Analyzing side channel leakage of masked implementations with stochastic methods

Abstract

Side channel cryptanalysis is a collective term for implementation attacks aiming at recovering secret or private keys from a cryptographic module by observing its physical leakage at run-time. Stochastic methods have already been introduced for first order differential side channel analysis. This contribution provides a compendium for the use of stochastic methods on masked implementations, i.e., on implementations that use internal random numbers in order to effectively prevent first order side channel attacks. Practical evidence is given that stochastic methods are also well suited for analyzing masked implementations, especially, as they are capable of combining several chosen components of different internal states for a multivariate side channel analysis. © Springer-Verlag Berlin Heidelberg 2007.