Heterogeneous IC Component Identification via EM Side-channels

Globalization of supply chain encompasses a mix of both trusted and potentially untrusted entities, raising significant security concerns. In this paper, we introduce an electromagnetic signal-based authentication methodology for integrated circuits within heterogeneous integrated circuits, addressing the heightened security vulnerabilities emerging from the globalization of the supply chain. Our proposed method involves exciting specific components to capture their unique EM signatures, thereby generating a comprehensive dictionary of component signatures. These signatures facilitate the clustering of signals and the detection of anomalies by enabling comparisons between test signals and a reference model derived from benign samples. We achieved perfect clustering performance for the components considered in HICs. Furthermore, utilizing the same setup, we developed an anomaly detection algorithm that achieves over 95% accuracy when the loop size of the injected code exceeds 100. While our focus is on HICs, the versatility of our method allows for straightforward application to a wide range of devices, including microelectronics and Internet of Things (IoT) devices, which are equally vulnerable to the complexities and risks associated with a diversified supply chain.