSE 461: Error Detection and Correction: Next Topic

This document discusses error detection and correction across network layers. It introduces concepts like adding redundant data through error detection and correction codes to detect and possibly repair bit errors caused by noise. Specific error detection techniques covered include parity, checksums, cyclic redundancy checks (CRCs), and Reed-Solomon codes. The document compares retransmission-based error recovery with forward error correction and discusses factors like burst errors that influence the best approach. Key metrics like Hamming distance are also introduced.

Uploaded by

Shahban ktk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
81 views

CSE 461: Error Detection and Correction: Next Topic

This document discusses error detection and correction across network layers. It introduces concepts like adding redundant data through error detection and correction codes to detect and possibly repair bit errors caused by noise. Specific error detection techniques covered include parity, checksums, cyclic redundancy checks (CRCs), and Reed-Solomon codes. The document compares retransmission-based error recovery with forward error correction and discusses factors like burst errors that influence the best approach. Key metrics like Hamming distance are also introduced.

Uploaded by

Shahban ktk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

CSE 461: Error Detection and

Correction

Next Topic

 Error detection and correction

 Focus: How do we detect and correct Application


messages that are garbled during Presentation
transmission? Session
Transport
 The responsibility for doing this cuts Network
across the different layers Data Link
Physical
Errors and Redundancy

 Noise can flip some of the bits we receive


 We must be able to detect when this occurs!
 Who needs to detect it? (links/routers, OSs, or
apps?)

 Basic approach: add redundant data


 Error detection codes allow errors to be recognized
 Error correction codes allow errors to be repaired too

Motivating Example
 A simple error detection scheme:
 Just send two copies. Differences imply errors.

 Question: Can we do any better?


 With less overhead
 Catch more kinds of errors
 Answer: Yes – stronger protection with fewer bits
 But we can’t catch all inadvertent errors, nor malicious ones

 We will look at basic block codes


 K bits in, N bits out is a (N,K) code
 Simple, memoryless mapping
Detection vs. Correction

 Two strategies to correct errors:


 Detect and retransmit, or Automatic Repeat reQuest.
(ARQ)
 Error correcting codes, or Forward Error Correction
(FEC)
 Retransmissions typically at higher levels (Network+).
Why?

 Question: Which should we choose?

Retransmissions vs. FEC

 The better option depends on the kind of errors and the


cost of recovery
 Example: Message with 1000 bits, Prob(bit error) 0.001
 Case 1: random errors
 Case 2: bursts of 1000 errors
 Case 3: real-time application (teleconference)
The Hamming Distance
 Errors must not turn one valid codeword into another valid
codeword, or we cannot detect/correct them.
 Hamming distance of a code is the smallest number of bit
differences that turn any one codeword into another
 e.g, code 000 for 0, 111 for 1, Hamming distance is 3
 For code with distance d+1:
 d errors can be detected, e.g, 001, 010, 110, 101, 011
 For code with distance 2d+1:
 d errors can be corrected, e.g., 001  000

Parity

 Start with n bits and add another so that the total


number of 1s is even (even parity)
 e.g. 0110010  01100101
 Easy to compute as XOR of all input bits

 Will detect an odd number of bit errors


 But not an even number
 Does not correct any errors
2D Parity

 Add parity row/column to array of


bits

0101001 1
 How many simultaneous bit errors 1101001 0
can it detect? 1011110 1
0001110 1
0110100 1
 Which errors can it correct? 1011111 0
1111011 0

Checksums

 Used in Internet protocols (IP, ICMP, TCP, UDP)


 Basic Idea: Add up the data and send it along with sum

 Algorithm:
 checksum is the 1s complement of the 1s
complement sum of the data interpreted 16 bits at a
time (for 16-bit TCP/UDP checksum)
 1s complement: flip all bits to make number negative
 Consequence: adding requires carryout to be added
back
CRCs (Cyclic Redundancy Check)
 Stronger protection than checksums
 Used widely in practice, e.g., Ethernet CRC-32
 Implemented in hardware (XORs and shifts)

 Algorithm: Given n bits of data, generate a k bit check


sequence that gives a combined n + k bits that are
divisible by a chosen divisor C(x)

 Based on mathematics of finite fields


 “numbers” correspond to polynomials, use modulo
arithmetic
 e.g, interpret 10011010 as x7 + x4 + x3 + x1

How is C(x) Chosen?


 Mathematical properties:
 All 1-bit errors if non-zero xk and x0 terms
 All 2-bit errors if C(x) has a factor with at least three
terms
 Any odd number of errors if C(x) has (x + 1) as a
factor
 Any burst error < k bits

 There are standardized polynomials of different degree


that are known to catch many errors
 Ethernet CRC-32:
100000100110000010001110110110111
Reed-Solomon / BCH Codes

 Developed to protect data on magnetic disks


 Used for CDs and cable modems too
 Property: 2t redundant bits can correct <= t errors
 Mathematics somewhat more involved …

Key Concepts

 Redundant bits are added to messages to protect


against transmission errors.
 Two recovery strategies are retransmissions (ARQ) and
error correcting codes (FEC)
 The Hamming distance tells us how much error can
safely be tolerated.

You might also like