基于能量迹预处理的AES算法相关功耗分析

发布时间：2018-01-18 03:35

本文关键词：基于能量迹预处理的AES算法相关功耗分析　出处：《中国科学技术大学》2017年硕士论文　论文类型：学位论文

【摘要】：旁路分析作为一种新兴的密码分析方法,已经越来越被广泛的相关学者所关注。与传统的密码分析学不同,旁路分析不要求研究者掌握深厚的数学功底。它利用密码算法在执行时泄露的一些旁路信息,例如执行时间差异、能量消耗/电磁辐射、Cache泄露等,通过分析这些旁路信息与中间密钥或中间计算之间的联系,从而破解密钥,攻破密码系统。由于原理简单易行,实验限制条件较少等因素,功耗分析成为了旁路分析中一种基础的分析方法。相关功耗分析利用密码算法执行时产生的功耗与中间计算过程中的子密钥信息之间的联系,结合统计学中相关性计算的理论,计算出功耗与密钥之间的相关性,从而破解出密码算法的密钥。本文结合前人提出的相关功耗分析方法,以AES算法为研究对象,对采集到的能量迹进行预处理,以完善整个相关功耗分析模型。具体从以下3个方面来实现:1、提出了一种基于差分的特征点提取方案,该方案既简单快捷又容易理解。本文利用不同的密钥在进行加密时产生的能量消耗之间存在微小的差别,通过多次进行随机明文与两个不同密钥加密之后求得两者均值的差分轨迹。在这个均值差分结果的轨迹中,幅度值较大对应的索引范围即为我们想要的特征点范围。2、改进了前人利用频域进行相位替换的能量迹对齐的方式。利用傅里叶变换中的时域平移特性,对前人的成果进行改进,使得对齐预处理方法更简便、更适用。3、引入小波分析的概念,第一次将其应用到旁路分析的能量迹预处理中。针对数字滤波器中傅里叶变换在处理非平稳信号的局限性,引入小波分析的概念,对我们采集到的能量迹进行信噪分离处理。通过上述三种预处理方案,对采集到的能量迹进行预处理之后,再通过相关功耗分析,能够以更快的速度,更少的时间和空间资源来分析出AES的密钥,并且成功率也得到一定的提高。
[Abstract]:As a new cryptographic analysis method, bypass analysis has been paid more and more attention by many scholars. It is different from traditional cryptography. Bypass analysis does not require researchers to have a solid mathematical background. It uses cryptographic algorithms to leak some bypass information during execution, such as execution time differences, energy consumption / electromagnetic radiation cache leaks and so on. By analyzing the relationship between the bypass information and the intermediate key or intermediate calculation, the key can be cracked and the cryptographic system can be broken. Power analysis has become a basic analysis method in bypass analysis. Correlation power analysis uses the power generated when the cryptographic algorithm is executed and the relationship between the sub-key information in the intermediate computation process. Based on the theory of correlation calculation in statistics, the correlation between power consumption and key is calculated, and the key of cryptographic algorithm is solved. Taking the AES algorithm as the research object, the collected energy traces are preprocessed to improve the whole model of power analysis. In this paper, a difference based feature point extraction scheme is proposed, which is simple, fast and easy to understand. In this paper, there is a small difference between the energy consumption generated by encryption with different keys. The differential locus of the mean value is obtained by encrypting the random plaintext and two different keys several times. In the trajectory of the difference result of the mean value. The index range corresponding to the larger amplitude is the desired range of feature points. 2, which improves the energy trace alignment in frequency domain, and makes use of the time-domain translation characteristic in Fourier transform. The previous achievements are improved to make the alignment preprocessing method easier and more applicable. 3. The concept of wavelet analysis is introduced. For the first time, it is applied to the energy trace preprocessing of bypass analysis. Aiming at the limitation of Fourier transform in digital filter to deal with non-stationary signal, the concept of wavelet analysis is introduced. Through the above three preprocessing schemes, after preprocessing the collected energy traces, and then through the analysis of power consumption, we can achieve faster speed. Less time and space resources to analyze the AES key, and the success rate is also improved.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN918.4

【相似文献】