Analysis And Research Of Side Channel Based On FPGA Block Cipher Algorithm

Side channel analysis which is the current hot research direction of cryptography,mainly uses the leakage information generated during the operation of the cryptographic chip to crack the key information in the encryption device by establishing the appropriate leakage model.With the rapid development of channel analysis,there is a great challenge to information security today.Therefore,it is of great significance to study side channel analysis methods to protect cryptographic chips from being hacked.In this paper,based on the FPGA cryptographic chip platform,AES and DES were used to the traditional side channel analysis method,and some improved new methods are put forward to further optimize the effect of cryptographic attack.The main work and innovation of this paper are as follows:(1)In this paper,a FPGA development board is used for hardware and software design,and an automatic electromagnetic acquisition platform is built.The platform has more functions,high collection efficiency and remarkable flexibility.In addition,this paper uses digital IC software to build a simulation collection platform for side channel power consumption.This platform does not need a lot of hardware equipment,and can complete the collection of side channel data only by PC.(2)In this paper,two unsupervised side-channel analysis methods,CPA and LRA,are used for side-channel analysis.The hamming distance model is established for AES and DES encryption algorithms,and all the keys are recovered by piece-wise attack method.This paper proposes an experimental method of joint attack method based on CPA and LRA.The experimental results show that the effect of joint attack is better than that of CPA and LRA alone.(3)In this paper,two supervised side-channel attack methods,TA and SVM multi-classification attack,are used for side-channel analysis.Aiming at AES and DES encryption algorithms,the hamming distance bit model is established to attack the sub-key of the last bit of the block key.In this paper,a new method of feature extraction in time-frequency domain is proposed.The extracted features are formed into feature groups by dynamic sliding window,and the PCA data dimension reduction method is used to reduce the dimension and then SVM multi-classification attacks are carried out.The experimental results show that the effect of SVM multi-classification attack after feature extraction is better than that of TA and SVM multi-classification attack.