异构集群中基于优先级的任务容错调度算法

A Priority-based Fault-tolerant Task Scheduling Algorithm on Heterogeneous Clusters

提出一种基于任务优先级的容错调度算法,任务的优先级并不是由用户指定,而是由任务的开始时间、执行时间和截止期限决定的。该算法能够容忍异构集群系统中一个节点故障,采用主动/被动副版本相结合的执行方式和重叠技术的备份方式,提高系统资源利用率。通过实验证明了该算法在容错调度中的可靠性和高效性。

We propose a task priority-based fault-tolerant scheduling algorithm, the priority of the task is not specified by the us- er, but decided by the start time, execution time and deadline of the task. The algorithm can tolerate a heterogeneous cluster sys- tem node fails, the use of active/passive executive backup mode and overlapping ways to improve the system resource utilization. The experimental results show the reliability and efficiency of this algorithm in fault-tolerant scheduling.