'Rank Correction': A New Side-Channel Approach for Secret Key Recovery

摘要

In this paper we present the "Rank Corrector" (RC), an empirical approach aiming at enhancing most Side Channel Attack (SCA). We show that during an SCA on a cryptographic algorithm like the Data Encryption Standard (DES), the rank of the secret key displays a specific behaviour with regards to other hypotheses. Hence the Rank Corrector algorithm is devised, in order to improve existing SCAs by exploiting such behaviours. With a profiling phase on a clone device, we precisely evaluate the set of parameters that ensure the adaptability of RC to a large range of cryptographic systems, and the possibility to discriminate the secret key from other hypotheses in an efficient manner. The main principle of RC is to detect and discard the false keys hypotheses when analysing the ranking evolution. This results in improving the rank of the secret key, thus accelerating the attack. The efficiency of our algorithm is assessed by performing a Differential Power Analysis (DPA) with and without the rank corrector. We observe a gain of at least 15% on the "Measurements To Disclosure" (MTD) criteria.