Using the joint distributions of a cryptographic function in side channel analysis

Abstract

The Side Channel Analysis is now a classic way to retrieve a secret key in the smart-card world. Unfortunately, most of the ensuing attacks require the plaintext or the ciphertext used by the embedded algorithm. In this article, we present a new method for exploiting the leakage of a device without this constraint. Our attack is based on a study of the leakage distribution of internal data of a cryptographic function and can be performed not only at the beginning or the end of the algorithm, but also at every instant that involves the secret key. This paper focuses on the distribution study and the resulting attack. We also propose a way to proceed in a noisy context using smart distances. We validate our proposition by practical results on an AES128 software implemented on a ATMega2561 and on the DPAContest v4 [32]. © 2014 Springer International Publishing Switzerland.