使用simgrid模拟环境,预先设定好资源的出错情况,分别使用简单重试、主动备份、被动备份、检查点四种容错策略对作业调度进行模拟。比较批量作业的平均执行时间和最终完成时间,分析容错策略对上述两种时间的影响,以期给网格环境下容错的方案设计提供指导。
In this paper,simgrid simulative environment is utilised to simulate jobs' schedule.The error conditions of the resources are set in advance,then four fault-tolerance policies were employed separately including simple retry,active backup,passive backup and checkpoint.The average runtime and final accomplishment time of the batch job were compared for analyzing the influence the fault-tolerant policies have on above two times.The purpose of this paper is to provide references to the design of fault-tolerant schemes on grid environment.