基于DPI技术的App应用协议分析系统的设计与实现

发布时间：2018-04-16 20:22

本文选题：DPI技术 + 推荐系统　；参考：《北京交通大学》2017年硕士论文

【摘要】：在互联网飞速发展的今天,数据业务急剧增长。随着信息技术的不断发展,各种新的应用层出不穷,但由于缺乏有效的技术手段,许多新的应用程序无法被感知和细化管理,导致网络运营十分困难。比如流量检测,传统应用用户使用热度的减退等一些问题。DPI(Deep Packet Inspection)技术的出现,为移动互联网运营商带来了曙光。本文设计并实现了 App应用协议分析系统,为IT企业内部的业务运营人员提供了 App应用识别、应用规则库的生成、应用规则机器入库和测试、App应用推荐和深度识别库等功能。本人通过统一 DPI技术分析App应用,充分了解运营人员的需求,提炼出了该系统的开发需求。该系统主要用于管理和维护使用DPI技术提取的应用规则,能够定期发布不同版本的规则库、测试规则库、加密和部署规则库,并通过规则库识别流量信息,通过对流量数据的处理,向用户推荐App应用。系统主要功能包括App应用分析、规则库生成、机器规则入库和检测、App应用推荐、规则库加密、深度识别库等功能模块。作者独自设计并开发实现了如下六个模块:(1)App应用分析与管理模块:包括App应用管理、应用组管理、应用搜索、应用协议分析等功能,重点实现了不同版本应用组、应用管理和应用规则分析。(2)规则创建和管理模块:包括应用规则管理、展示、纠错及查询功能,重点实现了 basic、http、ip三种应用识别规则的管理和xml规则预览。(3)机器规则入库和统计模块:对机器提取的应用规则入库、入库规则审核、应用规则信息统计等功能,重点实现了机器规则入库和审核。(4)规则库生成和测试模块:包括对三种规则的xml文件生成、规则测试、规则库信息管理等功能,重点实现了三种规则的xml文件生成。(5)App应用推荐模块:包括用户话单数据清洗、协同过滤算法实现、推荐算法优化等功能,重点实现了 App应用推荐。(6)规则库加密与深度识别模块:包括规则库加密解密、应用管理、应用内容展示、应用重点动作管理等。该系统的功能测试结果表明,App应用协议分析系统满足了对于常见手机应用规则管理、规则库管理、App应用推荐等功能要求,并通过分析系统运行时用户操作的实际情况,优化了规则库生成过程中对应用的动态管理。目前,该系统暂处于内部测试与试运行阶段,正式上线后,系统将为公司节省更大的人力成本并提高大数据方面的业务收入。
[Abstract]:In the rapid development of the Internet today, the rapid growth of data services.With the continuous development of information technology, various new applications emerge in endlessly, but due to the lack of effective technical means, many new applications can not be perceived and refined management, resulting in network operation is very difficult.For example, some problems such as traffic detection, the decline in the heat used by traditional users, and so on. The emergence of DPII Deep Packet Inspection technology has brought the dawn to mobile Internet operators.This paper designs and implements a App application protocol analysis system, which provides App application identification, application rule base generation, application rule machine input, application recommendation and depth recognition library for the business operators in IT enterprises.Through unified DPI technology to analyze App applications, I fully understand the needs of operators, and extract the development requirements of the system.The system is mainly used to manage and maintain the application rules extracted by using DPI technology. It can regularly publish different versions of rule base, test rule base, encrypt and deploy rule base, and identify traffic information through rule base.The App application is recommended to the user through the processing of the traffic data.The main functions of the system include App application analysis, rule base generation, machine rule entry and application recommendation, rule base encryption, depth recognition library and so on.The author has designed and implemented the following six modules: App application management, application group management, application search, application protocol analysis and so on.Application management and application rule analysis. 2) Rule creation and management module: including application rule management, presentation, error correction and query functions,The functions of three kinds of application recognition rules management and xml rule preview and statistics module are implemented, such as the input of the application rules, the audit of the rules, the statistics of the rules, and so on, and the functions of this module are as follows: (1) the key functions of this paper are as follows: (1) the management of the three kinds of application recognition rules and the previews of the xml rules.The module of generating and testing the machine rule base and auditing rule base is implemented, which includes the functions of xml file generation, rule test, rule base information management and so on, which include three kinds of rules, such as: xml file generation, rule test, rule base information management and so on.This paper focuses on the implementation of three kinds of rules of xml file generation. The application recommendation module includes user data cleaning, collaborative filtering algorithm implementation, recommendation algorithm optimization and other functions, including user data cleaning, collaborative filtering algorithm implementation, recommendation algorithm optimization and so on.The encryption and depth recognition module of App application recommendation. 6) rule base includes rule base encryption and decryption, application management, application content display, application key action management and so on.The function test results of the system show that the application protocol analysis system meets the functional requirements of common mobile phone application rule management, rule base management and application recommendation, and analyzes the actual situation of user operation while the system is running.The dynamic management of the application in the process of rule base generation is optimized.At present, the system is currently in the stage of internal testing and trial operation. After the system is officially launched, the system will save the company more labor costs and increase big data's business income.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.56

【参考文献】