摘要
网络异常会降低用户体验,而有效的异常识别和故障的根因定位,对于快速恢复服务并减轻损失来说至关重要,因此在智能运维领域,对网络异常识别方法的研究和应用将持续成为智能运维领域的热点和重点方向。本文针对无线网络网管指标量大、关联繁杂以及监控资源消耗高等问题,结合PCMCI算法和Tsfresh工具,找出网管指标之间的因果关联关系,并通过对网管指标时序数据的特征检测,识别出网络的异常,实现了无线网络优化平台的网管指标检测自动化、智能化,提升了异常识别的准确率和无线网络优化的响应效率,在识别网络异常情况和智能运维领域有着积极的影响。
Network anomalies can reduce user experience,and eff ective anomaly identifi cation and root cause location of faults are crucial for quickly restoring services and reducing losses.Therefore,in the fi eld of intelligent operation and maintenance,the research and application of network anomaly identification methods will continue to be a hot topic and key direction in the fi eld of intelligent operation and maintenance.In view of the large number of wireless network management indicators,complex associations,and high monitoring resource consumption,this paper combines the PCMCI algorithm and the Tsfresh tool to fi nd out the causal relationship between network management indicators and identify network anomalies through feature detection of network management indicator time series data,realizing the automation and intelligence of network management indicator detection of wireless network optimization platform,improving the accuracy of anomaly identification and the response efficiency of wireless network optimization,and has a positive impact on identifying network anomalies and intelligent operation and maintenance.
作者
黄继宁
刘洋
麦锦恩
林彬
HUANG Ji-ning;LIU Yang;MAI Jin-en;LIN Bin(China Mobile Group Guangdong Co.,Ltd.,Guangzhou 510623,China;China Mobile Group DesignInstitute Co.,Ltd.Guangdong Branch,Guangzhou 510623,China)
出处
《电信工程技术与标准化》
2024年第10期1-7,共7页
Telecom Engineering Technics and Standardization
