基于业务时间窗选取的告警聚类及关联方法  

A Method for Alarm Clustering and Correlation based on Business Time Window Selection

作　　者：杨剑[1] 蓝明超 YANG Jian;LAN Ming-chao(School of Information and Communication,National University of Defense Technology,Xi’an 710106,China)

机构地区：[1]国防科技大学信息通信学院,西安710106

出　　处：《光通信研究》2019年第1期33-36,共4页Study on Optical Communications

基　　金：军队装备预研基金资助项目(ZX-ZB-2016-003)

摘　　要：基于时间窗的告警事务集选取直接影响告警关联规则挖掘的有效性,同时,智能光网络告警的内在关联性使其具备一定的聚类关系,据此提出一种基于业务时间窗选取的告警聚类及关联方法。首先根据网络业务配置情况对多网元告警进行聚类,然后通过确定告警增长期进行时间窗划分,生成关联性更强的告警事务集,并进一步挖掘告警关联规则。该方法能够适应智能光网络多平面和多网元告警内在关联的特点,可有效压缩告警事务集,提高告警关联规则的准确度,为后续故障诊断提供支撑。最后运用实际网络数据验证了该方法的可行性和有效性。The selection of alarm transaction set based on time window directly affects the effectiveness of alarm association rule mining,and the internal correlation of Automatically Switched Optical Network(ASON)alarm makes it have a certain clustering relationship.Therefore,we propose a alarm clustering and correlation method based on business time window selection.The multi-network alarm is first clustered according to the network business configuration.Then the time window is divided to generate a more strongly correlated alarm transaction set by determining the alarm growth period.Finally,we also investigate the alarm association rules.This method can adapt to the characteristics of multi-plane and multi-network alarm internal correlation of intelligent optical network,which has the advantages of effectively compressing the alarm transaction set,higher accuracy of alarm association rules,and support for subsequent fault diagnosis.The feasibility and effectiveness of this method are verified by the practical network data.

关 键 词：智能光网络 网络业务 时间窗 告警聚类 告警关联 

分 类 号：TN915[电子电信—通信与信息系统]