Machine Learning Based Power Analysis for Simon with Secure Mechanism

The IoT requires the implementation of cryptographic techniques to ensure the confidentiality, integrity, and availability of communicated data. Therefore, lightweight block cipher with small circuit size, low power consumption, and low latency are attracting attention. Simon is a typical lightweight block cipher with small circuit size and computational security. On the other hand, side-channel analysis, which steals secret keys by using side-channel information during hardware implementation, has been reported as a threat. Also, various countermeasures have been proposed. Recently, deep learning side-channel analysis has been proposed as a more advanced analysis using deep learning. Therefore, it is very important to evaluate the tamper resistance of new analysis methods as well as conventional method to ensure the safety of the IoT. This study proposes a deep learning side-channel attack on Simon with noise countermeasure. The proposed method analyzes the round key computed in the final round function under the known ciphertext attack for Simon with loop implementation. Experiments on a real device showed that most partial keys can be analyzed using 5,000 waveforms, and the effectiveness of the proposed method was evaluated.