SCA with magnitude squared coherence

Abstract

Magnitude Squared Coherence (MSC) is a signal processing tool that indicates how well two time domain signals match one with the other by tracking linear dependencies in their spectral decomposition. Spectral Coherence ANalysis (SCAN) was the first way to use it as a Side-Channel Attack (SCA). This paper introduces two ways of using the Magnitude Squared Coherence in side-channel analyses. The first way is to use it as a distinguisher while the second consists in using it to transform the side-channel traces in a worthwhile manner. Additionally, an algorithm for fast computation of the SCAN is provided. © 2013 Springer-Verlag.