论文部分内容阅读
为了解决分布式计算系统回卷恢复容错的验证评估问题，设计一种分布式计算系统的回卷恢复容错算法的仿真机制，依据分布式计算系统回卷恢复容错的总体架构，将分布式计算系统中的节点任务过程使用离散事件模拟，在网络系统仿真工具的应用层增加支持多任务回卷恢复容错仿真的模块，并设计用于回卷恢复容错仿真的结构、功能模块和系统参数设定。结果表明本文提出的仿真机制能够实现分布式计算系统的回卷恢复容错算法的模拟验证，为不同容错算法间对比、改进与优化提供参照。
In order to solve the verification and evaluation problem of fault recovery in distributed computing system, a simulation system of rollback fault tolerance algorithm in distributed computing system is designed. According to the overall architecture of fault recovery in distributed computing system, the distributed computing system The node task process uses discrete event simulation to add modules that support multi-task wrap-around recovery fault-tolerant simulation in the application layer of network system simulation tools and design the structure, function modules and system parameter settings for roll-back recovery fault-tolerant simulation. The simulation results show that the simulation mechanism proposed in this paper can simulate the algorithm of rollback recovery fault tolerant in distributed computing system and provide a reference for the comparison, improvement and optimization of different fault tolerant algorithms.