一种网格环境下的动态故障检测算法  

A Dynamic Fault Detection Algorithm under Grid Environments

作　　者：田东[1] 陈蜀宇[2] 陈峰[2] 

机构地区：[1]重庆大学计算机学院,重庆400044 [2]重庆大学软件学院,重庆400044

出　　处：《计算机研究与发展》2006年第11期1870-1875,共6页Journal of Computer Research and Development

基　　金：教育部跨世纪优秀人才支持计划基金项目(NCET-04-0843);重庆市自然科学基金项目(2005BB2192)

摘　　要：针对现有网格系统出错几率较大、已有故障检测算法不能有效满足网格系统需求问题,提出了一种网格环境下的动态故障检测算法.根据网格系统的特点,基于不可靠故障检测思想,建立了网格系统模型和故障检测模型;结合心跳(heartbeat)策略和灰色预测方法,设计了一种动态心跳机制,并给出了预测模型和实时预测策略;提出了基于该动态心跳机制的网格故障检测算法,分析了算法的可靠性.仿真实验结果表明,该算法是正确、有效的,可用于网格环境下的故障检测.Aimed at the problems that grids are more prone to failures, and existing failure detection algorithms can not satisfy the unique requirement of grids, a dynamic failure detection algorithms is presented. According to the characteristics of grids and unreliable failure detection theory, the authors establish models for grid systems and fault detection respectively; Combining heartbeat strategy with grey prediction method, they also design a dynamic heartbeat strategy, and present the prediction model, as well as a real-time prediction strategy. Furthermore, on the basis of the dynamic heartbeat strategy, a fault detection algorithm is put forward for grid environments, and its dependability is analyzed. Finally, simulation results shows that the algorithm is valid and effective, and can be used for fault detection under grid environments.

关 键 词：网格 故障检测 灰色预测 心跳机制 

分 类 号：TP393.01[自动化与计算机技术—计算机应用技术]