基于空间资源的云化资源池故障根因智能推荐方案  

Intelligent recommendation scheme for root cause of cloud resource pool failures based on spatial resources

作　　者：许乾宸 李岩 章丹阳 庞雨薇 顾晓华 尤龙[1] 鲍一民 XU Qianchen;LI Yan;ZHANG Danyang;PANG Yuwei;GU Xiaohua;YOU Long;BAO Yimin(China Mobile Communications Group Jiangsu Co.,Ltd.)

机构地区：[1]中国移动通信集团江苏有限公司

出　　处：《江苏通信》2025年第1期91-96,共6页Jiangsu Communication

摘　　要：随着网络云化进程的快速发展,通信组网架构越来越复杂,传统基于固化经验的网络运维方式弊端逐渐凸显。如何提高面向海量告警的网络运维效率,缩短故障处理时长,成为当下研究的热点。本文通过解析云化资源池日常监控运维工作的流程及痛点,结合资源池空间资源、故障传播模型(告警发生时间、告警定位字段)等关键信息对资源池故障产生的海量告警进行汇聚,采用特征提取和建模分析方法,基于贝叶斯疑似度的启发式故障定位算法,提出了基于空间资源的网络云故障根因智能推荐方案。经过资源池故障实际案例的实践验证,该方案在日常运维过程中能够快速汇聚故障关联告警,定位故障根因,大幅缩短业务恢复的时长,有效提高网络云运维效率。With the rapid development of network cloudification,the communication network architecture is becoming increasingly complex.The drawbacks of traditional network operation and maintenance methods based onfixed experience are becoming prominent.How to improve the efficiency of network operation and maintenance for massive alarms and shorten the time of fault handling has become a hot research topic.This article analyzes pain points in daily monitoring and operation of cloud-based resource pools,and combines key information such as spatial resources and fault propagation models(alarm occurrence time,alarm location fields)to aggregate the massive alarms.By using feature extraction,modeling analysis method and fault localization algorithm based on Bayesian suspicion degree,a network cloud fault root cause intelligent recommendation scheme is proposed.Through practical verification with resource pool failure cases,this solution can quickly aggregate fault-related alarms,locate the root cause of faults,significantly shorten the time for business recovery,and effectively improve the efficiency of network operation and maintenance.

关 键 词：云化资源池 故障定位 告警聚类 贝叶斯疑似度 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术] TP391.3[自动化与计算机技术—计算机科学与技术]