Optimised malware detection in digital forensics

International Journal of Network Security & Its Applications (IJNSA), Vol.6, No.1, January 2014

Optimised Malware Detection in Digital Forensics
SaeedAlmarri and Dr Paul Sant
Institute for Research in Applicable Computing, University of Bedfordshire, Luton,
United Kingdom
Associate Dean, University Campus Milton Keynes, Milton keynes, United Kingdom

Abstract
On the Internet, malware is one of the most serious threats to system security. Most complex issues and
problems on any systems are caused by malware and spam. Networks and systems can be accessed and
compromised by malware known as botnets, which compromise other systems through a coordinated
attack. Such malware uses anti-forensic techniques to avoid detection and investigation. To prevent systems
from the malicious activity of this malware, a new framework is required that aims to develop an optimised
technique for malware detection. Hence, this paper demonstrates new approaches to perform malware
analysis in forensic investigations and discusses how such a framework may be developed.

Keywords
Denial of service (DOS), Wireshark, Netstat, TCPView, The Sleuth Kit (TSK), Autopsy, Digital Forensics,
Malware analysis, Framework

1. Introduction
Over the last decade, there have been noteworthy improvements in techniques to detect malware
activities [1]. Loading and distributing executable files over the Internet always presents a risk to
the overall security of the system [2]. Malware programmes can be installed by attaching hidden
malicious code in an innocuous file or application. The code can then be activated by a remote
programmer with the aim of threatening the existing system. According to a study by Islam et al.
on the risk of downloading [3], of more than 450,000 files downloaded, approximately 18%
contained malware programs. They also investigated whether different code investigation
techniques yielded the same results. Astonishingly, they found that there were many cases where
forensic investigatory tools were unable to detect the malware content of the infected files.
A significant amount of effort has been expended on developing techniques to perform robust
computer forensic investigations [6]. Such effort has focused on collecting, analysing and
preserving evidence of malware activities, for e.g. a study on botnets [4] and a study of
executable spyware and client-sided honeypots[5] also illustrated defensive mechanism for
securing a system both on the client and server side access. Other reports mentioned in [3][6]
have also focused on acquiring large and diverse samples of malware to enable researchers and
forensic experts to understand their nature and its rationale. Some existing tools like ERA
remover, conficker, etc. can execute hidden and anonymous files and monitor their behaviour.
These tools provide protection from all threats related to the malware functioning in the system.
According to reports by Kasama et al (2012), a single piece of malware can compromise and
infect the entire network system. Thus, protecting systems from unwanted malicious code can be
considered as one of the most critical concerns in information security [6].
DOI : 10.5121/ijnsa.2014.6101

01


Various protocols and services have been implemented to simulate the behaviour of an end user
in a compromised system, which assists researchers to analyse and relate malware behaviour in
order to develop a better detection tool. However, different encryption techniques and complex
programmes make it difficult to detect the activity of malware [1][3]. Commercially, there are
various tools existing, but they are not being used to detect in an optimal way. Hence, this makes
malware a serious threat to the domain of information technology and associated services [7].

1.1 Problem Statement
Malware is an intended program that performs malicious activity. It can be easily installed in any
computer with a malicious intent. In a most apt scenario [2], malware attacks by organised crime
syndicates largely target financial institutes and the banking industry, where they attack on their
online software services that deal with the system of monetary assets ownership. Cybercriminals
also enable malware attacks on customers and businesses dealing with financial institutions [2].
Bradford &Yegneswaran (2007) mentioned their report on the increase in malware activity
through the use of sophisticated tools and methods, making it complicated for its detection
[8].Also in a study by SANS institute (2012) illustration on the challenges in digital investigation
on malware was mentioned. According to the report, gathering and analysing large chunks of
data, obfuscated malware or malicious code are usually time consuming, costly and requires
different techniques The situation becomes even more complex when the same data are present
over a large network or various computational systems. This makes its challenging for an
investigator to gather evidence of illegal distribution with pre-existing forensic tools, pre-defined
malware detection systems and a limited duration at a workstation [9].
In cases where numerous systems or parties are involved in the same crime, the analysis of
independent digital data during the investigation can result in the loss of indispensable correlated
evidence. The loss occurs due to the incapability of digital forensics to craft correlations between
multiple malware cases. It is exceptionally difficult to detect malware and to obtain accurate data
due to numerous obfuscation techniques used by the programmers [10].
As mentioned in [10], this problem can be addressed by focusing on the various digital forensics
techniques currently used to extract, gather and analyse malware-infected systems. A new
correlational technique that focuses on typical features of malware activity and that is capable of
accumulating sufficient forensic evidence against the perpetrator and formulating it to produce
against the perpetrator in the court is needed. The goal of our research is to optimise the research
paradigm for malware analysis and to improve the investigatory experience by employing an
active approach to detect malware programmes. The next section of the report explains the
existing techniques for malware detection. With this in mind, we focus on developing a robust
framework for malware analysis and detection.

2. Malware Detection and Digital Forensics
Digital forensics and malware detection exhibit many similarities. Both involve techniques for
deep and extensive data mining. Comparisons between the two can be exploited to collect
evidence on any incident related to malware activities. Thus a similarity-based technique is
required to make the process of detection much quicker and easier for an investigator to get
results [11]. In contrast to digital forensics, studies on malware detection aim only to find
pathways and traces of malware activity and not the operation of the malware in a live system.
The first step should consider the technical principles of digital forensics, as it can formulate
elements to construct the research base of the investigation. Furthermore, it helps in robust
explorations of the traces of different types of malware present in the system [11].
2


2.1. The digital forensics process
rocess
The digital forensics process, which was introduced in 2001–2002 at the SANS Institute includes
2001
Institute,
specific categories that begin with the identification phase and end with the final decision stage
[12]. The different categories in digital forensics are illustrated in Figure 1.

Figure1.V
.Various categories in a digital forensics process

2.2.

File system analysis

The file system refers to the organisation of various records contained in a system’s database
organi
model. An analysis of the file system is a common process in digital forensics, as it provides
ile
valuable hidden information about system accessibility [13]. At the stage of data collection the
.
collection,
file system analysis extracts distinctive layers of data from the hard disk for later examination.
Figure 2 illustrates the layers of abstraction by Farmer and Venema [13]. It shows three layers: (a)
the user and the application view, (b) the file system view and (c) the hardware view. The data
view
obtained from these layers can be combined to discover unseen, obfuscated malicious files. The
process is advantageous for malware analysis in a case where data is kept hidden or encrypted
behind the data files.

Figure 2. V
. Various distinctive layers of abstraction [13]

3


2.3.

Characteristics of malware

In both digital forensics and malware detection, it is vital to acquire sufficient data and definitions
to make the detection system more efficient and the malware easier to recognise. Based on
knowledge obtained from analyses of various malware domains, it is possible to recognise
different patterns of malware, thus allowing an investigator to build a profile of the case.
Different techniques and methods for malware detection proposed in [14][15] can be used to
develop a new framework for a robust system. These techniques include methods such as time
stamping, entropy analysis (File type, hashing, etc.) and obtaining traces from keywords and
identifiers.

•
•

•

Time stamping: This method includes an analysis of the duration between the attack and the
initial phase of the investigation of the malware infection.
Entropy analysis: The most commonly infected files are executable files. In some cases, the
infection of these files is very difficult to detect. The infection may be hidden in the indexes,
recycle bins or system folders. An example of this is Windows/system32.
Keywords and identifiers: The simplest way of identifying malware code is via keywords,
called ‘identifiers’. These identifier scan be used to obtain data from IP addresses, email
addresses and other sources to get the required information mentioned in a communication
pattern. For example, ‘key logger’ can be a string used to find hints regarding any malware
attack on the system.

3. Approaches to Malware Detection and Analysis
In digital forensics, any investigation includes two types of functions: ‘D’, for malware detector,
and ‘P’, for the programme to be investigated. The maximum range for the D function relies on
the range of programme P, which is defined as SET {malicious, genial} [16].
The function of detector D is to scan the programme P to check whether the P is malicious or
genial. D detects the malware on the basis of the machine code, which stores the particulars, or
the ‘signature’, of the malware. The investigation may yield one of three outcomes: false positive,
false negative or the hit ratio [16].

•

•

•

False positive: A false positive is when a scanner detects malware in a non-infected file. This
means that the characteristics of a particular malicious code are not unique and hence the
code appears in a non-infected file.
False negative: This occurs when a detector fails to expose a particular malicious code in an
infected domain. This usually happens when a scanner lacks a signature sample of a
particular malware and hence fails to detect the infection.
Hit ratio: The hit ratio occurs when the detector detects the malware accurately. This is
possible only when the signature of a particular malicious code matches that of a stored
sample in the detector’s database.

Digital forensics distinguishes two types of malware detection methods: signature-based detection
methods and anomaly-based methods.

3.1.

Signature-based detection and analysis

Signatures are combinations of bytes, which are a constituent part of malware, or a malicious
code. Malware is categorised into three forms: basic, metamorphic and polymorphic.
Polymorphic is the most common. As this type of malware hides its identity and source of
4


generation, it becomes very hard for investigators to collect evidence against the plotter [16]. In
basic malware, the source, or the ‘initial point’, of the programme is reformed, and complete
control of the system is transmitted to the malicious payload. In polymorphic malware, the codes
are mutated, while the original data in the system remain intact. This type of malware virus is
replicated via a model-based engine, and thus the virus mutates every time the programme is
executed in the same system. Metamorphic malware is based on reprogramming, which
modulates the characteristics of the parent programme. This produces a new signature for each
child variant that is created at a later stage [16].

Problems
•
•
•
•

Extracting and dispensing the signature is very complex.
Signature-based detection and analysis requires the investigator to perform penetrating
research and analysis, which is only possible manually and in a controlled environment.
Bypassing a signature during the investigation is a very common problem. Hence, new
signatures should be updated every time before conducting the analysis.
The ever-growing repositories of malware signatures are becoming very difficult to manage.

3.2. Specification-based detection and analysis [17]
This type of detection system depends on the particular specification of the implementation and
the deployment of a particular application or system file. The process works by learning all
phases of the application development and assuming that any abnormality carries malicious code.
The process uses the concept of reverse engineering, which focuses on reaching the source of the
programme. Sometimes it takes multiple runs to debug the source coding.

Problems
The main problem with this technique is its accuracy. The whole process depends on pre-defined
rules to follow, which makes it hard to conduct practical analysis. Moreover, rules for the process
need to be updated on a regular basis, as it is difficult for an investigator to learn any of it.

4. Requirements for the New Framework
Based on the above conditions faced by investigators, a new framework is required to optimise
the results. This should be integrated into a system to detect and analyse malware efficiently and
effectively. To ensure that analysis is comprehensive, it should perform both as a detector and a
warning system. To understand the precision and the nature of the requirement, an extensive
market research is required in the field to gather information for analysis.

4.1 Market Research
To find the precise problem statement for the research, qualitative and quantitative approaches
were implemented using a questionnaire-based methodology. During the research, a questionnaire
was distributed through email to respective respondents. The questionnaire used a framework
based on a scale of items to assess data gathered from the respondents. A pilot study was
accompanied to authenticate the questionnaire, where 30 completed questionnaires were
collected. Around 43% of the respondents were professionals from IT security, 33% were
forensics experts, 17% were network administrators and 7% belonged to other IT fields. Below
are some of the sample questions:

5


1. How often you receive case studies related to Malware for investigation?
<5%
5% to 20%
20% to 50%
50 to 80%
80% Above

2. How many of the above cases were solved with accurate results
<5%
5% to 20%
20% to 50%
50 to 80%
80 Above

3. What is your opinion on level of skills and technological tools adopted in your
company to carry investigating process on malware?
Weak
Average
Strong

4. Do you think there is a need of a new customized tool for malware detection
and analysis?
Yes
No
Can’t say

4.2 Discussion on the Result
The result from the observations reflects the vulnerabilities in the existing tools and
methodologies present for malware detection. Although most of the experts were aware of the
criticality and complexity of malware detection, still no specific solution for accurate detection
was obtained. Some of the analysis report is explained in the following tables.
Table 1. Important factors from the analysis

Factors
Changes in Malware Landscape

Yes
52%

No
13%

Cant Say
35%

Organizational vulnerabilities
Expertise on Malware
Detection of malware by non-technical employees

47%
67%
17%

20%
27%
53%

33%
6%
30%

6

Table 2. Level of Skills in Investigation.

Weak
13%

Average
60%

Strong
25%

Table 3. Level of accuracy during investigation.

<5%
7%

5%to
20%
20%

20% to
50%
43%

50 to
80 %
27%

80 above
3%

Table 4. Choice of tools.

Open
Source
7%

Shareware

Commercial

0%

43%

Cant
Say
7%

Table 5. Need of new tool.

Yes

No

90%

3%

Cant Say
7%

Table 6. Preferable Methods.

Dynamic
57%

Static
33%

Other
10%

4.3. Result on Analysis
The results of the observations reflect the vulnerabilities in the existing tools and methodologies
present for malware detection. Although most of the experts were aware of the criticality and
complexity of malware detection, no specific solution for accurate detection was obtained. Some
important factors from the analysis were identified that provide a clear view of the problem
statement. The analysis is as follows:

•
•
•
•
•
•
•

•

The core problem faced by the investigators was to find the level of malware infection in the
file.
Most importantly, investigators required using other parameters, such as Malware signature
and behaviour in combination, for optimised detection.
Most of the infections were both critical and complex. Thus, there is a need for a prerequisite
method to reduce the associated complication.
Although there are tools that are capable of detecting malware and other malicious code, most
of them produce inaccurate or incomplete results.
The cost of the tool played an important role in its adoption.
Most of the investigators wanted to use tools based on both dynamic and static methodology.
It usually took less time for the investigators to detect the presence of malware, but more time
to investigate it forensically.
Managing and updating the malware database is a task that should be given priority.
7


5. Proposed Framework
The required framework is designed for developing an optimised malware detection and analysis
tool, which is based on the analysis of and the results obtained from the secondary research
conducted in the survey report. The entire research pattern is divided into three distinctive phases.
This distinctive framework defines the various policies, categorises the organisational assets and
creates a new environmental variable for the process. The framework is designed for developing
an optimised malware detection system that can confirm the malware infection using any standard
of deployment. It can also assist in developing a robust and redefined metrics model that can
work with a three-stage analysis covering all operational activities within the system. It is an
unbiased model that initialises its operation by first confirming the malware infection, and then by
analysing, detecting and proliferating it. The objective for developing this framework is to
understand the concepts of the detection costs associated with dealing with malware discovery
and analysis in an associated environment in order to provide accurate and precise information
that can be used to increase the performance matrices. The entire process is divided into three
phases:

Phase 1: Malware Acquisition Function,
Phase 2: Detection and Analysis
Phase 3: Database Operational Function.
Below is the diagram that defines the relationship between the different phases of the process.
Each phase is explained by keeping the process of implementation in mind.

8


Figure 3. Three
hree-phase process for malware detection and analysis

Phase 1: Malware acquisition function
cquisition
Phase 1 initiates its task by confirming the occurrence of malicious code in the data. It deals with
operations concerning the confirmation of the data that is infected by the malware. It also
includes the functionalities that contain components that can detect both active and passive
infections in the associated system. This phase includes the following components:
•

•

•

•

Notification: Acquisition can begin in a variety of ways; including a notification alert sent
from a third party (such as CMS-cash management systems or an online payment processor)
CMS cash
or it can be an alert from an endpoint suit. It can act as a platform to determine whether or not
the issues occurring in the system are realistic [21].
syste
Quarantine: The main objective of this component is to collect the damaged file from the
system and then remove it separately from the rest of the associated files. This activity can
prevent any duplication/replication of malicious code from being transferred into another
non-infected file [21].
Triage process: This process only works for identifying the critical nature of the issue that
occurred due to the malware infection. It contains all sorts of analysis processes to categorise
the problem by level of seriousness (high degree of seriousness or low degree of seriousness)
[22].
Infection confirmation: Up to this point in the process, we have obtained enough
information to know whether or not the selected data is infected. So, now the detection tool
9


makes a decision and the next appropriate action is determined [22].
Depending on the information that is found during Phase 1, the following actions can occur:
•
•
•
•

If the selected data is not infected, then the entire process can be stopped and things can
revert to normal.
If a malware infection is detected, we can decide whether to conduct an analysis of the
selected malware infection or jump directly to the database remediation.
If we decide to analyse the data, we must be sure to analyse it properly in order to obtain
more information about the infection [23].
If we identify the infection and do malware proliferation, we must check whether the selected
infection matches up to the database; if it does not, then operations need to be performed to
add the new information to update the existing malware profile [23].

Phase 2: Detection and analysis
Based on the results from the previous phase, it can now be confirmed that the files are infected
by the malware. Therefore, the question arises as to how we can go about obtaining information
about the activity that the phase was trying to perform on the system, whether the attack consisted
of an attempt to steal a password or an attack on the confidentiality of the data. If it is difficult for
an organisation to determine the exact level of threat, a complete investigation is required in order
to fully analyse the infected file.
Detection and analysis constitutes functional output from the first phase (malware acquisition).
After confirming the presence of a malware infection, the malicious code can be investigated and
its characteristics can be analysed [25]. Defining the characteristics helps in proliferation, which
can contribute to developing and updating the malware database (mentioned in Phase 3). The vital
testing sub-functions include:
Build a control environment
Before doing any analysis, it is important to develop a control environment that can be used to
perform a thorough investigation. This is mainly a one-time operative process, but with changes
in the requirements, one must engage in an on-going analysis in order to achieve the target goals
[24]. The following pre-defined guidelines are required:
•
•
•
•

•

Isolate the infected file: Keep the infected file away from the live data. The infected file should
be kept isolated in a separate network, as doing so can prevent a replication of the infection.
Provide a separate network: During the investigation, it is important to have access to a
separate Internet service provider system.
Testing tools: Tools used for creating a previous testing environment should be cleaned perfectly
before starting to work on them again.
Keep a log file: A log file should be considered to be a special preference in the investigation, as
it carries important information about the malware activities. Log files record each step that the
malicious code used to infect the data file.
Sandbox: It is always preferable to perform testing on a separate sandbox, as it becomes easy to
conduct a specific investigation on the memory file that may contain a malicious infection.
Static Analysis
This function conducts a core analysis process on the suspected file. All results identified by this
function can be added to the database carrying the malware proliferation function. Malware also
10


leaves evidence behind; thus, a thorough investigation of the executable files can be used to
identify future threats to the system. Activities can be monitored internally in the system in a
variety of ways, including network accessibility and performing operations on the file system and
on the internal storage. To optimise the static analysis process, the following guidelines need to
be followed (mentioned in [25]):
•

•

•

•

Initial analysis: This includes investigating the obvious symptoms of attack in order to
identify the nature of the infected file. First, we try to match up the fingerprint of the file to
identify any existing malware profile in the database. This can usually be done using MD5
hashing.
Classifying the format of the file: While analysing the Windows files, every executable file
contains a schema, which may yield specific information about the cracker and its intention
for the attack. Analysis of information such as version information, menus and calling
functions can reveal a lot of evidence [25].
Analysing text strings: It is always important to scan the isolated text strings that are present
in the infected file. Using various keywords in the built-in search field can help in obtaining
the associated evidence.
Debugging: The last step for static analysis is to conduct a process of disassembly. This helps
determine the patterns that can assist in identifying the attacker [25].

Dynamic analysis
As mentioned above, static analysis is an initial way to investigate malware in an infected file.
We need to maintain the goal of creating a robust malware profile that can prevent all future
malware attack attempts. Although malware infection can be detected at the earliest stage, one
still may not be able to identify the unauthorised modulations that occurred and that compromised
the system after the attack. So, here we also apply the dynamic analysis process. The dynamic
analysis typically gathers information on the following (mentioned in [23]):
•

•

•

•

Memory: Malware can bring changes to the buffer flow and can also tamper with the main
memory of critical programmes. Thus, a deep analysis of the volatile memory can assist in
obtaining good information about how/why malware activities were accessing internal
memories.
Investigation registries: This involves looking for the modifications that were done in the
registry keys. The victimised machine can illustrate the exact changes that occurred in the
machine, which can help the investigator detect easy changes in system.
Investigation processes and running services: This can show all the processes that were
either started or stopped intentionally by the malware. Analysing the processes illustrates
some of the initial processes that can suggest the track record (origin) of the infected file in
the system.
Looking for virtual machines: Keeping an eye on the attached virtual machine (VM) can
help an investigator continue the research process. Malware can go dormant after VM is
detected, so the investigator is required to check for the presence of any VM on the hardware
or a VM that is running dissolutely inside the system.

Dynamic analysis can be further differentiated into three types:
•
•
•

Network analysis
Device scanning
File content analysis

While the static function can assist in analysing the features of executable files, the dynamic
11


function reads the metadata, which include information about the size of the file and its format
and information about the library management system. These data can be accessed to detect
malware in the infected file.

Phase 3: Database operational function
Database management is required to ensure that samples of the malware and information about
the malware are available. This information can be used to update the profile of a particular
malware, which can contribute to the process of performing an analysis and detecting the
mechanism. Any failure in malware notifications (in Phase 1) can be remediated to the database
operation in order to monitor the stage of reinfection of the malware in the system [26].
Malware profiling includes the following:
•
•
•
•

Adding the aggregated information.
Packaging the profile to make it more reliable for screening and scanning.
Categorising the profiles so as to make it convenient for other investigators to access the
information.
Revisiting the malware profile periodically in order to prevent malware reinfection (an
updated malware profile is another important function to consider).

6. Implementation of the Approach
By emphasising the described approaches in section 5, an experimental operation and evaluation
can be undertaken to develop an optimised technique based on pre-defined correlational
methodologies for robust malware detection. The following steps can be undertaken during the
experiment:

6.1File content analysis
A file is an important part of the system, as it carries data processed by the user that demonstrates
all activities done on the system before it was compromised. Hence, analysing file content can
provide a major clue about the source and the characteristics of the malware.
One major problem faced by all previous detection techniques is the inability of the system to
recognise obscure malware (e.g. malicious codes that are fully encrypted). In such cases,
analysing the file content is always preferable compared to the previously mentioned approaches.
File content analysis can be done either by calculating the entropy of the file or by analysing the
particular string structure [15].
The entropy calculation method can be used to examine compressed files that are encrypted and
hence modulated into different formats [15]. In the proposed scenario, the entropy calculation for
any file type can be done with the ‘Ent tool’, which presents the difference in the sizes between
an original file and an infected file in a tabular formation. String method analysis involves
searching for particular keywords. Such analysis can be undertaken with the ICAT tool (from
TSK), where TSK stands for The Sleuth Kit and Autopsy Browser (which acts as GUI for TSK)
and can be employed to search for strings [18]

6.2Network analysis
A large amount of malware is distributed over the network to create redundancy and to ensure
12


effective propagation. An optimised network should be able to detect malware and other
malicious codes. Trojan is the most commonly distributed malware. In this study, the netstat
command and network packet analysis are discussed as follows:
•

Netstat command

This performs the analysis on the TCP/IP network protocol. The conditions of the connection
before and after deployment of malware on the system will provide a comparison result between
the two. Deploying TCPView (for listing TCP and UDP endpoints) and Port Explorer (for
exploring and analysing network sockets) can generate logs and records, which will be matched
up with the original Listening PID to detect malware functioning [19].
•

Network packet analysis

This technique investigates network traffic flowing on the network channels. Malicious code
showing unusual activity enables evidence to be collected on the server side [20]. For performing
the analysis, the Wireshark tool will be deployed on a virtual machine to explore the traffic on the
suspected host. The main advantage of using a packet sniffer is to obtain knowledge about the
DNS protocol packets and connections established between the machine and the IP address
generator [20].
The proposed approach provides a robust framework for detecting malware. The following
are its characteristics:
•
•
•

It avoids allowing data chunks to accumulate, making it easier for investigators to flexibly
detect malware without wasting much time.
It focuses on distributed networks, as most other methodologies avoid getting into network
data.
With this approach, it is feasible to track the source of the malware, which can further prevent
the chance of other malware attacks.

7. Conclusion
This paper discussed a new framework and approach for malware detection. Open source tools
can be used in the extraction, investigation and analysis of the data. The report focuses on
malware detection techniques used in digital forensics and data mining. The techniques
mentioned in Section 4 can be used to perform experiments; hence, results pertinent to live
systems will be generated. The overall outcome of the study is multi-fold, where the intention is
to discuss various efficient and optimised techniques for use with malware detection. By
recognising the faults in the existing systems, the new framework can overcome the limitations
and thus can assist an investigator in obtaining evidence of malware to get an optimised result.

References
[1]

Carrera,E. &Erdelyi. (2004). ‘Digital genome mapping: Advanced binary malware analysis’, Virus
Bulletin, pp. 175-186.

[2]

BITS (2011) ‘Malware Risks and Mitigation Report’ ITS/The Financial Services Roundtable 2011,
Retrieved from http://www.nist.gov/itl/upload/BITS-Malware-Report-Jun2011.pdf

[3]

Islam, N.,Anand, R., Jaeger, T.,&Rao, J.R. (2009). ‘A flexible security system for using
Internet
content’,Software,
IEEE,
Vol.
14,
No.5,
pp.
52,59.
Retrieved
from
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=605931&isnumber=13290
13

[4]

Jehyun, L., Jonghun, K., Hyo, S., &Heejo, L. (2010). ‘Tracking multiple C&C botnets by analysing
DNS traffic.Secure Network Protocols (NPSec)’,6th IEEE Workshop, pp.67, 72.

[5]

Hayatle, O.,Otrok, H., & Youssef, A. (2012).‘A game theoretic investigation for high interaction
honeypots’,Communications (ICC), 2012 IEEE International Conference, Vol. 10, pp.66626667.Retrieved
from

[6]

Kasama, T., Yoshioka, K., Inoue, D., & Matsumoto, L. (2012). ‘Malware detection method by
catching their random behaviour in multiple executions’,Applications and the Internet (SAINT),
IEEE/IPSJ 12th International Symposium, Vol. 2, pp.262,266. Retrieved from

[7]

Mehdi S, Tanwani A, Farooq M (2009), “IMAD: in-execution malware analysis and detection”,
Proceeding of the 11th Annual conference on Genetic and evolutionary computation, pp: 1553-1560

[8]

Barford,P.,&Yegneswaran,V (2007). An inside look at botnets. Special Workshop on Malware
Detection,Advances in Information Security, IEEE Journal.

[9]

Zahn K (2012), “Case Study: 2012 DC3 Digital Forensic Challenge Basic Malware Analysis
Exercise”, SANS Institute Reading Room site, Retrieved from https://www.sans.org/readingroom/whitepapers/malicious/case-study-2012-dc3-digital-forensic-challenge-basic-malware-analysisexercise-34330

[10] Lim Y; Ryu H; Choi K; Park C; Park W; Kook K (2012) "A Study on Malware Detection System
Model Based on Correlation Analysis Using Live Response Techniques," Information Science and
Applications (ICISA), 2012 International Conference, vol., no., pp.1,6, 23-25
[11] Park J, Kim M, Noh B, Joshi J (2006), “A Similarity based Technique for Detecting Malicious
Executable files for Computer Forensics”, information Reuse and Integration, 2006 IEEE
International Conference, vol., no., pp.188,193, 16-18
[12] Ryder, K. (2002). ‘Computer forensics – We’ve had an incident, who do we get to investigate?’SANS
Institute, GSEC Certification Assignment Version 1.3.
[13] Farmer, D. &Venema, W. (2004).Forensic Discovery.Addison Wesley Professional, USA
[14] Casey E (2002), “Error, Uncertainty, and Loss in Digital Evidence”, International Journal of Digital
Evidence
Summer
2002,
Volume
1,
Issue
2,
Retrieved
from
https://utica.edu/academic/institutes/ecii/publications/articles/A0472DF7-ADC9-7FDEC80B5E5B306A85C4.pdf
[15] Davis T (2009), “Utilizing Entropy to identify Undetected Malware”, White paper by Guidance
Software,
cyber
security
solution,
Retrieved
from
http://image.lifeservant.com/siteuploadfiles/VSYM/99B5C5E7-8B46-4D14A53EB8FD1CEEB2BC/43C34073-C29A-8FCE-4B653DBE35B934F7.pdf
[16] Neelakantan, S.&Rao, M. (2008). ‘Threat-aware signature based intrusion-detection approach for
obtaining network-specific useful alarms’,Internet Monitoring and Protection. The Third International
Conference,
Vol.
2,No.
3,
pp.80,85.
Retrieved
from
[17] Vinod P, Laxmi V, “Survey on Malware Detection Methods”, Department of Computer Engineering,
Malaviya
National
Institute
of
Technology,
Retrieved
from
http://www.security.iitk.ac.in/contents/events/workshops/iitkhack09/papers/vinod.pdf
[18] Dowling A (2006), ‘The Sleuth Kit v2.01 and Autopsy Forensic Browser Demonstration’, White
paper
on
TSK
manual,
Retrieved:
http://www.sjones.co.nz/downloads/Files/Forensics/TSK_v201_Demonstration.pdf
[19] Vigna, G., & Kemmerer, R.A. (2008).‘NetSTAT: A network-based intrusion detection approach’,
Computer Security Applications Conference, Vol. 7, No. 11, pp.25,34.
[20] Broadway, J., Turnbull, B., & Slay, J. (2008). ‘Improving the analysis of lawfully intercepted network
packet data captured for forensic analysis’,Availability, Reliability and Security, 2008.
ARES
08. Third International Conference, Vol. 4, No. 7, pp.1361,1368.
14

[21] Park, S. (2012).‘Malware expert: Execution tracking’, Cybercrime and Trustworthy Computing
Workshop (CTC), 2012 Third, Vol. 48, No. 55, pp.29-30.
[22] Blount, J.J.,Tauritz, D.R., Mulder, S.A. (2011). ‘Adaptive rule-based malware detection semploying
learning classifier systems: A proof of concept’, Computer Software and Applications Conference
Workshops (COMPSACW), 2011 IEEE 35th Annual, Vol., No., pp.
110,115, 18-22.
[23] Adeel, M.&Tokarchuk, L.N. (2011). ‘Analysis of mobile P2P malware detection framework through
Cabir&Commwarriorfamilies’, Privacy, security, risk and trust (passat), 2011 IEEE Third
International Conference on SocialComputing (Socialcom), pp.1335,1343, 9-11.
[24] Chen, L., Liu, B.,Hu, H. &Zheng, Q. (2012). ‘A layered malware detection model using
Trust, Security and Privacy in Computing and Communications (TrustCom), 2012 IEEE
International Conference, pp.1259,1264, 25-27.

VMM’,
11th

[25] Cavallaro L, Saxena P, R. Sekar (2008), “On the Limits of Information Flow Techniques for Malware
Analysis and Containment”, Computer Science Department University of California at Berkeley,
Retrieved from http://www.comp.nus.edu.sg/~prateeks/papers/saxena-dimva08.pdf
[26] Moser, A.,Kruegel, C.&Kirda, E. (2007), ‘Limits of static analysis for malware
Computer Security Applications Conference, 2007. ACSAC 2007. Twenty-Third
No., pp. 21,430, 10-14.

detection’,
Annual,
Vol.,

Authors
SaeedAlmarri received his B.S.c in Information system from Ajman University of Science & Technology,
College of Information Technology in 2008. He got his Master degree in Business information system from
University of Bedfordshire, UK, in 2011. Currently, he is a PhD student at University of Bedfordshire with
a thesis entitled " Malware Detection and Analysis".
Dr. Paul Sant completed his PhD from King's College, London in 2003 with a thesis entitled
"Algorithmics of edge-colouring pairs of 3-regular trees" and prior to this, a BSc. in Computer Science
from the University of Liverpool (1999). Paul is an active member of the British Computer Society and a
Chartered Information Technology Professional (CITP) as well as being a fellow of the Higher Education
Academy. In January 2013 Paul was appointed to a seconded position of Associate Dean working on a
University Campus project. He still maintains strong links with the Department, and is still research
attractive, being a local PI on the EU funded ECENTRE project.

15

Optimised malware detection in digital forensics

More Related Content

What's hot (17)

Similar to Optimised malware detection in digital forensics (20)

Recently uploaded (20)

Optimised malware detection in digital forensics