Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Digital Investigation 9 (2013) 167–169

Contents lists available at SciVerse ScienceDirect

Digital Investigation
journal homepage:

Editorial

Experimental design challenges in digital forensics

Applied research is a major theme in this issue of the measurements from within the environment being stud-
Journal. The included papers cover a range of evidential ied. Although we strive for certainty, as a practical matter
areas: Windows Mobile 7, Frostwire, Digsby, JumpLists, we have to accept a high degree of probability as a sub-
IconCache, and Stegdetect. In addition to providing a stitute for unequivocal proof, and we must work to miti-
deeper understanding of specific digital artifacts, these gate the factors that can cloud our view of the underlying
contributions exhibit various approaches to conducting truth.
experiments and interpreting their results. To increase the chances of success, when formulating an
Designing good experiments is hardly a trivial under- experiment to determine the cause or meaning of digital
taking, and has been the focus of brilliant minds at least artifacts, thoughtful planning is required to eliminate
since the scientific revolution. Along with the intricacies of irrelevant phenomena, and to evaluate individual causes
scientific experimentation in general, experiments in digi- separately. Planning includes the arrangement of the
tal forensics pose novel challenges, some of which are experimental environment to eliminate unwanted in-
discussed further below. The applied research papers in fluences, figuring out how to control and assess each vari-
this issue attempt to tackle many of these challenges in able separately, and developing scientific protocols to
order to avoid mistakes and increase scientific rigor in govern experimentation.
digital forensics. Read these papers with this in mind: give
experimental design in digital forensics the attention it 2. Complex influences
deserves.
The creation of an experiment and the design of appa-
ratus to measure results is a form of intervention into what
1. Scientific experiments is being studied. For instance, the presence of security
measures such as a firewall can inhibit certain aspects of
Applied research is a fundamentally important aspect of the process being studied. The use of a system monitoring
digital forensics, providing reliable knowledge of computer tool to collect data during experimentation can introduce
processes to support conclusions concerning what activ- errors into an experiment. Specific configuration settings of
ities occurred on a device or network. In digital forensics, a computer program or operating system can influence the
flawed test results can lead to incorrect decisions, poten- findings. The version of the operating system can
tially resulting in the loss of a person’s livelihood, liberty or completely change the outcome of an experiment. In one
life. To reduce the risk of incorrect conclusions, attention case, digital investigators concluded that file deletion and
must be given not just to the results of applied research, but disk wiping had occurred on the basis of experiments they
also to the underlying experimental design to ensure that performed using Windows NT, when in fact the arrange-
results are reliable and repeatable. ment of data on the evidential disk was caused by normal
The scientific method is heavily dependent on the usage of Windows 98. To reduce the risk of mistakes in
ability to gather accurate observations and to test working applied research, care must be taken to create an experi-
hypotheses. In digital forensics, accurate observation and mental environment that does not bias observations.
testing can be undermined by the inextricable linkage In order to obtain the clearest view of cause and effect
between experimentation and technology. Even when when conducting experiments, it is also desirable to isolate
conducting studies of the physical world, there can be each significant variable and test it individually, while
intrinsic interactions between the experiment and what is holding the other variables fixed. In some circumstances, it
being studied. Undoubtedly, the most famous example of may even be necessary to create the equivalent of a control
this is the Heisenberg uncertainty principle of quantum group for a given experiment by conducting tests that help
mechanics. Conducting experiments in digital forensics distinguish normal usage of the computer system from the
often has the unavoidable complexity of having to make particular process being studied. In digital forensics,

1742-2876/$ – see front matter ª 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.diin.2013.02.002
168 Editorial / Digital Investigation 9 (2013) 167–169

important variables can include operating system version mind that systematic error or bias in an experiment can be
and configuration, file system version, program version and difficult to detect, even between multiple runs of the same
configuration, client–server interactions, and specific op- experiment. Ideally, different people will take different
erations performed during an experiment. Although it can experimental approaches to test a theory – the more ways a
be troublesome to set up separate experiments to test one hypothesis is tested, the higher will be our confidence in its
variable at a time, in digital forensics, the trouble is usually veracity.
worthwhile when weighed against the dangers of report- Inevitably, there will be differences between multiple
ing incorrect results. runs of an experiment. Some of these differences may be
To complicate matters, experiments in digital forensics important, others irrelevant. One run of an experiment may
can encounter unforeseen situations that make it more encounter unique errors and data corruption that only
difficult to maintain control of the process. A program may become apparent when the experiment is repeated.
crash or certain triggering events may only occur under Random errors are always to be expected when conducting
specific circumstances, leading to variations in functionality experiments and can generally be excluded as a causal
under different experimental conditions. In some situations, factor by running an experiment multiple times and per-
a process will only exhibit certain behavior after a given forming statistical analysis on the results. Running the
period of time or when a certain trigger occurs. As a result, same experiment at different times may result in differ-
experiments conducted over a period of days may exhibit ences in date–time stamps that are materially irrelevant to
different results from those conducted over several months. the overall results. However, running multiple experiments
In order to control such influences as much as feasible, over a long period of time may lead to new findings, such as
when designing an experiment, careful thought must be automatic resetting of counters, trimming of file contents,
given to which interactions are necessary, which influences or automatic deletion of data. Therefore, it is also necessary
are irrelevant, and which factors could create a disturbance to define what falls within the parameters of reproduc-
that might distort the results. ibility and which differences are to be expected.

3. Reverse engineering 5. Reporting

When designing an experiment, a methodical approach The ability to apply and independently verify research
to uncovering conditions and triggers that influence the findings depends heavily on adequate reporting of the
operation of a process is preferable. Therefore, in some experimental setup and results. Not knowing the version of
situations, applied researchers may wish to become application or operating system that was used can make it
familiar with the programmatic logic of the code, since it is difficult to assess whether the findings are associated with
this that fundamentally drives the process. the current version, an older one, or both. Simply providing
In digital forensics, when a particular computer pro- a list of digital artifacts and summary of their meaning can
gram or operating system feature is being studied, it may lead to confusion, particularly when some information may
be possible to gain insight into the inner workings. In- be wrong. To ensure that results are verifiable by others,
depth analysis of the program being studied can reveal applied research reports should detail how experiments
data structures and encryption keys that help unlock the were set up and conducted. Although it may not be
content and meaning of experimental results. Under such necessary to publish the complete written protocols used to
circumstances, malware forensics can provide a framework conduct experiments, it is a good practice to keep them on
for applied research in digital forensics. Malware analysts file for future reference.
commonly reverse engineer computer programs to deter- There is also a need for completeness when reporting
mine their functionality and to uncover special conditions, outcomes of experimentation. In addition to describing the
encryption and other nuances in the code. successes, report the failures, because these can reveal
However, such in-depth dissection of a program or crucial characteristics of the process or weaknesses in the
operating system feature is not always feasible. The time or experiment. When strange errors or peculiar outcomes
skills that are readily available may not be sufficient to occur during an experiment that has not been seen before,
support such an analysis. In some cases, the full application such events may turn out to be significant and should be
may be inaccessible, such as client server or obfuscated/ documented accordingly. These events may be a clue that
protected code. further research and testing is required to determine their
cause and meaning.
4. Reproducibility
6. Interpretation of findings
Digital investigators survive on their reputations, and
must make every effort to verify experimental findings for A degree of scientific skill is also required when inter-
themselves. An error in a forensic report or expert testi- preting the results of an experiment. An interpretation
mony will be attributed to the digital investigator, regard- cannot be considered sound until an honest attempt has
less of whether the error was actually due to a mistake in been made to eliminate other plausible explanations. This
someone else’s work. Therefore, to be useful in digital fo- process is called falsification and is a crucial component of
rensics, experiments need to be reproducible. One must the scientific method.
run the same experiment multiple times to verify results, Given all of the challenges associated with experi-
and to exclude errors and extraneous variations. Keep in mental design, it is necessary to think critically about
Editorial / Digital Investigation 9 (2013) 167–169 169

what can and cannot be concluded. Stating the limita- reproduced, the results can be applied and validated within
tions of one’s experimental setup provides digital in- the defined context. However, a growing number of cases
vestigators with a realistic expectation of the results and have had flawed conclusions presented in court based on
their application. In addition, clearly stating these limi- incorrect or misinterpreted experimental results. There-
tations can motivate other researchers to conduct further fore, it is necessary for those conducting applied research in
experiments to deepen our understanding of digital digital forensics to ensure that their experiments are reli-
evidence. able and reproducible, and to clearly communicate the
limitations of their findings.
7. Conclusions
Eoghan Casey
Johns Hopkins University Information Security Institute,
The usefulness of applied research is greatly increased
216 Maryland Hall, Baltimore, MD 21218, United States
by solid experimental design and documentation. Provided
E-mail address: eoghan@jhu.edu
the experimental setup is defined clearly enough to be

You might also like