云计算环境下基于行为预测的博弈信任模型  

Game-theoretic Trust Model Based on Prediction of User Behavior in Cloud Computing Environment

作　　者：刘莹[1,2] 罗永龙[1,2] 乔云峰[1,2] 陈付龙[1,2] 赵传信[1,2] 

机构地区：[1]安徽师范大学计算机科学与技术系,安徽芜湖241003 [2]安徽师范大学网络与信息安全工程技术研究中心,安徽芜湖241003

出　　处：《小型微型计算机系统》2015年第10期2327-2333,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61370050)资助;安徽省自然科学基金项目(1308085QF118)资助

摘　　要：为解决云计算环境中用户和云服务提供商之间信任危机问题,提出一种基于行为预测的博弈信任模型.该模型设立第三方云审计代理,利用贝叶斯网络预测云服务提供商的行为信任等级,并采用博弈机制根据用户和云服务提供商的行为信任等级监测他们的行为.博弈模块将信任预测结果结合支付矩阵进行博弈分析,计算出用户和云服务提供商与云环境的混合纳什均衡策略,依据混合纳什均衡策略调整云审计代理对用户和云服务提供商的监测概率.实验表明本模型可以预测云服务提供商的行为类型,约束不可信用户与云服务提供商的恶意行为,降低实时监控的开销.It is difficult for the users and the service providers to establish a trust relationship in cloud computing environment. We propose a game-theoretic trust model based on prediction of user behavior to solve the problem of trust crisis. The model sets up a third party cloud audit agency. The audit agency forecasts the behavior trust of the cloud service providers using Bayesian network, and monitors the users and the service providers according to their behavior trust level using the game mechanism. The game module conducts game analysis by combining the predicted results with the payoff matrix, calculates the Mix Nash equilibrium strategy between the user and service provider and the cloud environment. Then the audit agency controls the monitoring probability for the users and the service providers on the basis of Mix Nash equilibrium strategy. The experimental show that the model can predict the behavior type of service providers, constrain the malicious behavior of the un-trusted users and service providers, and reduce the real-time monitoring overhead.

关 键 词：云计算 云审计代理 信号博弈 行为预测 行为监控 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]