Advances in computer network, digital video compression and storage technologies have made video-on-demand (VOD) services a reality. Using this service, people could get the video he needs quickly and expediently, however, high bandwidth and real-time features of video data require video server and network with high performance. Too many users may cause bottleneck problem in server I/O bandwidth or network bandwidth, and it is proved to be a major limiting factor in the widespread usage of VOD service.To resolve the shortcomings of the traditional VOD system and integrate the distance education, the paper considers two aspects of modern VOD: channel allocation scheme and distribute system. In the dissertation, a theoretical model has been put forward with full consideration of human factor and technological background of on-demand from the practice in VOD application. Some important techniques such as resource scheduling, video storage and distribute are deeply researched and several new ideas have been presented as well as some innovations.The paper puts forward a distribute VOD system model that suit for distance education, and discuss some aspects of the model:1) Optimized Storage Scheme. A parameterσis used to test the scheme rational or not.2) Video Distributing Scheme based on a binary tree. The paper presents arithmetic to build the binary tree.3) Integrate unicast and multicast to form a mixed transmission scheme. The paper makes use of the different request probability of the videos to resolve the abuses of traditional channel scheduling scheme. MTS is proved that it has little client wait time and precedes the traditional scheme by simulation. The distribute VOD system model that the paper presented can satisfy the distance education very well, and it resolved the abuse of traditional VOD system. |