See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228864187

An Introduction to Digital Forensics

Article

CITATIONS READS

2 6,955

3 authors, including:

John P. Abraham
University of Texas Rio Grande Valley
15 PUBLICATIONS   75 CITATIONS   

SEE PROFILE

All content following this page was uploaded by John P. Abraham on 19 June 2014.

The user has requested enhancement of the downloaded file.

 An Introduction to Digital Forensics
Irma Resendez, Pablo Martinez, and John Abraham
University of Texas Pan American
[email protected]

Abstract

As new technologies develop criminals find ways to apply these technologies to

commit crimes. With the explosion of web technologies almost all major businesses in
the world have web presence thus exposing their data to legitimate and illegitimate users.
Computers have become intrinsic part of our lives. Businesses have streamlined their
operation saving millions of dollars because of the web technologies and services.
Neither the businesses nor the consumers can live without these technologies. Because of
the intricate involvement of computer technology in all aspects of our lives, it also has
become legal evidence in both civil and criminal cases.

Computer evidences admitted in courts could be any file or fragment recovered from
the storage devices such as email, browsing history, graphics, photographs, or application
documents. These files may be undeleted or deleted. Deleted file recovery would require
special techniques. Computer professionals trained in digital forensics preserve and
retrieve evidence in a non-destructive manner. Evidence may be recovered from any
storage medium installed in digital equipment such as computers, cameras, PDAs, or cell
phones. All forensic work should be done with care including documenting clear chain
of custody in order for the evidence to be admissive in a court of law.

Introduction

With the proliferation of computers in our everyday lives, the need to include
computer contents or traces as part of formal evidence has become inevitable.
Computerized devices are part of our world in the form of laptops, desktop computers,
servers, etc., but there are also many other storage devices that may contain forensic
evidence. Devices such as memory cards, personal digital assistants, and video gaming
systems, are among a myriad of devices that have the ability to accept input, provide
output, and also store data. It is this data or the usage of these devices that is at the center
of computer forensics.

According to Marcella and Menendez, cyber forensics, e-discovery, digital, forensics,

computer and computer forensics mean relatively the same thing yet none has emerged as
a defacto standard (Marcella, Menendez 2008). They further present a working definition
of computer forensics as ―the science of locating, extracting, and analyzing types of data
from different devices, which specialists then interpret to server as legal evidence (p5).‖
They further state that computer forensics can also be defined as ―the discipline that
combines elements of law and computer science to collect and analyze data from
computer systems, networks, wireless communications, and storage devices in a way that
is admissible as evidence in a court of law (Marcella, Menendez 2008).‖ Computer
forensics aims to attain as much information from electronic devices or media by
utilizing sound forensic techniques that may be admissible in court. This includes
concise and sound forensic techniques including a clear chain of custody and
documentation. There are two different areas that must be considered when collecting
digital forensic evidence. The first is the process of collecting the evidence without
altering its contents and ensuring it is admissible in court. The second area is the actual
use of law enforcement grade sound forensic practices that results in the collection being
admissible in court. It is not the intent of this document to address the latter. As the
authors are computer science professionals, it is the intent of this document to provide a
concise overview of the process of collecting forensic evidence and review different
methods, tools, and challenges during forensic analysis and collection.

Techniques – Live Incident response

Live incident response collects all of the relevant data from the system that will be
used to confirm whether an incident occurred. Live incident response includes collecting
volatile and nonvolatile data. Volatile data is information we would lose if we walked up
to a device and disconnected the power cord. Nonvolatile data includes data that would
be very useful to collect during digital forensic collection such as system event logs, user
logons, and patch levels, among many others. Live incident response further includes the
collection of information such as current network connections, running processes, and
information about open files. Live incident response is typically collected by running a
series of commands that produces data that would normally be sent to the console but is
sent to a different storage device or a forensic workstation. A forensic workstation is a
machine that the forensic investigator considers trusted and is used to collect data from a
suspected host computer. There are various ways to transmit the data to the forensic
workstation during forensic collection. Some of the most common tools to collect data
include commercial software such as EnCase and FTK (forensic toolkit from
accessdata.com) but there are many other open source and commercial tools that can be
part of our arsenal during collection.

Some of the most common tools include Netcat and Cryptcat. Netcat is the
considered by many as the Swiss army knife of live incident response collections. Netcat
crates TCP channels can be executed in listening mode like a telnet server, or in
connection mode like a telnet client. Netcat also includes MD5 check sum capability
which, as we will cover later, is an essential part of any sound digital collection order to
prove the data was not altered during collection. Netcat works by booting to a shell and
―listening‖ on a particular TCP port in verbose or listening mode. Once the forensic
workstation is in ―listening‖ mode via Netcat, the data is sent from the target host to the
particular port the forensic workstation is listening to. Cryptcat is a variant of Netcat that
is worth mentioning in this document as it encrypts all of the data across the TCP
channel. It uses all of the same commands and command-line switches as Netcat but
enhances Netcat by providing secrecy and authentication—which are now very important
in the age of eavesdropping and wire tapping.
Volatile vs. Nonvolatile data

Some of the volatile data that should be collected includes system date and time,
current network connections, open TCP and UDP ports, which executables are opening
UDP and TCP ports, cached NETBIOS name table, users currently logged on, the
internal routing table, running processes, scheduled jobs, open files, and process memory
dumps. This list is not all inclusive as a forensic investigator must consider any and all
possible variables during collection. However, one thing that all these have in common is
that they would be lost if the power were removed from your target machine. While there
are various methods of collecting volatile data, some of the most common and non-
commercial methods will be presented:
a. System date and time – can be collected by using the time and date
commands at the prompt.

b. Current network connections – can be collected by using the netstat

command with the –an flags to retrieve all of the network connections and
see the raw IP addresses instead of the FQDN.
c. Open TCP and UDP ports – use netstat –an. Also use FPort tool
(www.foundstone.com) in order to link the open ports to the executables
that opened them.

d. Users currently logged on – run PsLoggedOn from

www.sysinternals.com. This tool will return the users that are currently
logged onto the system or accessing the resource shares. PsExec can also
be used.
e. Internal routing table – netstat with the –rn switch.
 f. Running processes – it is important to identify running processes as they
could contain backdoors. Use pslist from PsTools suite.
g. Running Services – use PsService from PsTools suite.
h. Scheduled jobs – run at command at command prompt
i. Open files – run Psfile from pstools
j. Process Memory dumps – process memory dumps may provide critical
investigative material of a volatile nature. These include Command line
utilized by the intruder and remotely executed console commands and
their resultant output. Memory dumps can be captured using
userdump.exe provided by windows. It is available from Microsoft at:
http://www.microsoft.com/downloads/details.aspx?FamilyID=E089CA41-
6A87-40C8-BF69-
28AC08570B7E&displaylang=en&displaylang=en

Nonvolatile data includes information we would like to acquire including system

version and patch level, file system time and date stamps, registry data, the auditing
policy, a history of logins, system event logs, user accounts, IIS logs, and any suspicion
files. These can be acquired using the following methods:

a. System Version and Patch level – knowing which patches have been applied
to a server or workstation will enable us to narrow our initial investigation to
areas of high probability. Use PSInfo from pstools to query system
information. Switches include –h to show installed hotfixes, -s to show
installed software, and –d to show disk volume information.
b. File System Time and Date stamps (pg 31) – UnxUtils from
http://sourceforge.net
c. Registry Data – Leads we want to capture include programs executed on
bootup and entries created by intruder’s tools. We use RegDmp to capture the
complete registry.
http://www.softpanorama.org/unixification/registry/microsoft_registry_tools.s
htm#RegDmp
d. The Auditing Policy - Windows has auditing disabled by default. Use
auditpol included in Microsoft resource kits
e. System event logs – Psloglist from sysinternals
f. User accounts – use pwdump utility to dump user accounts.
http://foofus.net/fizzgig/pwdump
g. IIS Logs
ii. Most attacks happen over port 80
iii. You cannot block what you must allow in
iv. C:\winnt\system32\logfiles – IIS logs
UNIX Live Response

Any forensic investigator should be prepared to encounter non-windows operating

systems such as DOS, Linux, and UNIX. This section will concentrate on UNIX live
response. In order to collect volatile data, we can utilize the following commands during
a UNIX live response:

a. System date and time – date

b. Current network connections – netstat –an
c. Open files – Lsof lists open files, displays the files that the process has
open and using lsof-n lists raw IP addresses.
d. Running Processes – ps-aux will display the open processes
e. Internal routing table – netstat –rn
f. Mounted file systems – mount, df

Nonvolatile data during UNIX live response is geared at collecting similar

information as previously listed for Microsoft Windows:

a. System version and patch level – uname –a. In Linux and UNIX the patch level
is difficult to retrieve because it is dependent on the Linux distribution used to
create the system.
b. Users Currently Logged On - Users currently logged on are saved on:
/var/run/utmp log.
c. A history of Logins - Saved in the var/log/wtmp binary log.
d. System logs - In Unix, the syslog daemon listens for messages from either local
programs or other servers on the Internet and logs them according to the
/etc/syslog.conf configuration file.
i. In this case, the only relevant logs are /var/log messages and
/var/log/secure.
ii. Using Netcat we transfer the logs to our forensic workstation.
iii. Logs: contain five columns
1. Date
2. Time
3. Machine name
4. Process that initiated the event, along with the process
number
5. Message that was logged

Network Based Evidence (NBE)

Investigators collect 4 types of network based evidence including full content data,
session data, alert data, and statistical data full content data involves collecting all data
transmitted over a wire. It consists of the actual packets, typically including headers and
application information, seen on the wire or in the airwaves. It records every bit present
in a data packet. When looking at headers, analysts can identify subtle aspects of packets.
The presence of certain options or the values of certain fields can be used to identify
operating systems. Full content data presents the greatest opportunity for analysis, but it
also requires the greatest amount of work.

I. Session Data
a. Includes data that was created or modified during a particular user’s
session.

II. Alert Data

a. Alert data is created by analyzing NBE for predefined items of interest.
b. Example: a client using IP address 192.168.0.42 has queried the remote
procedure call (RPC) port mapper on a server, 192.168.0.40. This could
be the precursor to an attack because an intruder could use the information
to identify vulnerable RPC services on the server. The alert is a judgment
made by the network IDS (intrusion detection) that the packet it saw is a
query of the port mapper service.

III. Statistical data

a. Provides a big picture perspective
b. In the data world, statistical data has traditionally been used to measure
the health and performance of a network.

Common Forensic Analysis Techniques

When you conduct forensic analysis, there are a few steps that must be executed in
nearly every type of investigation to prepare the data for analysis. For instance, it is
usually recommended to recover any deleted files and add them to the analysis. It is also
advantageous to reduce the data set collected to the smallest number of files and add
them to the analysis for efficient review. Another step that should be incorporated into
the analysis is string searching to identify relevant files and fragments of relevant files.

Recovering deleted files is of crucial importance. Particularly in cases where the

target computer in question has been utilized by a savvy computer user or a suspect who
may have wanted to delete traces of their digital footprints. In order to avoid doing work
twice, recovering deleted files should be done first. There are several different options
and solutions that provide similar results which will be listed below. It is worth
mentioning that at the time of this writing, the authors of this document tested all
solutions presented here and discovered an interesting and extremely important fact that
should be of great importance during collection particularly when collection and
recovered data is to be presented as legal evidence. Although commercial software was
the easiest to use, it was unable to recover all deleted files. In some instances, utilizing
open source tools produced the recovery of additional deleted files after a first pass with a
commercial product. During our forensic tests, it was also found that in every instance,
commercial software was able to recover most of the deleted files when compared to
open source tools. It is then concluded that the best way to undelete the most data from a
target host is to utilize a combination of commercial and open source tools. The tools
tested at the time of this writing are the following:

I. EnCase & FDK– Windows-based commercial products that seems to be

dominating the forensic market at the time of this writing. The main
advantage to recover undeleted files using these tools is that they do not have
to be installed on the target host. They can search and undelete files based on
a forensic duplication (clone of the target host’s hard drive).

II. Linux Kernel – it is an open source tool that enables us to take a forensic
duplication and make it act like a real hard drive under Linux.
a. ftp://ftp.hp.nasa.gov/pub/ig/ccd/enhanced_loopback
b. Once the Linux kernel has been loaded and it recognizes our forensic
duplication as a mounted volume, we need to recover the deleted files.
The most used tool used to be The Corner’s Toolkit t
http://www.fish.com/tct but the limitation that it only recovers files from
Windows made it less popular. The most popular at the time of this
writing is TASK and The Sleuth Kit at: http://www.sleuthkit.org.

Production of Time Stamps and Other Metadata for Files

After undeleting files, the next step in forensic analysis is attaining metadata and time
stamp of files. Metadata includes full title names, file sizes, MD5 hashes, among others.
Commercial solutions such as EnCase and FDK produce metadata information by
default. In lieu of commercial solutions, there are open source tools such as GNU find
that produce the same information.

After attaining metadata, our next chore is attaining the MD5 checksum (define). As
in the case of metadata, commercial solutions produce the MD5 checksum by default and
there is little or no intervention from the forensic analyst. There are also open source
solutions one of which is the fls command included in the Sleuth kit.
Removing known files

All of the aforementioned steps may include thousands of files that we may need to
investigate. There are usually thousands of files associated with both operating systems
and application programs that, upon an integrity check by the forensic analyst, can be
removed in order to alleviate the load in the forensic analyst. Since during collection we
don’t know if the files we collect from alleged application programs and operating
systems are of trusted origin, we must compare the MD5 hashes of every file against a
known good installation (usually part of a forensic analyst’s arsenal) and good set of
hashes. If malicious code or an altered file exists masqueraded as a legitimate operating
system file, comparing the suspect file against the legitimate file’s MD5 checksum will
produce a red flag.

In order to attain a reputable source of files, we can either install the operating system
ourselves in a controlled environment or use somebody else’s work. Installing the
operating systems may present a great task due to the increasingly large number of
operating systems. An alternative is using a list of MD5 hashes that may be available for
this purpose. A well known source of such information is the National Software
Reference Library (NSLR) provided by the National Institute of Standards and
Technology (NIST) at http://www.nsrl.nist.gov. Comparing the MD5 hashes using opens
source solutions involves tedious manual work including removing quotations from all
file names, comparing lists, and writing a script (usually a Pearl script) to compare and
identify the anomalies. Commercial solutions such as EnCase have the ability to import
existing hash sets such as the ones provided by NSLR with very little user effort.

Web Browsing Activity Reconstruction

The last piece of forensic analysis that will be covered is the collection of web
browsing activity and its reconstruction. In this area, nothing beats commercial solutions.
EnCase, FTK and IE history at http://www.pillipsponder.com provide the necessary tools
to collect and reconstruct Web browsing activity.

Conclusion

Although still considered in its infancy, computer forensics has been on the forefront
of many recent legal investigations. As we digitize anything from video to live
television, digital forensics will not only inevitably become increasingly important to
legal investigations, but it may soon become a science with legislated methods. The most
important conclusion from this document should be the fact that there are no tools that
 are infallible. It is the job of a computer forensics scientist to utilize different tools in
order to recover and capture digital evidence in a forensically sound manner.

References

Marcella, Albert J, Menendez, Doug (2008). Cyber Forensics. Boca Raton, FL:
Auerbach Publications Taylor & Francis Group.
Gialanella, David. (2008) New Tech, Old Problem. ABA Journal, 94(8), 35.

View publication stats