基于移动数据业务的DPI系统的研究及应用

发布时间：2018-05-31 17:43

本文选题：数据业务 + 流量经营　；参考：《北京邮电大学》2014年硕士论文

【摘要】：随着移动互联网的普及,电信运营商正面临着前所未有的机遇与挑战。作为移动数据业务的提供者,运营商不仅要通过建设高质量的通信网络来保障用户体验,更需要通过业务发展来完成企业经营模式的转型。但是,由于受到以微信为代表的OTT (Over The Top)业务的影响,运营商在加大网络建设投入的同时,正面临数据流量“增量不增收”的窘境,业务经营入不敷出。目前,业内机构普遍认为,流量经营是帮助运营商打破经营困局的有效方法。本文从流量经营的需求出发,针对移动数据业务设计并实现了一种深度包检测(Deep Packet Inspection, DPI)系统。通过对即时通信、图像影音等移动应用的分析,系统建立了DPI特征库和应用知识库,从而可对移动数据流量进行识别和信息挖掘。该系统的实现,为基于移动应用的流量经营提供了基础。论文首先通过统计数据分析了移动数据业务经营需求,并研究了DPI技术的应用场景。在此基础上,从功能定义、部署方案和系统架构三个方面对基于移动数据业务的DPI系统进行总体设计。具体地,系统由数据采集、数据处理、数据挖掘和信息展示四个主要的功能模块组成。通过移动应用的分析,系统定义并建立了DPI特征库和应用知识库,并以此为依据,通过Python语言实现了数据流量的处理,包括流量来源标识、用户行为信息挖掘和结果展示。目前,论文所实现的系统已经得到了实际的应用,为电信运营商实现流量经营提供了有效的参考。
[Abstract]:With the popularity of mobile Internet, telecom operators are facing unprecedented opportunities and challenges. As a provider of mobile data service, operators not only need to build high-quality communication network to protect the user experience, but also need to complete the transformation of business model through business development. However, due to the influence of the OTT over The Top-service represented by WeChat, the operators are facing the dilemma of "no incremental increase in data flow" while increasing investment in network construction. At present, industry organizations generally believe that traffic management is to help operators break through the operational difficulties of an effective method. Based on the requirement of traffic management, this paper designs and implements a deep packet detection Packet Inspection, DPI) system for mobile data services. Based on the analysis of mobile applications such as instant messaging, image audio and video, the DPI signature database and application knowledge base are established, so that mobile data flow can be recognized and mined. The realization of the system provides the foundation for the traffic management based on mobile application. Firstly, the paper analyzes the business demand of mobile data service through statistical data, and studies the application scenario of DPI technology. On this basis, the DPI system based on mobile data service is designed from three aspects: function definition, deployment scheme and system architecture. Specifically, the system consists of four main functional modules: data acquisition, data processing, data mining and information display. Based on the analysis of mobile applications, the DPI signature database and application knowledge base are defined and established. Based on this, the data flow processing is realized by Python language, including traffic source identification, user behavior information mining and result display. At present, the system implemented in this paper has been applied in practice, which provides an effective reference for telecom operators to realize traffic management.
【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】：TN929.5;TP393.01

【参考文献】