一种新的网络告警相关系统的体系结构

New architecture for alarm correlation in network systems

事件相关通过对事件在时间和空间上进行相关处理 ,对告警进行事件关联 ,就可以改变故障管理中“数据丰富 ,信息贫乏”的局面 ,有效地进行故障定位。基于数据挖掘方法 ,给出了一种新的告警相关系统框架 ,根据时序关系 ,自动找出隐藏在大量事件中的规律性 ,形成关联规则 ,解决知识获取这一瓶颈。它克服了传统事件关联方法的缺点 ,不仅可以自主获取先验知识 ,而且解决了多故障定位问题 。

Fault management requires a sophisticated mechanism to associate events to meaningful fault report which can be addressed by event correlation. A novel architecture based on data mining for discovering rules from events database was introduced. The basic idea was to search for the regularity in events according to their time and space orders. Compared with traditional event correlation methods, it could acquire knowledge initiatively,solve the problem of multi-fault location, and predict the failure of services.