针对时间限制严格的DAG(有向无环图)类型网格工作流提出了一种新的基于排队论时间预测模型的网格工作流调度算法。按照用户提交的工作流执行时间的要求,利用Little公式计算任务在各候选资源上执行时间超过规定时间概率的大小,然后选择最佳的候选资源执行工作流任务。仿真结果证明了该算法的有效性。
To the workflow Scheduling of DAG-based workflow in which time constraints are rind, a new algorithm based on the model of Time-Prediction of Queuing Theory was put forward. According to the requirement of workflow execution time submitted by users, this algorithms found critical paths first and obtained probability for actual execution time extending beyond a limit set by users using Little formula, and then made choice of appropriate resource for workflow task. The simulation shows the validity of algorithm.