结合协方差与变异系数的密码芯片能量泄漏评估模型

Energy Leakage Evaluation Model of Cryptographic Chip Combining with Covariance and Variation Coefficient

在信息技术安全性评估通用准则中,必须使用具体的侧信道分析方法来评估密码芯片工作时的能量泄漏情况。为降低评估过程对侧信道分析方法的依赖性,通过分析能量迹各点之间的关系,构建一种基于协方差矩阵变异系数的能量泄漏评估模型。利用协方差矩阵度量能量迹各点间的线性关系,并引入变异系数衡量矩阵内各元素的差异程度,从而评估芯片工作时的能量泄漏情况。基于DPA contest v2、SAKURA-G、DPA contest v4和ATMEGA2560数据集的实验结果验证了该模型的有效性。与现有密码芯片能量泄漏评估模型相比,其具有操作简单和可行性强的优点,可使实验人员无需深入了解侧信道分析知识背景即可快速评估芯片能量消耗情况。

The existing security evaluation criteria Common Criteria(CC)need a specific side-channel analysis method to assess the energy leakage of a working chip.In order to reduce the dependence on specific channel analysis methods,this paper proposes an energy leakage evaluation model based on variation coefficient by analyzing the relationship of points on power traces.In this model,the covariance matrix is used to measure the linear relationship among the points and the coefficient of variation is introduced to measure difference of the elements in the covariance matrix,so as to evaluate the energy leakage.Experimental results based on DPA contest v2,SAKURA-G,DPA contest v4,and ATMEGA2560 datasets verify the effectiveness of the model.Compared with the existing energy leakage evaluation models of cryptographic chip,this model has the advantages of simple operation and high feasibility.It enables researchers without a thorough understanding of side channel analysis.