A hypothesis-based approach to digital forensic investigations

Brian D Carrier, Purdue University

Abstract

This work formally defines a digital forensic investigation and categories of analysis techniques. The definitions are based on an extended finite state machine (FSM) model that was designed to include support for removable devices and complex states and events. The model is used to define the concept of a computer's history, which contains the primitive and complex states and events that existed and occurred. The goal of a digital investigation into make valid inferences about a computer's history. Unlike the physical world, where an investigator can directly observe objects, the digital world involves many indirect observations. The investigator cannot directly observe the state of a hard disk sector or bytes in memory. He can only directly observe the state of output devices. Therefore, all statements about digital states and events are hypotheses that must be tested to some degree. Using the dynamic FSM model, seven categories and 31 unique classes of digital investigation analysis techniques are defined. The techniques in each category can be used to test and formulate different types of hypotheses and completeness is shown. The classes are defined based on the model design and current practice. Using the categories of analysis techniques and the history model, the process models that investigators use are formally compared. Until now, it was not clear how the phases in the models were different. The model is also used to identify where assumptions are made during an investigation and to show differences between the concepts of digital forensics and the more traditional forensic disciplines.

Degree

Ph.D.

Advisors

Spafford, Purdue University.

Subject Area

Computer science

COinS