基于相关性分析的网络故障预测

在网络管理领域,相关性分析愈来愈发挥出重要的作用。与传统的专家系统方法相比,相关性分析等数据挖掘方法,不仅能够有效克服知识获取、更新困难的瓶颈,而且,能够从海量网络管理信息中,快速挖掘出先前未知的却有潜在价值的信息和模式。通过对故障、告警数据的讨论分析,文章研究利用相关性分析,从历史告警序列数据中,挖掘潜在的相关性规则,讨论并定义了相关性的类型、相关性规则的表示语言、相关性规则的生成算法,最后讨论了运用相关性规则进行网络故障预测的方法。

Network Fault Prediction Based on Alarm Correlation

Alarm correlation plays a more and more important role in network management domain.As compares with expert systems,alarm correlation,as a data mining method,not only can overcome the bottleneck of knowledge retrieval and knowledge update,but also can discover potential and valuable information and patterns from massive alarms.This paper discusses the indispensability and practicability of using alarm correlation in network fault prediction via the discussion and definition about fault and alarm data.Then it discusses the creation algorithm of correlation patterns and correlation rules based on the discussion of correlation type's definition and correlation rule's presentation.After that,it discusses the method and process to predict network fault using correlation rules.