基于代理的网格用户信任模型  

An agent-based trust model for grid users

作　　者：韦凯[1] 唐星[2] 刘欣欣[1] 

机构地区：[1]华南理工大学计算机科学与工程学院,广东广州510006 [2]广西大学商学院,广西南宁530004

出　　处：《广西大学学报（自然科学版）》2011年第2期297-302,共6页Journal of Guangxi University（Natural Science Edition）

基　　金：广东省科学基金研究项目(10451064101005826);中央高校基本科研业务费专项资金项目(2009ZM0112)

摘　　要：针对传统网格计算的信任管理模型中存在的诸如网格服务提供者(GSP)难以有效评估和管理数量庞大的用户的信任度,难以抵御wh itewash ing攻击保障自身安全等问题,一个基于代理的网格用户信任模型被提出以解决上述问题。模型由自治域的用户信任代理(UTA)管理域内用户的信任度,GSP则判断UTA的评估结果。实验表明该模型符合网格架构的结构特点,克服了传统信任推荐方法的弱点,使GSP对网格用户的信任评估更加准确合理。An agent-based trust evaluation model is proposed in this paper,targeting at the problems in traditional trust model for Grid users,such as difficulty for the Grid Service Provider（GSP） to conduct effective evaluation and management of the trust degree in a circumstance of a huge number of Grid users,and the difficulty for resistance against the whitewashing attack.In the proposed model,the User Trust Agent（UTA） is used to evaluate and manage user＇s trust,and then the evaluation result of UTA is judged by GSP.Experiments showed that the proposed model,which meets the requirement of Grid architecture,was able to overcome the weakness of the traditional evaluation system and was cable of achieving reasonable and accurate evaluation result for the GSP.

关 键 词：网格计算 自治域 主观信任 whitewashing攻击 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]