Rank Estimation Methods in Side Channel Attacks
Christine van Vredendaal
Abstract:
Suppose we have an implementation that uses a key $k*$. Given side-channel attack results on this implementation, the rank of $k*$ is defined as the number of keys $k$ that have a higher probability of being the one used in the implementation according to the side-channel results. A higher rank therefore means less information leakage and a more secure key.
Side-channel attacks are a vital aspect of a security evaluation framework. Interpreting the results to give a security level is however not always so easy. Rank estimation algorithms give a method to compute an interval for the rank of a key and this can be used to estimate the security level of an implementation.
In this talk I will give the theoretic background to rank estimation and delve into algorithms to compute the rank in symmetric key cryptography such as AES. If time allows I will also speak about what you can do in public key crypto, more specifically ECC.