Optimal Linear Codes for Inner Product Masking


Citation: Wei Cheng, Olivier Rioul, Jean-Luc Danger and Sylvain Guilley. Optimal Linear Codes for Inner Product Masking. Journée de l’ED IP Paris 2019, Paris, France, Dec. 10, 2019. [Poster]

Basically, my Ph.D thesis consists of two subjects. The first one is about the information-theoretic measure of information that Y brings on X, where Y is the (general) leakage of X. For instance, Y is the side-channel leakages of (computing, storing, etc) secret(s). In this poster, we presented the application of Rényi entropy for measuring Hamming weight leakages Y of the sensitive variable X (as in the schematic figure).

The second one is the construction and analysis of code-based masking scheme against side-channel analysis and fault injection analysis. These code-based masking includes IPM (Inner Product Masking), DSM (Direct Sum Masking), etc. In this paoster, we quantified the side-channel resistance of IPM with different linear codes by using SNR (Signal-to-Noise Ratio) and MI (Mutual Information). In addition, we propose a straightforward way to choose the optimal linear codes for IPM in the sense of maximizing its side-channel resistance.