基于DPI的流媒体流量监控系统的研究与实现

发布时间：2018-04-29 11:33

本文选题：P2P + 流媒体　；参考：《江西理工大学》2014年硕士论文

【摘要】：作为近几年来互联网最热门的技术，P2P（Peer to Peer）流媒体得到了全面的发展，与文件共享系统相比，，流媒体系统的边下载边播放特点使其日益流行。但是在流媒体普遍应用的同时也出现了许多负面的影响如耗费网络带宽、不良信息传播、知识产权、信息安全等方面的问题。因此，对流媒体应用软件及网站流量进行识别与进一步管理变得尤为重要。为了能够有效的识别和控制网络中流媒体应用流量，防止网络用户过度使用流媒体应用造成网络资源匮乏等一系列问题，本文提出一种基于DPI的流媒体流量监控系统，系统可有效实现对局域网中流媒体应用的实时识别与控制，有效的解决上述问题，协助网络管理员更加安全高效的管理网络。针对以往流量监控系统流量识别无法具体到应用、采用的匹配算法效率较低、流量的限流与阻断效果不够明显等缺陷，本文在对几种基本单模式和多模式匹配算法研究的基础上，创新性的提出一种AC-BM模式匹配算法，使其既有多模式串匹配的特点又可应用BM算法中跳跃移动的思想，并将其应用于流量监控系统的识别模块，结合DPI技术对相关流量进行深度解析，极大提高了系统对流媒体流量识别的效率；与此同时，本文在流量控制模块采用一种新型的直路串接与旁路干扰相结合的控制方式，对比以往系统仅采用单一控制方法，此种控制方式的优势在于模块中部署的直路控制设备，可对一些较容易识别的流媒体应用和一些较难构造干扰报文包的协议类型直接进行处理，使得旁路干扰部分只需对那些较为复杂且能够构造干扰报文包的流媒体流量进行处理，这样使得流量控制模块的控制效率极大提升。设计完系统各个模块后，文中最后对流量监控系统的整体功能与性能进行了全面的测试，通过验证，本文系统可对近百种流媒体应用精确识别，具有较低的漏识别、误识别率，下发相关控制策略后，系统可有效对流媒体应用流进行阻断、限流操作，测试结果表明整个流控系统的识别准确性、控制效果均能满足设计的要求。目前，本文系统仅能对较为常用的近百种种流媒体应用进行识别与控制，随着网络规模的不断增大，流媒体技术的不断发展，流媒体应用必然会层出不穷，但随着对流媒体流量的不断研究与分析，系统的功能不断完善，流量监控系统必然能实现对所有流媒体流量进行识别与控制，从而更加广泛的应用于网络管理领域。
[Abstract]:As the most popular Internet technology in recent years, P2PU Peer to Peer streaming media has been fully developed. Compared with file-sharing system, streaming media system is becoming more and more popular because of the characteristics of streaming media system while downloading and playing. However, there are many negative effects such as consuming network bandwidth, bad information dissemination, intellectual property, information security and so on. Therefore, it is very important to identify and further manage streaming media application software and website traffic. In order to effectively identify and control the streaming media application traffic in the network and prevent network users from overusing streaming media applications, this paper proposes a streaming media flow monitoring system based on DPI. The system can effectively realize the real-time identification and control of streaming media applications in LAN, solve the above problems effectively, and help network administrators to manage the network more safely and efficiently. In the past, the flow identification of the flow monitoring system can not be applied, the efficiency of the matching algorithm is low, the current limit and blocking effect of the flow is not obvious enough, and so on. In this paper, based on the research of several basic single pattern and multi-pattern matching algorithms, an innovative AC-BM pattern matching algorithm is proposed, which not only has the characteristics of multi-pattern string matching, but also applies the idea of jumping movement in BM algorithm. And it is applied to the identification module of the flow monitoring system, combined with the DPI technology to analyze the depth of the related traffic, which greatly improves the efficiency of the system to identify the streaming media traffic, at the same time, In this paper, the flow control module adopts a new type of control mode, which combines straight circuit serial connection with bypass interference, and only uses a single control method compared with the previous system. The advantage of this control mode lies in the direct control equipment deployed in the module. Some streaming media applications which are easy to identify and some protocol types which are difficult to construct interference packets can be processed directly, which makes the bypass interference only need to deal with the streaming media traffic which is more complex and can construct interference packets. In this way, the control efficiency of the flow control module is greatly improved. After the design of each module of the system, the overall function and performance of the flow monitoring system are tested in the end. Through the verification, the system can accurately identify nearly 100 kinds of streaming media applications, and has low leakage recognition and error recognition rate. After sending off the related control strategy, the system can effectively block the streaming media application flow, limit the current operation, the test results show that the whole flow control system identification accuracy, the control effect can meet the design requirements. At present, this system can only identify and control nearly 100 kinds of streaming media applications. With the increasing of network scale and the continuous development of streaming media technology, streaming media applications are bound to emerge in endlessly. However, with the continuous research and analysis of streaming media traffic, the function of the system is constantly improved, and the flow monitoring system can identify and control all streaming media traffic, which is more widely used in the field of network management.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【参考文献】