针对目前服务网格资源管理中存在的信任机制与调度机制分离的缺陷,基于网格信任模型与信任效益函数,讨论了信任Qos增强的计算服务调度问题.分析传统调度算法的缺陷,提出了基于信任关系的网格服务调度算法.该算法在保证服务性能QoS要求的同时兼顾信任QoS的要求.对基于信任关系的网格服务调度算法性能进行了多角度分析和大规模仿真实验.结果表明:该算法较传统的基于性能QoS的调度启发式和另外两种基于信任QoS的调度启发式而言,不仅具有较优的平均信任效益、总信任效益和较小的最终服务期,而且在失效服务数和系统吞吐率等方面具有较好的综合性能.
At present existing scheduling algorithms in service grid largely ignore the impact of trust mechanism. It is rather doubtful to adopt these algorithms in a realistic environment. Based on the grid trust model and trust utilization functions, a computational service scheduling problem based on trust QoS enhancement is proposed. A novel trust-QoS enhanced heuristic based on trust relationship is put forward. The algorithms are evaluated with large scale simulation. Simulation results demonstrate trust-driven heuristics perform better than time-driven heuristic. The algorithm based on trust relationship achieves better integrative performance on performance- QoS, trust-QoS, etc. than trust-driven Min-Min and Sufferage heuristic algorithms.