CS 665 - Network Forensics

Tuesdays - 6:15pm-8:45pm - BC640

DRAFT SYLLABUS -AUG 27, 2009

Network forensics involves the identification, preservation, and analysis of evidence of attacks in order to identify the attackers and document their activity with sufficient reliability to justify appropriate technological, business, and legal responses. This course, however, only focuses on the technological and not on the legal components of the topic. Much emphasis is on the network traffic analysis aspect, not on the host aspect. The technical aspect addresses analysis of intruder types and the intrusion process, review of network traffic logs (pcap, flow records) and profiles and their types, identification of attack signatures and fingerprints, application of data mining techniques, study of various traceback methods, and the extraction of information (e.g. from malware, including botnet traffic) acquired through the use of network analysis tools and techniques. The class will not only cover the subjects in theory but instead also provide the students with an extensive hands-on experience.  The class will involve a fair amount of programming. Those who take the class are expected to be able to program in C/C++, have some basic knowledge of linear algebra, graph theory and statistics, be able to work with Mathematica and Matlab, and be familiar with network basics and programming, as well as Unix-like operating systems.

Instructor:

There will be group projects for this class. In general, programming sections of a project should compile and run on the Unix lab (Burchard) or Cybersecurity lab (Babbio) machines. For projects dealing with Windows/MacOS or other OSes, you must get the permission of the instructor.

Project I: A 5 to 8-page paper summarizing your project findings plus a 15-minute presentation to the class followed by 10-15 minutes of discussion.

Project II: A 15-page paper summarizing your project findings, plus any programming appendices, and a 30-minute presentation followed by 15-30 minutes of discussion.