实时分布系统的动态容错调度算法

A dynamic fault-tolerant scheduling algorithm for a real-time distributed system

分布系统中如何保证具有容错需求任务的实时性是一个具有挑战性的课题,目前针对不同的应用背景已产生了一些有意义的研究成果,但当任务的截止时间不足以让主、副两版本分别单独运行时,尚未作处理.这里提出了基于PB的改进时间冗余的启发式动态容错调度算法FTPB,在算法中根据任务的开始时间、执行时间、截止期限之间的内在关系,分别采用副版本主动复制与被动复制策略,使系统在不增加额外资源的前提下,既有较好的容错能力,又有效地利用CPU资源,同时还能提高任务的接受率.模拟实验结果表明:算法FTPB优于同类其他算法,具有较好的实用价值.

In a real-time distributed system ,it is critical to have a fault-tolerant task execution scheme to meet the tasks' deadlines even in the presence of one processor failure. Although interesting scheduling algorithms are available, including the algorithm in which a task's primary copy is not executed with its backup copies at the same time.Thus far,research could not handle task deadlines that are insufficient to run the primary copies and backup copies respectively. A new dynamic fault-tolerant algorithm FTPB based on PB is offered here.In this algorithm , an active and a passive replica are utilized according to the relationship between Ri+Ci+Vi and Di. Simulation results show that the FTPB algorithm is effective, has better fault-tolerant ability, and can improve efficiently the CPU utilization and tasks acception ratio.