容错云计算网络中自适应故障检测方案研究

Research on adaptive fault detection scheme in fault-tolerant cloud computing networks

当前大多数故障检测(FD)技术无法针对动态网络条件自动调整它们的检测服务参数,导致无法应用于实际环境中。针对容错云计算网络中的故障检测特性进行研究,提出了一种通用的容错云计算网络自适应故障检测(SFD)算法;基于上述通用算法,对当前故障检测器进行优化,提出了一种自适应累积型故障检测器,并对SFD算法的部署进行了研究。基于七种典型的广域网案例,比较评估了该故障检测算法与当前其他算法的性能。实验结果表明,该算法可以实现故障的自适应检测,满足用户需求,拥有较高的系统性能。

Most existed fault detector （FD） schemes didn＇t automatically adjust their detection service parameters for the dy- namic network conditions, thus they couldn＇ t he used for actual application. This paper explored FD properties with relation to the fault-tolerant cloud computing networks, and proposed a general self-tuning failure detection method in the fault-tolerant cloud computing networks. Based on the above general method, this paper proposed an automatic self-tuning failure detector （SFD） to optimize the existed FDs and explored the implementation of SFD. Finally, it comparatively evaluated the proposed failure detection scheme and existed schemes by extensive experiments in seven representative wide area network （WAN） ca- ses. The experimental results show the properties of the different FDs, and demonstrate that the proposed scheme can adaptive- ly detecte the fault to obtain corresponding services and satisfy user requirements, while maintaining good performance.