摘要
告警关联分析是通信网故障管理的重要手段,通过告警关联分析网络维护人员能够进行告警压缩和故障定位。它是智能运维的关键技术之一,因此全面而系统地对告警关联分析的步骤和算法进行总结非常必要。本文首先对基于机器学习的通信网告警关联分析的步骤进行了介绍,然后对告警预处理、告警关联分析算法、告警关联规则生成等步骤中目前学者研究的主要问题和方法进行了分析比较。接下来对基于大数据处理框架的告警关联分析实现进行了对比,最后对告警关联分析的研究进行了总结和展望。
Alarm correlation analysis is the main technology for fault management in telecommunication networks.By means of alarm correlation analysis,the network maintainer easily reduces the redundant alarms and locates the root alarm.It is also one of the key technologies for AIOps.Therefore,it is significant to summarize the procedure and algorithm for alarm correlation analysis.As a first step,this paper introduces the general procedure for machine learning-based alarm correlation analysis in telecommunication network.In the following,the current research status for several key steps like alarm preprocessing,correlation analysis algorithm,and correlation rule generation are introduced and compared in detail.Then,the implementation of alarm correlation analysis based on big data processing framework has been summarized.Finally,the conclusion is drawn and the future research direction of alarm correlation analysis is also proposed.
作者
丁宏
周宏林
DING Hong;ZHOU Honglin(China Mobile Group Sichuan Co.,Ltd.,610000,Chengdu,China;DEC Academy of Science and Technology Co.,Ltd.,611731,Chengdu,China)
出处
《东方电气评论》
2021年第1期77-84,88,共9页
Dongfang Electric Review
关键词
告警关联分析
关联规则挖掘
机器学习
数据挖掘
故障管理
alarm correlation analysis
correlation rule mining
machine learning
data mining
fault management