基于模板和KNN算法的能量分析攻击对比研究

doi:10.11959/j.issn.1000-0801.2022080

电信科学 ›› 2022, Vol. 38 ›› Issue (4): 121-129.doi: 10.11959/j.issn.1000-0801.2022080

基于模板和KNN算法的能量分析攻击对比研究

靳济方, 刘承远, 范晓红, 段晓毅, 刘嘉瑜

北京电子科技学院电子与通信工程系，北京 100070

修回日期:2022-03-03 出版日期:2022-04-20 发布日期:2022-04-01
作者简介:靳济方（1972− ），女，北京电子科技学院电子与通信工程系副教授，主要研究方向为电路与系统安全
刘承远（1999− ）,男，北京电子科技学院硕士生，主要研究方向为信息安全
范晓红（1979− ），女，北京电子科技学院电子与通信工程系讲师，主要研究方向为信息安全
段晓毅（1979− ），男，北京电子科技学院电子与通信工程系副教授，主要研究方向为信息安全
刘嘉瑜（1999− ），女，北京电子科技学院在读，主要研究方向为信息安全
基金资助:
北京市高精尖学科建设基金资助项目(20210032Z0401);北京市高精尖学科建设基金资助项目(20210032Z0401)

Comparative study of power analysis attacks based on template and KNN algorithm

Jifang JIN, Chengyuan LIU, Xiaohong FAN, Xiaoyi DUAN, Jiayu LIU

Department of Electronics and Information Engineering, Beijing Electronic Science and Technology Institute, Beijing 100070, China

Revised:2022-03-03 Online:2022-04-20 Published:2022-04-01
Supported by:
The High-Tech Discipline Construction Funds of China(20210032Z0401);The High-Tech Discipline Construction Funds of China(20210032Z0401)

摘要/Abstract

摘要：

能量分析攻击至今仍是针对密码芯片最具威胁的攻击方法之一，针对传统的模板分析攻击和KNN算法的攻击进行对比研究，对比模板攻击和机器学习中的KNN优缺点。首先对皮尔逊相关系数、互信息和最大信息系数、距离相关系数 3 种降维方法进行了研究；然后对比了相同数量功耗曲线下，特征点数量对两种能量分析的成功率等性能的影响；同时研究了不同降维技术在相同功耗曲线数量和不同功耗曲线数量时对两种能量分析攻击的影响。结果表明，模板攻击在运行速度、占用内存方面优于KNN算法攻击，而在攻击成功率和鲁棒性方面，KNN算法攻击具有更好的表现。

关键词: 能量分析攻击, 模板攻击, KNN算法, 特征点

Abstract:

Power analysis attack is still the most threatening type of side channel attack on cryptographic hardware.The template analysis attack with the attack of KNN algorithm was compared.Firstly, three dimensionality reduction methods of Pearson correlation coefficient, mutual information and maximum information coefficient and distance correlation coefficient were studied.Then, the effects of the number of feature points on the attack success rate of the two power analysis attacks under the same number of power consumption curves were compared.At the same time, the effects of different dimensionality reduction techniques on the two power analysis attacks when the number of power curves is the same and different.The results show that the template attack is better than the KNN algorithm attack in running speed, memory occupation and robustness, and the KNN algorithm attack has better performance in attack success rate.

Key words: power analysis attack, template attack, KNN algorithm, feature point

中图分类号:

TM344.1

靳济方, 刘承远, 范晓红, 段晓毅, 刘嘉瑜. 基于模板和KNN算法的能量分析攻击对比研究[J]. 电信科学, 2022, 38(4): 121-129.

Jifang JIN, Chengyuan LIU, Xiaohong FAN, Xiaoyi DUAN, Jiayu LIU. Comparative study of power analysis attacks based on template and KNN algorithm[J]. Telecommunications Science, 2022, 38(4): 121-129.

图/表 12

图1

图2

图3

图4

图5

图6

图7

表1

图8

图9

图10

表2

参考文献 15

[1]	KOCHER P , JAFFE J AND JUN B . Differential power analysis[J]. Lecture Notes in Computer Science, 1999: 388-397.
[2]	AO J R , ROHATGI P . Template attacks[J]. International Workshop on Cryptographic Hardware and Embedded Systems, 2002: 13-28.
[3]	BRIER E , CLAVIER C , OLIVIER F . Correlation power analysis with a leakage model[C]// Cryptographic Hardware and Embedded Systems.[S.l.:s.n.], 2004.
[4]	CHRISTIAN R , ELISABETH O . Practical template attacks[C]// Proceedings of the 5th international conference on Information Security Applications (WISA'04).[S.l.:s.n.], 2004: 440-456.
[5]	ARCHAMBEAU C , PEETERS E , STANDAERT F X ,et al. Template attacks in principal subspaces[C]// Cryptographic Hardware and Embedded Systems.[S.l.:s.n.], 2006.
[6]	GIERLICHS B , LEMKE-RUST K , PAAR C ,et al. Stochastic methods[C]// Cryptographic hardware and embedded systems.[S.l.:s.n.], 2006.
[7]	STANDAERTF X , ARCHAMBEAU C . Using subspace-based template attacks to compare and combine power and electromagnetic information leakages[C]// Cryptographic Hardware and Embedded Systems.[S.l.:s.n.], 2008.
[8]	HOSPODAR G , GIERLICHS B , DE MULDER E ,et al. Machine learning in side-channel analysis:a first study[J]. Journal of Cryptographic Engineering, 2011,1(4): 293-302.
[9]	LERMAN L , MEDEIROS S F , VESHCHIKOV N ,et al. Semi-supervised template attack[J]. IACR Cryptology EPrint Archive,2012, 2012:82.
[10]	KIM J , PICEK S , HEUSER A ,et al. [J]. Make some noise.unleashing the power of convolutional neural networks for profiled side-channel analysis, 2019: 148-179.
[11]	马向亮, 王宏, 李冰 ,等. 基于能量分析技术的芯片后门指令分析方法[J]. 电子学报, 2019,47(3): 686-691.
	MA X L , WANG H , LI B ,et al. A power analysis method against backdoor instruction in chips[J]. Acta Electronica Sinica, 2019,47(3): 686-691.
[12]	BENADJILA R , PROUFF E , STRULLU R ,et al. Deep learning for side-channel analysis and introduction to ASCAD database[J]. Journal of Cryptographic Engineering, 2020,10(2): 163-188.
[13]	BENADJILA R , PROUFF E,STRULLUR ,et al. Deep learning for side-channel analysis and introduction to ASCAD database[J]. Journal of Cryptographic Engineering, 2020,10(2): 163-188.
[14]	CAI X M , LI R F , KUANG S J ,et al. An energy trace compression method for differential power analysis attack[J]. IEEE Access, 2020(8): 89084-89092.
[15]	DUAN X Y , CHEN D , FAN X H ,et al. Research and implementation on power analysis attacks for unbalanced data[J]. Security and Communication Networks, 2020:5695943.

	TA	KNN（k=7）
PCC	0.868	0.923
MIC	0.897	0.940
Dcorr	0.857	0.926

基于模板和KNN算法的能量分析攻击对比研究

Comparative study of power analysis attacks based on template and KNN algorithm

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 15

相关文章 0

Metrics

推荐阅读 0

	攻击成功率	鲁棒性	运行速度	占用内存
模板攻击	低	弱	快	小
KNN算法攻击	高	强	慢	大