网络流量分类方法研究及流量特征分析

发布时间：2018-05-29 23:42

本文选题：网络流量分类 + 分类器评价　；参考：《北京邮电大学》2014年硕士论文

【摘要】：近几年,互联网基础建设和内容服务均快速发展,随之带动了互联网宽带业务的发展。据我国互联网信息中心CNNIC测算,截至2013年6月底,我国网民规模达5.91亿,互联网普及率为44.1%。产生网络流量的根本原因是互联网中的网络应用。实现网络流量的正确识别并分析网络流量的特征,是我们深入理解网络状况,用户行为,互联网现状的前提条件,无论是对网络管理员还是对用户和服务提供商,都具有重要的意义。网络流量分类技术,是分析网络流量特征和增强网络可控性的基本手段之一,广泛运用于流量监控、网络安全检测、用户行为分析、计费管理或其他网络活动中。在对固定互联网的特征进行分析研究的同时,随着移动互联网的迅猛发展,对移动互联网流量的特征分析也显得尤为重要。不仅用户数量激增,同时随着线速越来越高,网络流量越来越大,在线的商用设备每天都能产生TB级甚至更庞大的流量数据。网络流量已呈现大数据特征,对海量数据的存储和分析已成为网络流量特征分析的重要研究内容。为了方便对网络流量进行分析,我们开发了网络流量分类与分类系统(Traffic Analysis and Classification System, TACS)以及海量数据分析系统LogAnalyser。本文首先介绍网络流量分类及分析方法,在简要介绍TACS系统之后,对网络流量分类中分类器评价标准进行了深入研究。在此基础上,本文对真实的网络数据进行了基于DFI(Deep Flow Inspection)的流量分类并对已有分类器评价标准进行了改进。针对海量网络流量数据的分析,整体介绍了自主开发的分析系统。最后,依据TACS和LogAnalyser,本文对固定互联网和移动互联网的网络流量数据进行了不同角度的流量特征分析。
[Abstract]:In recent years, Internet infrastructure and content services have developed rapidly, which has led to the development of Internet broadband services. According to CNNIC, China's Internet Information Center, by the end of June, 2013, the number of Internet users in China reached 591 million, and the Internet penetration rate was 44.1%. The fundamental reason for the network traffic is the network application in the Internet. Realizing the correct identification of network traffic and analyzing the characteristics of network traffic is a prerequisite for us to deeply understand the network situation, user behavior, and the current situation of the Internet, whether for network administrators or users and service providers. Are of great significance. Network traffic classification technology is one of the basic methods to analyze network traffic characteristics and enhance network controllability. It is widely used in traffic monitoring, network security detection, user behavior analysis, billing management and other network activities. At the same time, with the rapid development of mobile Internet, it is particularly important to analyze the characteristics of mobile Internet traffic. Not only does the number of users surge, but also as the speed of the line increases and the network traffic becomes larger and larger, online commercial devices can generate terabytes or even larger traffic data every day. Network traffic has been characterized by big data, and the storage and analysis of massive data has become an important research content of network traffic feature analysis. In order to facilitate the analysis of network traffic, we have developed Traffic Analysis and Classification System, TACS), a network traffic classification and classification system, and LogAnalyser. a massive data analysis system. This paper first introduces the classification and analysis method of network traffic. After a brief introduction of TACS system, the evaluation standard of classifier in network traffic classification is studied deeply. On this basis, this paper classifies the real network data based on DFI(Deep Flow inspection and improves the evaluation criteria of existing classifiers. Based on the analysis of massive network traffic data, this paper introduces the analysis system developed by ourselves. Finally, according to TACS and Log Analyser, this paper analyzes the traffic characteristics of fixed and mobile Internet traffic data from different angles.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【参考文献】