Research For Application Of Distributed VoD System In Distance Education

Video-on-Demand (VoD) system is an important application of streaming media in distance education that supplies user with a real-time video-on-demand through network with the advantage of breaking the limit of time and space to improve the efficiency and quality of modern education. However, with the rapid development of VoD system, the problems of I/O bandwidth in video server and network transmission are getting more and more obvious with concurrent visit of a large amount of user, which is an obstacle to large-scale application of VoD in distance education, as well as the expansion of the scale is a great challenge to techniques of architecture, storage and service policy of video server. Our research deploys architecture and service policy of video server. The major contribution of this thesis includes:Related techniques and current state are argued. On the basis of deep research into the basic technology, application characteristic and current situation of VoD system, the critical technology in video server and application system is analyzed, and the demand and problem of application in distance education is argued while the critical way to solve the problem of the VoD system is proposed such as improvement in system architecture and scheduling policy.An extensible distributed multilayer architecture is designed. Architecture in large-scale VoD system is one factor of influencing the system performance. Adopting distributed architecture is a good way to improve performance and efficiency and satisfy requests of a mass of concurrent user. The advantage and disadvantage of different architecture are compared, then the abstract model is presented and the proposed multilayer structure and the VoD flow are described in detail.A packet-based period patching scheduling algorithm (PBP~2) is proposed. Static and dynamic stream scheduling algorithms are argued, patching stream scheduling algorithm and period patching stream scheduling algorithm are discussed, then an advanced algorithm based on batching whose idea is increasing waiting time of user reduces the consumption of bandwidth is proposed, and the pseudo-code of the advanced algorithm is described and analysed.The simulation of PBP~2 is analyzed and tested. The results illustrate that PBP~2 can reduce bandwidth consumption, and the amount of patching stream on each sharing stream period is reduced compared to the period patching stream algorithm under the same frequency of program request and request arrival and the same useful bandwidth.