【摘要】 本文从有效利用资源的角度出发,提出了一种以最小化处理机数目为优化目标的Best-Fit启发式容错调度算法。该算法采用主/副版本备份技术和副版本的主动运行方式与被运行方式相结合的方法,将实时任务的主版本和副版本调度到不同处理机上运行;并且按照Best-Fit启发式策略为实时任务主版本寻找“最佳满足”处理机,使尽可能多的实时任务副版本以被动方式运行。算法既保证了系统的实时性和容错性,也节约了处理机。分析和仿真结果均证明了算法的有效性。更多还原

【Abstract】 This paper proposes a fault-tolerant scheduling algorithm based on best-fit heuristics, which aims at minimizing the number of processors to achieve efficient utilization of resources. By the use of the primary/backup approach and the integration of passive backup and active backup, the primary copy and the backup copy of a realtime task are scheduled on two different processors. The best-fit heuristics try to find" best-fit" processors for scheduling primary copies such that as many backup copies as possible can be executed passively. This not only assures that real-time tasks complete before the deadlines even in the presence of processor failures, but also decreases the number of processors. Both analysis and simulation results show the effectiveness.更多还原