New side-channel and techniques for hardware trojan detection

Abstract

The thesis introduces a new physical side-channel, which we call the backscattering side-channel, and propose novel hardware Trojan (HT) and counterfeit integrated circuit (IC) detection techniques that exploit the backscattering side-channel. These techniques are capable of detecting different types of inactive HTs and counterfeit ICs on multiple circuit benchmarks while tolerating manufacturing variation. For the last decade, demand for effective HT and counterfeit IC detection techniques has risen considerably. Numerous HT and counterfeit IC detection techniques have been published and side-channel analysis based approaches are among the most widely used. However, the problem with existing side-channels is that they do not provide enough resolution bandwidth, and information about the operation of electronic circuitry to detect small dormant hardware Trojan and small-changed counterfeit ICs. In addition, most previously proposed techniques do not take into account manufacturing variation, test on very few benchmarks, or rely on an unrealistic assumption of having a golden (HT-free or trusted-IC) sample. Motivated by these problems, our research focuses on introducing a new side-channel, i.e., the backscattering side-channel, and proposing novel techniques for HT and counterfeit IC detection using the new side-channel. We observe that the backscattering side-channel is especially suitable for HT detection because it has high bandwidth and spatial resolution, and its signal carries information about the current state of on-chip impedance. To summarize, this work has 1) introduced a new backscattering side-channel, theoretically and experimentally proved the concept and existence of the side-channel, 2) developed new techniques for detection of dormant hardware Trojans and counterfeit ICs using the new backscattering side-channel, 3) modeled and compared the backscattering, electromagnetic (EM), and power side-channels and their performance in detecting malware and hardware Trojans , 4) developed novel clustering based techniques that can assist reverse engineering based methods for HT detection in a large population of integrated circuits, and 5) developed novel golden-chip-free HT detection techniques using backscattering side-channel using circuit impedance models.