一种基于DPI的CMNet城域网用户模型流量预测方法  

A DPI based Traffic Prediction Method of CMNet Man User Model

作　　者：张忠杰 ZHANG Zhong-jie(Jilin Jida Communication Design Institute Co.,Ltd.,Changchun 130012,China)

机构地区：[1]吉林吉大通信设计院股份有限公司,吉林长春130012

出　　处：《通信电源技术》2020年第8期110-112,共3页Telecom Power Technology

摘　　要：DPI(Deep Packet Inspection)是一种基于数据包的深度检测技术,针对不同的网络应用层载荷(如HTTP、DNS等)进行深度检测。文中对DPI系统进行简要概述,分析了基于DPI系统在确定业务流量预测模型相关指标参数方面的优势和准确性。借助DPI手段,对CMNet城域网出口节点进行定量分析,可准确区分业务、用户类型及业务流向,为流量预测模型的建立提供基础数据,增加业务预测的准确性,避免流量预测偏离实际值过大或过小对网络建设规模及投资产生的影响。DPI(Deep Packet Inspection)is a kind of deep detection technology based on data packets,which is used to detect different network application layer loads(such as HTTP,DNS,etc.).This paper gives a brief overview of DPI system,and analyzes the advantages and accuracy of DPI system in determining the relevant index parameters of business flow prediction model.With the help of DPI,the quantitative analysis of CMNet man export nodes can accurately distinguish services,user types and business flows,provide basic data for the establishment of traffic prediction model,increase the accuracy of business prediction,and avoid the impact of too large or too small deviation of traffic prediction from the actual value on network construction scale and investment.

关 键 词：DPI 流量预测 CMNET 

分 类 号：TP393.1[自动化与计算机技术—计算机应用技术]