Logo IMG
HOME > PAST ISSUE > Article Detail


Digital Forensics

Modern crime often leaves an electronic trail. Finding and preserving that evidence requires careful methods as well as technical skill

Simson L. Garfinkel

2013-09GarfinkelF1.jpgClick to Enlarge ImageSince the 1980s, computers have had increasing roles in all aspects of human life—including an involvement in criminal acts. This development has led to the rise of digital forensics, the uncovering and examination of evidence located on all things electronic with digital storage, including computers, cell phones, and networks. Digital forensics researchers and practitioners stand at the forefront of some of the most challenging problems in computer science, including “big data” analysis, natural language processing, data visualizations, and cybersecurity.

Compared with traditional forensic science, digital forensics poses significant challenges. Information on a computer system can be changed without a trace, the scale of data that must be analyzed is vast, and the variety of data types is enormous. Just as a traditional forensic investigator must be prepared to analyze any kind of smear or fragment, no matter the source, a digital investigator must be able to make sense of any data that might be found on any device anywhere on the planet—a very difficult proposition.

From its inception, digital forensics has served two different purposes, each with its own difficulties. First, in many cases computers contain evidence of a crime that took place in the physical world. The computer was all but incidental—except that computerization has made the evidence harder for investigators to analyze than paper records. For example, financial scam artist Bernard Madoff kept track of his victims’ accounts by using an IBM AS/400 minicomputer from the 1980s. The age of the computer helped perpetuate his crime, because few people on Wall Street have experience with 25-year-old technology, and it created an added complication after Madoff was arrested, because investigators had few tools with which to make sense of his data.


Today personal computers are so ubiquitous that the collection and use of digital evidence has become a common part of many criminal and civil investigations. Suspects in murder cases routinely have their laptops and cell phones examined for corroborating evidence. Corporate litigation is also dominated by electronic discovery of incriminating material.

The second class of digital forensics cases are those in which the crime was inherently one involving computer systems, such as hacking. In these instances, investigators are often hampered by the technical sophistication of the systems and the massive amount of evidence to analyze.

Digital forensics is powerful because computer systems are windows into the past. Many retain vast quantities of information—either intentionally, in the form of log files and archives, or inadvertently, as a result of software that does not cleanly erase memory and files. As a result, investigators can frequently recover old email messages, chat logs, Google search terms, and other kinds of data that were created weeks, months or even years before. Such contemporaneous records can reveal an individual’s state of mind or intent at the time the crime was committed.

But whereas pre-computer evidence, such as handwritten letters and photographs, could be reproduced and given to attorneys, judges, and juries, computerized evidence requires special handling and analysis. Electronic data are easily changed, damaged, or erased if handled improperly. Simply turning on a consumer GPS may cause the device to delete critical evidence. Additionally, computers frequently harbor hidden evidence that may be revealed only when specialized tools are used—for example, a digital camera may appear to have 30 photos, but expert examination may show another 300 deleted photos that can be recovered. (When a device “erases” a file, it doesn’t clear the memory space, but notes that the space is available; the file may not be really deleted until a new one is written over it.)

Because they can look into the past and uncover hidden data, digital forensics tools are increasingly employed beyond the courtroom. Security professionals routinely use such tools to analyze network intrusions—not to convict the attacker but to understand how the perpetrator gained access and to plug the hole. Data recovery firms rely on similar tools to resurrect files from drives that have been inadvertently reformatted or damaged. Forensic tools can also detect the unintentional disclosures of personal information. In 2009 the Inspector General of the U.S. Department of Defense issued a report stating that many hard drives were not properly wiped of data before leaving government service.

Digital evidence can even be examined to show that something did not happen. Here they are less powerful, for the well-known reason that the absence of evidence is not the evidence of absence. In May 2006 a laptop and external hard drive containing sensitive personal information of 26.5 million veterans and military personnel was stolen from an employee at the U.S. Department of Veterans Affairs. After the laptop was recovered in June 2006, forensic investigators analyzed the media and determined that the sensitive files probably had not been viewed.

One way to make such a judgment is by examining the access and modification times associated with each file on the hard drive. But someone taking advantage of the same forensic techniques could have viewed the laptop files without modifying those timestamps, so the investigators really determined only that the files had not been opened by conventional means.

These examples emphasize that the possibilites of digital forensics are bounded not by technology but by what is cost-effective for a particular case. Convictions are frequently the measure of success. In practice there is a considerable gap between what is theoretically possible and what is necessary; even though there may be an intellectual desire to analyze every last byte, there is rarely a reason to do so.

comments powered by Disqus


Subscribe to American Scientist