Machine-Learning-Based Side-Channel Leakage Detection in Electronic System-Level Synthesis
Lu Zhang, Dejun Mu, Wei Hu, Yu Tai · IEEE Network · 2020
Security critical hardware devices in CPS must experience security testing including side-channel leakage detection, which aims to provide earlier assessment before coming into service. Due to the exponential nature of ESLS, performing fast and efficient leakage detection is a critical task in the context of security evaluation for DSE. However, the traditional method of leakage detection is always time-consuming and costly. To overcome this limitation, this article investigates the strategies of leveraging the emerging ML techniques to achieve faster and more effective leakage detection. A clustering-based methodology is proposed as an alternative method for first-order leakage detection. We further confirm the effectiveness and rapidity of the proposed method by launching clustering-based analysis on seven different benchmarks. The proposed method provides almost the same quality of results as the non-specific t-test while substantially reducing the efforts on security evaluation in CPS.