An Aggregation And Acceleration System Of Multi-sourced Streaming Media Data

In recent years, video-sharing websites such as YouTube, Youku, Tudou,are expanding rapidly with the popularization and development of the network. To a successful video websites, there are two important points: colorful video contents and nice user experience. Most video websites only shows their own videos, and the B/S structure will face a enormous pressure with huge amounts of users.To aggregate videos of the Internet and improve quality of playing, we develop the CloudMedia system. This system consists of four components: crawl, website, client and tracker. The crawler and the website are used to aggregate videos from major video websites and show them. The client and the tracker can provide P2P acceleration. Crawl is used to get the video pages'URLs and their information such as videos'titles, tags and introductions. The website shows the video pages. When users visit the video page on the CloudMedia website, the client downloads the video data from date server. The client gets other peers information by tracker. If other users are watching the same video, or have watched it, the client will get the video date from those users. CloudMedia uses different methods to get data from different video websites and provide P2P acceleration.On the CloudMedia website, multi-sourced videos could be browsed. It can help users get resources together. What is more, to make playing more smoothly, client can provide a wide P2P acceleration. Our work also lays the foundation for CloudMedia ads recommending, searching based on users actions and social tagging.CloudMedia achieved all the features of aggregation and acceleration modules. The test result shows that this system could afford excellent service. Compared with the existed video website's accelerator,it increases by 30% at P2P acceleration effect.