Streaming Video Data Monitoring And Analysis On Network Streaming Media Technology

With the rapid development of internet application and web transmission technology, video streaming media service with high data rate as its typical characteristic has adopted advanced CDN and P2P technology, which resolved the problem of transmission bandwidth bottlenecks. However, with the spread of massive video data on the network, the management of video contents and copyright gets into trouble show in thress aspects:firstly, for a great variety of condensed data formats and mixed standards, it’s hard for the management to decode video data; secondly, most protocols of data encapsulation and transmission use proprietary standards, that make data traffic identification more difficult; thirdly, these protocols use mixed standards and change irregularly. Hence different protocols need different methods of traffic identification and the methods need alterations irregularly. All of these make it harder for the management to automatically extract and rebuild video contents on the network. In order to settle this problem, the thesis studies the technologies of monitoring, extraction and rebuilding for video data packets.This thesis takes the process of H.264 video data transmission in C/S model as research subject, proposes and implements an algorithm of RTP packets recognition which encapsulates H.264 video data. On that basis, it implements a rebuilding algorithm which can extract, store and combine H.264 video data. Meanwhile, the thesis makes a deep analysis on PPLive in P2P model, and gets the characteristic of PPLive video data transmission and the proprietary packed formats of UDP and TCP which have video data. Based on these characteristics, the thesis proposes and implements PPLive video packets of the Gateway recognition and rebuilding algorithms. By using these algorithms, the thesis also proposes and implements a video data monitoring and rebuilding system in C/S and P2P model separately. The experimental results show that the RTP video data monitoring system in C/S model can recognize RTP packets well which have H.264 video data, rebuild original video accurately and play back it at the same time; The PPLive video monitoring system in P2P model can support for multi-users data monitoring and playback, and the average accuracy rate of video rebuilding is about 86%. That proves the feasibility and practicability of algorithms proposed by the thesis which can provide automatic management of video contents and copyrights with efficient video contents.