Paper 2019/818

X-DeepSCA: Cross-Device Deep Learning Side Channel Attack

Debayan Das and Anupam Golder and Josef Danial and Santosh Ghosh and Arijit Raychowdhury and Shreyas Sen

Abstract

This article, for the first time, demonstrates Cross-device Deep Learning Side-Channel Attack (X-DeepSCA), achieving an accuracy of $>99.9\%$, even in presence of significantly higher inter-device variations compared to the inter-key variations. Augmenting traces captured from multiple devices for training and with proper choice of hyper-parameters, the proposed 256-class Deep Neural Network (DNN) learns accurately from the power side-channel leakage of an AES-128 target encryption engine, and an N-trace ($N\leq10$) X-DeepSCA attack breaks different target devices within seconds compared to a few minutes for a correlational power analysis (CPA) attack, thereby increasing the threat surface for embedded devices significantly. Even for low SNR scenarios, the proposed X-DeepSCA attack achieves $\sim10\times$ lower minimum traces to disclosure (MTD) compared to a traditional CPA.