计算机网络协同防御系统信任启动模型  

Trust bootstrapping model for computer network collaborative defense system

作　　者：余洋[1] 夏春和[1] 原志超[1] 李忠[1] 

机构地区：[1]北京航空航天大学计算机学院,网络技术北京市重点实验室,北京100191

出　　处：《浙江大学学报（工学版）》2016年第9期1684-1694,共11页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(61170295)

摘　　要：为了解决计算机网络协同防御(CNCD)中的信任启动问题,研究信任启动过程,并给出基于信任类型的信任启动模型.基于防御代理(Agent)的行为模式划分信任类型,并使用博弈论分析防御任务的信任效用和防御成本之间的关系,据此给出执行防御测试任务的约束条件.通过执行防御测试任务,确认防御Agent的信任类型,并设计赋对应值和加权均值2种方法来计算新加入的防御Agent的初始信任值.模拟实验结果表明,相对于其他模型,基于信任类型的信任初始化模型具有更低的防御任务失败率,同时在防御任务的执行方面具有更好的适应性.The process of trust bootstrapping was investigated, and the trust type based trust bootstrapping model was proposed in order to solve the problem of trust bootstrapping in computer network collaborative defence （CNCD）. First, the division of trust type based on the behavior model of defense Agent was dis-cussed. Then, the relationship of trust utility and defense cost was analyzed according to the game theory. Based on the above analysis,the constraints of executing defense test task were proposed. The trust type of defense Agents was identified by executing defense test tasks. The initial trustworthiness of defense A-gents was assigned by two means, including assigning the corresponding values and weighted mean. The simulated experiment results show th a t, relative to other models, the proposed model has lower failure rate of defense tasks, and better adaptability in the aspect of defense task execution.

关 键 词：协同防御 信任启动模型 信任类型 信任效用 防御代理(Agent) 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]