CAO Jiahua,WU Zhen,WANG Yi,et al.Side Channel Attack of S-box Power Randomization based on CNN-BPR[J].Journal of Chengdu University of Information Technology,2022,37(01):16-20.[doi:10.16836/j.cnki.jcuit.2022.01.003]
基于CNN-BPR的S-Box功耗随机化侧信道攻击
- Title:
- Side Channel Attack of S-box Power Randomization based on CNN-BPR
- 文章编号:
- 2096-1618(2022)01-0016-05
- Keywords:
- side channel attack; template attack; convolutional neural network; loss function; Bayesian personalized ranking
- 分类号:
- TP309
- 文献标志码:
- A
- 摘要:
- S-Box功耗随机化是一种对抗侧信道攻击的防御方案,该方案将设备加密过程中S-Box输出值功耗泄露的位置进行随机化处理,降低了中间值与能量消耗的相关性,使得基于固定位置进行能量分析的代价大幅增加。具备平移不变性的卷积神经网络在侧信道攻击上取得了显著的效果。为进一步提高其对S-Box功耗随机化防御方案的攻击能力,基于贝叶斯个性化排序的思想,提出一种更符合侧信道攻击原理的CNN-BPR模型。实验结果表明,与Softmax交叉熵损失模型相比,CNN-BPR模型在使用全部训练能迹用于模板攻击时,成功恢复密钥所需要的攻击能迹数量能够减少3%,当使用60%的训练能迹用于模板攻击时,减少的攻击能迹数量能够达到27%。
- Abstract:
- S-box power randomization is a defense scheme against side channel attack. This scheme randomizes the location of power leakage of S-box output value in the process of device encryption, and reduces the correlation between intermediate value and energy consumption, and greatly increases the cost of energy analysis based on fixed location. Convolutional neural network with translation invariance has achieved remarkable results in side channel attack. In order to further improve its attack ability against S-box power randomization defense scheme, a CNN-BPR model more in line with the principle of side channel attack is proposed based on the idea of Bayesian personalized ranking. The experimental results show that, compared with softmax cross-entropy loss model, when CNN-BPR model uses all training energy traces for template attack, the number of attack traces required to successfully recover the key can be reduced by 3%, and when 60% of the training energy traces are used for template attack, the number of attack energy traces can be reduced by 27%.
参考文献/References:
[1] Kocher P C.Timing attacks on implementat-ions of Diffie-Hellman,RSA,DSS,and other systems[C].Annual International Cryptology Conference,1996:104-113.
[2] Mangard S.A Simple Power-Analysis(SPA)Attack on Implementations of the AES KeyExpansion[C].Information Security and Cryptology-ICISC 2002,2002:28-29.
[3] Benhadjyoussef N,Machhout M,Tourki R.Optimized power trace numbers in CPA attacks[C].2011 8th International Multi-Confere-nce on Systems,Signals and Devices(SSD),2011:1-5.
[4] 杜之波,吴震,王敏.针对基于SM3的HMAC的能量分析攻击方法[J].通信学报,2016,37(5):38-43.
[5] Kocher P,Jaffe J,Jun B.Differential poweranalysis[C].Advances in Cryptology,1999:388-397.
[6] Standaert F-X,Archambeau C.Using subspace-based template attacks to compare andcombine power and electromagnetic information leakages[C].Cryptographic Hardware and Embedded Systems-CHES 2008,2008:411-425.
[7] Veyrat-Charvillon N,Medwed M,Kerckhof S.Shuffling against Side-Channel Attacks:AComprehensive Study with Cautionary Note[C].Advances in Cryptology-ASIACRYPT 2012,2012:740-757.
[8] Okeya K,Sakurai K.A Second-Order DPAAttack Breaks a Window-Method Based Co-untermeasure against Side Channel Attacks[C].Information Security,2002:389-401.
[9] 杨欢,吴震,王燚.侧信道多层感知器攻击中基于贝叶斯优化的超参数寻优[J].计算机应用与软件,2021,38(5):323-330.
[10] Cagli E,Dumas C,Prouff E.ConvolutionalNeural Networks with Data AugmentationAgainst Jitter-Based Countermeasures[C].Cryptographic Hardware and Embedded Systems-CHES 2017,2017:45-68.
[11] Benadjila R,Prouff E,Strullu R.Deep learning for side-channel analysis and introduction to ASCAD database[J].Journal of Cryptographic Engineering,2020:163-188.
[12] Zaid G,Bossuet L,Dassance F.RankingLoss: Maximizing the Success Rate in DeepLearning Side-Channel Analysis[C].Cryptographic Hardware and Embedded Systems-CHES 2021,2021:25-55.
[13] Wang Y,Wang Q,Guo X,et al.Optimization and Performance Analysis of Extreme Learning Machine by L2-Norm Regularization[M].Cham:Springer,2021:405-413.
[14] Dave VS,Zhang B,Chen PY.Neural Brane:Neural Bayesian Personalized Ranking forAtributed Network Embedding[J].Data Scienceand Engineering,2019,4(2):119-131.
[15] François-Xavier Standaert,Malkin T,YungM. A Unified Framework for the Analysisof Side-Channel Key Recovery Attacks(extended version)[C].Advances in Cryptology-EUROCRYPT 2009,2009:443-461.
相似文献/References:
[1]万武南,陈 俊.针对Montgomery模幂算法的选择明文SPA攻击[J].成都信息工程大学学报,2016,(04):348.
WAN Wu-nan,CHEN Jun.A Simple Power Analysis Attack on the Montgomery
Modular Exponentiation Algorithms[J].Journal of Chengdu University of Information Technology,2016,(01):348.
[2]万武南,陈 俊.针对双重掩码模幂算法的二阶互相关功耗分析攻击[J].成都信息工程大学学报,2016,(04):353.
WAN Wu-nan,CHEN Jun.A Second Order Cross Correlation Power Analysis Attack on
Double Blinding Exponentiation Algorithms[J].Journal of Chengdu University of Information Technology,2016,(01):353.
[3]匡晓云,黄开天,兰 天,等.针对SM4密码算法的模板攻击[J].成都信息工程大学学报,2021,36(05):499.[doi:10.16836/j.cnki.jcuit.2021.05.004]
KUANG Xiaoyun,HUANG Kaitian,LAN Tian,et al.Template Attack Against SM4 Cryptographic Algorithm[J].Journal of Chengdu University of Information Technology,2021,36(01):499.[doi:10.16836/j.cnki.jcuit.2021.05.004]
备注/Memo
收稿日期:2021-09-27
基金项目:“十三五”国家密码发展基金资助项目(MMJJ20180224); 四川省重点研发资助项目(2019YFG0096)