网格环境下基于信任模型的动态级调度  

Dynamic Level Scheduling Based on Trust Model in Grid Computing

作　　者：袁禄来[1] 曾国荪[1] 姜黎立[1] 蒋昌俊[1] 

机构地区：[1]同济大学计算机科学与工程系,上海201804

出　　处：《计算机学报》2006年第7期1217-1224,共8页Chinese Journal of Computers

基　　金：本课题得到国家自然科学基金(60173026);中国教育部科研重点项目基金(105071);上海高校网格技术E-研究院项目(200301-1)资助.

摘　　要：网格用户、资源和服务的不确定性潜在地影响网格应用任务的正常执行,这样使得设计既能减小应用任务执行时间又能减小欺骗可能性的调度算法十分困难．参考社会学的人际关系信任模型,建立网格节点信任推荐机制,并利用D-S理论对推荐证据进行综合分析,从而定义出基于不确定性推理理论的信任度计算函数．将该函数并入DLS算法得到“可信”动态级调度算法(TDLS),从而在计算调度级别时考虑网格节点的可信程度．仿真结果证实,提出的TDLS算法以小的时间花费为代价,能有效提高任务在信任方面的服务质量需求．The uncertainty of Grid users, resources and services may play a negative affect on the execution of Grid tasks, which makes it difficult to design a scheduling algorithm to minimize execution time and cheat probability of Grid tasks. Referring to the social trust relationship, the authors introduce a trust model that is based on the uncertainty reasoning theory （D-S theory）. In addition, by combining the trust model and Dynamic Level Scheduling （DLS） algorithm, the authors propose a novel scheduling algorithm that is called Trustworthy and Dynamic Level Scheduling （TDLS）. The algorithm takes the Grid nodes＇ trust degree into account when calculating the scheduling-level of task-node pairs. Simulations prove that the algorithm can efficiently satisfy the QoS requirement in trust, with costing a few more time.

关 键 词：网格计算 可信调度 信任模型 不确定性推理 DLS算法 

分 类 号：TP338[自动化与计算机技术—计算机系统结构]