【摘要】基于時(shí)間冗余的容錯(cuò)實(shí)時(shí)調(diào)度算法，其可調(diào)度條件是以任務(wù)的最壞執(zhí)行時(shí)間和為任務(wù)預(yù)留容錯(cuò)操作的時(shí)間為分析基礎(chǔ)，且任務(wù)拒絕率高，系統(tǒng)資源利用率低。該文提出的基于資源回收的容錯(cuò)單調(diào)比率調(diào)度，以容錯(cuò)實(shí)時(shí)調(diào)度算法FT-RM為基礎(chǔ)，利用任務(wù)的實(shí)際執(zhí)行時(shí)間低于最壞執(zhí)行時(shí)間的特性，收集已完成任務(wù)釋放的資源供等待任務(wù)使用，在保證系統(tǒng)容錯(cuò)要求的同時(shí)，提高了任務(wù)吞吐量和系統(tǒng)資源利用率，并通過模擬實(shí)驗(yàn)分析算法性能，研究了影響算法調(diào)度性能的若干因素。
關(guān) 鍵 詞 容錯(cuò); 實(shí)時(shí)系統(tǒng); 容錯(cuò)單調(diào)比率調(diào)度; 資源回收

Fault-tolerant real-time scheduling algorithm through time redundancy, whose schedulability bound is based on the worst execution time of tasks and the time reserved for tasks’ fault-tolerant operation, has high rejection rate and low resource utilization. This paper presents fault-tolerant rate-monotonic scheduling with resource reclaim, based on FT-RM, to improve resource utilization and tasks throughoutput. This algorithm makes use of the attribute that the actual execution time of task is shorter than the worst execution time, to reclaim and reuse the resource released by the ended task. At last, the performance simulation and analysis of the scheduling algorithm are presented and factors that affect performance are investigated.