EMSCALe: Component Decomposition by Leveraging Electromagnetic Sidechannels

We propose “EMSCALe: Component Decomposition and Attestation by Leveraging Electromagnetic Side-channels”, a novel approach to accurately identify functions of integrated circuits (ICs) and detect deviation from ordinary activities by monitoring the unintended electromagnetic (EM) signal emanated from heterogeneous ICs (HICs). EMSCALe will collect EM signals using purpose-designed antennas, then pre-process and reveal the information embedded into the EM signals about different aspects of the components. EMSCALe will use models of software behavior and software-system-hardware interactions to form hypotheses about the signal models of the individual ICs. During the test stage, the framework will then update these hypotheses by matching the expected signals for each hypothesis to the observed signals. Having multiple hypotheses helps to cope with the presence of interrupts, input/output activity, cache misses, branch miss-predictions, other events, etc., and will allow EMSCALe to maintain high accuracy and fidelity while identifying components of an HIC of interest. These events significantly change the emanated EM signals in a way that is seemingly random. However, EMSCALe can account for these deviations and even use them to improve component identification. The basic concepts and initial feasibility of EMSCALe for IoT-class devices was demonstrated by the founders of Aether Argus Inc. in the DARPA LADS program. We will build upon that work to demonstrate the ability to identify the individual IC components and detect anomalies (hardware Trojan injected or counterfeit) in real-time on devices of interest. Aether Argus Inc. is a technology startup created by faculty members at Georgia Tech and Virginia Tech, bringing together technical expertise, know-how, and intellectual property created through participation in the DARPA LADS program in their academic capacity. This prior work represents a “best of breed” on EME-based analysis for defensive purposes, consistently demonstrating the dominant results across the program. We will leverage our experience and initial results to “hit the ground running” in conducting the proposed research.