基于演化博弈的Serverless移动目标防御决策方法  

Moving Target Defense Decision Method Based on Evolutionary Game in Serverless Computing

作　　者：刘轩宇 张帅 LIU Xuanyu;ZHANG Shuai(Information Engineering University,Zhengzhou 450001,China;Purple Mountain Laboratory,Nanjing 210023,China)

机构地区：[1]信息工程大学,河南郑州450001 [2]紫金山实验室,江苏南京210023

出　　处：《信息工程大学学报》2024年第2期205-212,共8页Journal of Information Engineering University

基　　金：国家自然科学基金资助项目(62072467);国家重点研发计划资助项目(2021YFB1006200,2021YFB1006201)。

摘　　要：“无服务器”(Serverless)是一种新型计算范式,具有轻量、敏捷的特点,由于其架构的特殊性,导致在原有云安全问题的基础上又引入一些新的安全威胁。针对在未知攻防场景下难以准确选择最优化防御策略的问题,从虚拟化层和应用层两个方面根据移动目标防御(Moving Target Defense,MTD)多样化的思想提出防御策略。将具有玻尔兹曼探索的Q-Learning算法与复制动态方程结合,从有限理性的角度出发构建具有探索机制的演化博弈模型,使防御者在反复的攻防对抗中不断进行试错—探索—实施,最终获得最优化防御策略,获取最大收益。实验表明,引入探索机制的演化博弈模型具有可预测性,且演化博弈均衡点具有较强的稳定性。Serverless is a new computing paradigm with lightweight and agile characteristics,due to the particularity of its architecture,some new security threats are introduced on the basis of the original cloud security issues.To address the problem that it is difficult to accurately select the optimal defense strategy in unknown offensive and defensive scenarios,a defense strategy is proposed based on the diversified ideas of MTD from the aspects of virtualization layer and application layer.The Q-Learning algorithm with Boltzmann exploration is combined with the replication dynamic equation to construct an evolutionary game model with an exploration mechanism from the perspective of bounded rationality.Defenders can continuously carry out trial and error,exploration,implementation in repeated offensive and defensive confrontations,and finally obtain the optimal defense strategy and the maximum benefit.Experiments show that the evolutionary game model introducing the exploration mechanism is predictable and has strong stability at the equilibrium point of the evolutionary game.

关 键 词：无服务器 移动目标防御 演化博弈 Q-Learning复制动态方程 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]