Towards Side Channel-based Analysis for Embedded Systems Through Synthetic Control Flow Fingerprinting
Dr. Constantinos (Costas) Kolias
Associate Professor
Department of Computer Science
University of Idaho
Abstract: Embedded systems, such as Programmable Logic Controllers (PLCs), are critical components of essential infrastructure. However, protecting these systems presents significant challenges due to their limited resources. Traditional security measures, like antivirus software and host-based intrusion detection tools, are often unsuitable for such platforms. The need for innovative protection methods in this field has been highlighted by high-profile incidents such as Stuxnet, Industroyer, and BlackEnergy. Recently, the scientific community has explored the analysis of analog signals, such as the electromagnetic (EM) emissions generated by devices during normal operations. This approach, known as side-channel analysis (SCA), has traditionally been used offensively—for instance, to extract cryptographic keys. However, SCA techniques can also serve defensive purposes. By examining EM signals or power consumption patterns, it becomes possible to identify malicious changes in embedded system code remotely, without requiring additional software or hardware. Despite these advancements, a significant gap remains between laboratory research and practical applications. Many SCA studies are conducted in highly controlled environments, overlooking real-world challenges like the cost of collecting training data and the complexity of capturing all potential execution states (i.e., alternate code paths). Recent developments in generative models optimized for producing high-fidelity analog signals from machine code embeddings offer a potential solution. These synthetic signals could enable offline fingerprinting of embedded system execution, addressing practical limitations and enhancing system protection strategies.
BIO: Constantinos Kolias joined the Computer Science Department at the University of Idaho in Fall 2018. Prior to this, he served as a Research Assistant Professor in the Computer Science Department at George Mason University under the guidance of Prof. Angelos Stavrou. Dr. Kolias’s research focuses on security and privacy embedded systems, and critical infrastructures. He is particularly involved in designing intelligent Intrusion Detection Systems (IDS) using side-channel analysis. His work promotes a groundbreaking code-to-analog signal synthesis framework powered by generative AI techniques, enabling large-scale offline testing of software implementations for side-channel vulnerabilities. Additionally, Dr. Kolias is active in wireless network security. In 2015, he developed and released the AWID dataset, the first wireless dataset specifically designed for wireless security research, which has since been widely used as a benchmark by academic institutions and organizations worldwide. He is also exploring the application of Domain Adaptation principles to create synthetic training datasets tailored to specific applications and networks, aiming to advance the field of security research further.
Tuesday, December 3, 2024, 12:15-1:15 pm, ZOOM Only
Join Link: https://clarkson.zoom.us/j/97763004044?pwd=fReadMi2o7OYVOIOgYm5yAuGGnbmdy.1
*Co-Sponsored by IEEE student branch and HKN
________________________________________________________________
Electrical and Computer Engineering ● CLARKSON UNIVERSITY ● Potsdam, New York 13699-5720