Cross-Correlation Based Trace Segmentation for Clustering Power Analysis on Public Key Cryptosystems

Simple Power Analysis (SPA) is a technique that directly analyzes the power consumption information collected during the execution of cryptographic algorithms. It is primarily based on the fact that different key values in public key cryptosystems (PKC) correspond to distinct operations, reflected in the power traces, allowing for key recovery. Effective segmentation of the power trace significantly enhances the efficiency of SPA, reducing the difficulty of key retrieval. This paper introduces a semi-automated Cross-Correlation Based Trace Segmentation method. We experimentally validated the segmentation method in scenarios involving smart cards, USB keys, and microcontrollers simulating unmanned aerial vehicle cryptographic modules. The results demonstrate the method’s high effectiveness in segmenting power traces of PKC.