Title:  Network based Fingerprinting Techniques for Industrial Control Systems

Committee: 

Dr. Beyah, Advisor        

Dr. Owen, Chair

Dr. Copeland

Abstract:

The objective of the proposed research is to develop novel fingerprinting-based intrusion detection techniques that leverage the unique timing aspects of industrial control systems, both physical and in software. Industrial control systems present a unique set of security challenges due to legacy equipment and real-time system requirements preventing regular patching and the capability for cryptographic protocols. Therefore, minimally invasive intrusion detection and fingerprinting techniques must be used to secure these systems where compromises can lead to catastrophic physical damage. First, network traffic from live power distribution substations will be extensively studied and characterized, which will guide the development of the proposed fingerprinting techniques. Then, physical fingerprinting techniques proposed in previous work will be extended to a more general application by removing several limiting assumptions about the network architecture and capabilities of the end devices. Finally, timing based fingerprinting methods will be developed to detect when user applications on programmable logic controllers have been modified through both accidental and malicious means.