差分侧信道密码分析中泄露模型的线性回归分析

发布时间：2018-01-17 23:29

本文关键词：差分侧信道密码分析中泄露模型的线性回归分析　出处：《同济大学学报(自然科学版)》2014年02期 　论文类型：期刊论文

【摘要】：从统计学线性回归模型的角度研究密码设备差分侧信道分析攻击中泄露模型的建模及估计,在不需对设备信息泄露有提前了解的情况下,得出线性回归泄露模型,克服了传统泄露模型的局限性.首先,分析能耗泄露的随机模型从而构建线性回归模型,然后用最小二乘估计和最小一乘估计两种方法求解线性回归模型的系数,最后基于八位控制器PayTV-AES智能卡平台实现能耗泄露的建模及系数估计.通过对两种求解方法结果的比较,提出最小二乘估计比最小一乘估计更适合用于泄露模型的线性回归分析;通过对被估模型系数曲线的分析,提出线性回归分析可以用于测量数据的预处理,以提高泄露模型建模效率.
[Abstract]:From the point of view of statistical linear regression model, this paper studies the modeling and estimation of the leak model in the channel analysis attack of the differential side of the cryptographic equipment, without the need to know the device information in advance. The linear regression leakage model is obtained, which overcomes the limitation of the traditional leakage model. Firstly, the stochastic model of energy leakage is analyzed to construct the linear regression model. Then the coefficients of the linear regression model are solved by the least square estimation and the least one multiplication estimation. Finally, based on the 8-bit controller PayTV-AES smart card platform, the energy leakage modeling and coefficient estimation are realized. The results of the two methods are compared. The least square estimation is more suitable for the linear regression analysis of the leaky model than the least square estimation. By analyzing the coefficient curve of the estimated model, it is proposed that the linear regression analysis can be used to preprocess the measured data in order to improve the modeling efficiency of the leak model.
【作者单位】：同济大学中德学院;同济大学数学系;
【基金】：国家自然科学基金(60903033) 中央高校基本科研业务费专项资金(0905915)
【分类号】：TN918.4
【正文快照】： 密码芯片实现中,密码算法基于某个物理设备并采用软件或硬件方式实现,物理设备会与其环境发生物理交互作用.攻击者有可能主动策划并检测这种交互作用,进而产生有助于密码分析的信息,比如运行时间、能耗、电磁信息等侧信道信息,对密码芯片实施侧信道攻击[1-2].相对于传统密码分

【相似文献】