网络告警知识发现研究与实现

Study and Implementation of Knowledge Discovery in Network Alarm

文章研究企业网络告警数据中的知识发现问题,设计并实现了以Apriori算法为核心的网络告警关联规则发现系统。系统试运行结果表明,该系统能够有效发掘隐藏在海量告警数据背后、不易为网络管理人员所知的告警及故障模式知识。将发现的新知识应用到告警关联/故障诊断专家系统,有效突破了专家系统“知识获取”瓶颈,显著增强了专家系统推理和诊断网络故障的能力。

Knowledge discovery in enterprise network alarm data is studied and an association rule discovery system for network alarm is implemented based on Apriori algorithm.Preliminary test shows,knowledge regarding alarm and fault pattern,which is always hided behind extremely large alarm data and is difficult to be understood by network management operators,can be discovered effectively.Application of these rules to alarm correlation and/or fault diagnosis expert systems will break the bottle-neck of knowledge acquisition and improve expert system's capabilities of reasoning and diagnosing network abnormalities as well.