摘要
将分布式系统的任务分配算法与处理器局部调度算法相结合,提出一种主动备份的、基于EDF的分布式系统实时容错调度算法,其特点是主/副版本执行时间可以重叠。给出了该调度算法的任务集可调度的充分条件、任务集可调度所需最小处理器个数的计算方法。模拟结果比较了主动备份容错调度算法与被动备份容错调度算法,结果表明主动备份算法效率更优。
The paper presents an initiative-backup fault-tolerant real-time scheduling algorithm for distributed systems. In the scheduling, processor＇s local scheduling algorithm is EDF and the fault-tolerance＇s technique is primary/backup copies. The execution time of the primary/backup copies may be overlapped. The schedulability of task set is analyzed and the method of calculating the minimal number of processor is given. The simulation compares the initiative-backup algorithm with the passive-backup algorithm, the result shows that the initiative-backup algorithm is more effective than the passive-backup algorithm.
出处
《计算机工程》
CAS
CSCD
北大核心
2006年第4期76-78,共3页
Computer Engineering
基金
国家"十五"预研基金资助项目(41315040106)
关键词
分布式系统
主动备份
容错调度
主版本/副版本
EDF算法
Distributed system
Initiative-backup
Fault-tolerant scheduling
Primary/backup copies
EDF algorithm
