一种分布计算系统自适应故障侦测方法  

An Adaptive Failure Detection Method for Distributed Computing Systems

作　　者：蔡京平[1] 贾云得[1] 

机构地区：[1]北京理工大学计算机科学工程系,北京100081

出　　处：《小型微型计算机系统》2007年第1期136-139,共4页Journal of Chinese Computer Systems

基　　金：国家"八六三"计划基金项目(2003AA735051)资助.

摘　　要：面向高可靠智能应用的分布计算系统,首先提出一组故障侦测服务的QoS度量标准,其次给出一种自适应故障侦测方法.该方法使用一个无需统计行为的高度动态的计算方法,动态地估算心跳消息超时时限,并协商改变心跳消息的发送周期,以适应分布计算系统计算节点和网络状态变化,提高故障侦测服务的QoS.模拟实验表明,该方法能够适应分布计算系统状况的变化,在侦测的实时性和正确性上提供较好的平衡.This paper proposes a set of OoS metrics for failure detection service, based on the distributed computing systems of high reliable intelligent applications. An adaptable heartbeat failure detection method is then present. This method dynamically estimates the heartbeat detection timeout using an dynamic algorithm of non-probabilistie behavior. It changes the sending interval of heartbeat according to the processor load and transmission delay of the system. Simulation results show that the failure detector can achieve a compromise between a good detection time and the need of avoiding false detections. It can improve the failure detection QoS of the distributed computing systems.

关 键 词：故障侦测 高可靠智能应用 分布计算系统 心跳法 适应性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]