移动边缘计算网络下基于静态贝叶斯博弈的入侵响应策略研究  

Research on intrusion response strategy based on static Bayesian game in mobile edge computing network

作　　者：范伟[1,2] 彭诚 朱大立 王雨晴[1,2] FAN Wei;PENG Cheng;ZHU Dali;WANG Yuqing(Institute of Information Engineering,Chinese Academy of Sciences,Beijing 100093,China;School of Cyber Security,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院信息工程研究所,北京100093 [2]中国科学院大学网络空间安全学院,北京100049

出　　处：《通信学报》2023年第2期70-81,共12页Journal on Communications

基　　金：国家重点研发计划基金资助项目(No.2019YFB1005204)。

摘　　要：针对移动边缘计算环境下边缘节点资源受限、入侵过程难以被准确检测且缺乏有效应对外部入侵的入侵响应策略的问题,提出了一种适用于移动边缘计算环境的入侵检测网络结构,建立了基于静态贝叶斯博弈的入侵响应决策模型,模拟边缘节点与外部入侵者的网络交互行为,并对博弈过程中攻击者和防御者选择不同行为的概率进行了预测。入侵响应决策模型综合考虑系统资源、响应成本以及检测率、误报率和漏报率等因素,在兼顾入侵检测系统资源消耗及边缘节点隐私保护的基础上,对入侵检测系统的响应决策进行优化。实验分析了影响入侵响应决策的因素,为具体应用提供了实验依据。In the mobile edge computing(MEC)environment,the resources of edge nodes are limited.It is difficult to detect the intrusion process accurately,and there is no effective intrusion response strategy to deal with external intrusions.An intrusion detection network structure suitable for mobile edge computing environment was proposed and an intrusion response decision model based on static Bayesian game was established to simulate the network interaction behavior between edge nodes and external intruders.The probability of attackers and defenders in the game process was predicted respectively.The influence of the system resource,the cost of intrusion response,the detection rate and false alarm rate were considered comprehensively by the intrusion response decision model.The response decision of the intrusion detection system was optimized on the basis of the considering both resource consumption of the intrusion detection and the privacy protection of the edge nodes.The factors that affected the decision-making of intrusion response were analyzed,and the experimental basis for the specific application was provided.

关 键 词：移动边缘计算 静态贝叶斯博弈 贝叶斯纳什均衡 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]