Randomness Assessment of an Unpredictable Random Number Generator based on Hardware Performance Counters
Randomness Assessment of an Unpredictable Random Number Generator based on Hardware Performance Counters
复制标题
基于硬件性能计数器的不可预测随机数生成器的随机性评估
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
A. Suciu
中科院分区：
文献类型：
--
作者：
K. Marton;Alexandra Zaharia;Sebastian Banescu;A. Suciu
。硬件性能计数器捕获和测量处理器内的硬件事件，提供有关运行系统或用户级程序时处理器内各种微架构组件的内部行为和性能的详细指示。我们之前研究过性能计数器捕获的硬件事件为生成不可预测的随机数序列提供熵源的潜力，现在我们通过各种延迟类型和简单的操作扩展了[1]中引入的生成器模型，这些延迟类型和简单的操作对处理器的内部行为产生了很大的影响。使用统计和视觉随机性测试对不同配置的生成器产生的序列的随机性质量进行了彻底评估，这些测试突出了生成器的特性和局限性，并显示了新设计所实现的改进。
. Hardware performance counters capture and measure hardware events within the processor, providing a detailed indication on the internal behavior and performance of the various micro-architectural components within the processor while running system or user level programs. We have previously investigated the potential of hardware events captured by performance counters to provide entropy sources for generating unpredictable random number sequences, and now we extend the generator model introduced in [1] by various delay types and simple operations which show high impact on the internal behavior of the processor. The randomness quality of sequences produced by the generator in different configurations are thoroughly assessed using statistical and visual randomness tests which highlight the characteristics and limitations of the generator and show the improvements achieved by the new design.