Power Analysis Attack Based on FCM Clustering Algorithm

Gao Shen, Qiming Zhang, Yongkang Tang, Shaoqing Li, Ruonan Zhang


With the development of information technology over the world for decades, information security is called an important issue in today's society. As a hardware carrier for information security, cryptographic chips have become an important issue in academia. In this paper, we propose a power analysis attack based on FCM(Fuzzy C-Means) clustering algorithm. Our method clusters the energy traces according to their intrinsic similarity, and classifies the energy traces according to the Hamming distance energy model. The correct key is found by comparing the similarity between the clustering result and the classification result. In order to eliminate noises, ICA (Independent Component Analysis) is involved. The simulation experiment uses AES cipher algorithm as the attack object. HSPICE simulation is performed on the first round of 8-bit s-box circuit under 40nm process. Experimental results show that our method is effective.


Full Text:



  • There are currently no refbacks.