电信科学 ›› 2014, Vol. 30 ›› Issue (6): 152-156.doi: 10.3969/j.issn.1000-0801.2014.06.025

• 运营技术广角 • 上一篇    

基于大数据分析的网络异常检测方法

曹旭,曹瑞彤   

  1. 中国联合网络通信集团有限公司 北京 100033
    中国联合网络通信集团有限公司 北京100033
  • 出版日期:2014-06-15 发布日期:2014-06-15

Network Anomaly Prediction Method Based on Big Data

Xu Cao,Ruitong Cao   

  1. China United Network Communications Co.,Ltd.,Beijing 100033,China
  • Online:2014-06-15 Published:2014-06-15

摘要:

随着当前网络结构的日益复杂,网络故障出现的概率逐渐增加,网络管理面临更大的挑战。网络管理的一个重要任务是通过对网络设备运行状况的监测,及时发现网络异常,提前进行处置,以降低故障带来的损失。当前业界已经对故障诊断进行了大量研究,但仍然存在告警丢失、虚警严重等问题。为此,提出了一种基于大数据分析的网络性能预测方法,通过改变网络管理中“先报警,再处理”的故障处理流程,实现网络故障的预警功能。本文的研究内容对于提升网络的可用性和可靠性具有一定的意义。

关键词: 网络管理, 故障预测, 大数据, 时间序列分析

Abstract:

With the development of the network technology and expansion of the network size,the probability of network anomaly is greatly increased,which makes network management face greater challenges.An important task of network management is network anomaly detection in time based on the data got by network management system (NMS); the performance of network management will be great improved if the network anomaly could be detected before it really happen.A method of network anomaly prediction based on big data was presented,the traditional process act as “first alarm,then solve”was replaced by anomaly prediction process.The result is significant for improving network availability and reliability.

Key words: network management, anomaly prediction, big data, time series analysis

No Suggested Reading articles found!