首先描述QoS调度问题,建立QoS需求模型;然后通过分析任务的依赖性,提出时间花费、资源价格和可靠性三种QoS参数的映射机制;最后针对网格环境的新特征,提出一种以优化用户效用为目标,基于QoS的关联任务调度算法(QBDTS_UO)。仿真实验结果表明,该算法能以较小的时间花费为代价,有效满足用户的QoS需求,并能大大提高网格资源的使用率。
This paper firstly described QoS scheduling problem and established a QoS demand model. Then proposed a QoS mapping mechanism for three QoS parameters which included completion time, execution cost and reliability in basis of analy. zing the relationship among tasks, Finally, proposed a computationally efficient static scheduling heuristics which was called QoS-based dependant task scheduling for utility optimization (QBDTS_UO). Simulation results show that QBDTS_UO is capable of meeting diverse QoS requirements for users, while can greatly increase the usage of grid resources.