###
DOI:
Journal of Software:2007.18(10):2584-2594

Internet服务故障管理:分层模型和算法
黄晓慧,邹仕洪,褚灵伟,程时端,王文东
(北京邮电大学,网络与交换国家重点实验室,北京,100876)
Internet Services Fault Management: Layering Model and Algorithm
HUANG Xiao-Hui,ZOU Shi-Hong,CHU Ling-Wei,CHENG Shi-Duan,WANG Wen-Dong
()
Abstract
Chart / table
Reference
Similar Articles
Article :Browse 3372   Download 3405
Received:April 26, 2006    Revised:February 15, 2007
> 中文摘要: 在基于时间窗口的Internet服务故障管理中,时间窗口大小设置不合适会给算法准确度带来影响.为了降低这种影响,分析了Internet服务故障管理中存在的问题,提出了分层故障管理模型,采用图论技术进行故障诊断,选择二分图作为各层的故障传播模型.提出了基于时间窗口的故障诊断算法--多窗口故障诊断(multi-window fault diagnosis,简称MFD),该算法通过综合考虑相邻时间窗口之间的关联关系,在一定程度上降低了因时间窗口大小设置不合适而给算法准确度带来的影响.仿真结果证明了MFD算法的有效性和效率.
Abstract:In window-based Internet service fault management,improper time window size setting will affect the fault diagnosis algorithm. In order to reduce the impact,challenges of Internet service fault management are analyzed in this paper,and a layering model is recommended. Bipartite graph is chosen to be the fault propagation model (FPM) for each layer. A window-based fault diagnosis algorithm MFD (multi-window fault diagnosis) is proposed for the bipartite FPM. MFD takes the correlation of adjacent time windows into account. As a result,it can reduce the impact of improper time window size setting. Simulation results prove the validity and efficiency of MFD.
文章编号:     中图分类号:    文献标志码:
基金项目:Supported by the National Basic Research Program of China under Grant Nos.2003CB314806, 2006CB701306 (国家重点基础研究发展计划(973)); the National Natural Science Foundation of China under Grant Nos.60603060, 90604019, 60472067 (国家自然科学基金) Supported by the National Basic Research Program of China under Grant Nos.2003CB314806, 2006CB701306 (国家重点基础研究发展计划(973)); the National Natural Science Foundation of China under Grant Nos.60603060, 90604019, 60472067 (国家自然科学基金)
Foundation items:
Reference text:

黄晓慧,邹仕洪,褚灵伟,程时端,王文东.Internet服务故障管理:分层模型和算法.软件学报,2007,18(10):2584-2594

HUANG Xiao-Hui,ZOU Shi-Hong,CHU Ling-Wei,CHENG Shi-Duan,WANG Wen-Dong.Internet Services Fault Management: Layering Model and Algorithm.Journal of Software,2007,18(10):2584-2594