Research On The Key Technologies Of The P2P Streaming Media Monitoring

In recent years, with the rapid development of the p2p(peer-to-peer) network technology and multi-media information processing technology, p2p streaming media transaction has become another favorite application of Internet users, followed by p2p file-sharing application. However, while p2p streaming media is providing high-quality media services to the Internet users, it also brings a great deal of problems, among which the security problem is especially noticeable. How to monitor efficiently for large-scale p2p streaming media network is a critical issue to explore in present international p2p technology research area.Due to the protocols of current p2p streaming media systems are all private, there are so limited research on their traffic identification, together with their inherent characteristics such as distributedness, anonymity and dynamicity, thus it is difficult to monitor p2p streaming media efficiently only by traditional passive monitoring way, so there exist many challenging issues that need to be addressed, including how to accurately identify and filter the traffic of p2p streaming media in real time, how to monitor the anomaly of p2p streaming media in real time and so on. These challenges actually constitute the key problems in p2p streaming media monitoring research area. The dissertation thoroughly studies these problems and gives some efficient proposals.First, the dissertation studies the method of p2p streaming media traffic identification. Based on the research of typical p2p streaming media communication mechanism, we extracte the basic feature set of p2p streaming media which includes the important IP address feature, protocol definition feature and data transfer feature, then we propose a new strategy and algorithm to identify the p2p streaming media using the integrated features, finally, we use the experiment to verify its accuracy and confirm the effectiveness of this method. In additions, the dissertation proposes a flexible and muti-stage monitoring model to monitor the traffic of p2p streaming media in local network segment.Second, the dissertation studies the method of p2p streaming media anomaly monitoring from its topology. First of all, with the goal of building p2p streaming media topology fast and accurately, the dissertation proposes two algorithms: BFS-Crawler and DFS-Cralwer, and then take the case of PPLive, we desigen a topology crawler for PPLive network named AT-Crawler, through the topology measurement experiment, we compare and analyze the difference between the above two algorithms, and we also take the topology visualization experiment using collected data. Subsequently, the dissertation defines the anomaly of p2p streaming media concretely, and then proposes a model for monitoring the anomaly of p2p streaming media based on its topology, which can provide technical guidance for the information security of p2p streaming media.Finally, the dissertation is aimed at the design and implementation of prototype system framework for monitoring p2p streaming media, and we propose a framework with the combination of active detection and passive identification, on the basis of systematic analysis for requirements. Furthmore, we analyze and design the functional modules in detail, consequently, it lays the foundation in the aspect of modeling framework for setting up the p2p streaming media monitoring system orient the Internet.The work of this dissertation can provide a good theoretical and pratical foundation for p2p streaming media monitoring in near future.