智能监控应用平台告警关联大数据分析算法研究

Big data analysis algorithm research on alarm correlation of communication system

本文以支撑故障管理体系为目标,构建基于Spark/Hadoop的电信网络智能监控应用平台并实现对海量告警数据的挖掘与分析,研究基于距离的告警关联大数据分析算法,实现告警自关联及故障快速定位。通过对实际数据的测试及分析,该方法能够实现告警关联的自动化、智能化,聚合原有关联规则以提升规则有效性,提升了告警压缩率和故障命中率。

Based on the general requirement of the existing network centralized alarm monitoring and fault handling,the technology architecture of intelligent monitoring system based on Spark/Hadoop was proposed which can dig and analyze massive alarm data.A distance-based alarm big data correlation analysis algorithm was mainly studied.Through analyzing the actual data,this method can automatically and intelligently correlate the alarms and aggregate the rules of alarm correlation so as to enhance the validity of the rules and improve the alarm compression ratio and the trouble-shooting ratio.