| Summary: | 碩士 === 國立花蓮教育大學 === 學習科技研究所 === 95 === Recent advances in communication and computer technology have made transmission rate of the Internet faster and faster. Next-generation networks will support transmission rate that are orders of magnitude higher than current rates. Because of the deployment in the Internet, the explosive increase in commercial usage of the Internet has resulted in a rapid growth in demand for video deliver technologies. For examples, on-demand home entertainment gives customers the freedom to view movies from remote sites at any time in their comfort own home, distance learning provides opportunities for students to take courses taught at remote locations according to their individual needs and time constraints, and digital video library lets remote users research and view videos from a large video library.
In this thesis, we incorporate network-aware concept into our work and attempt to adjust suitable settings under different network environments so that the throughput for video-on-demand (VoD) can be predicted and is acceptable in heterogeneous peer-to-peer (P2P) networks. The issue of network status detection and prediction is mainly addressed in this work to satisfy user’s need. A solution for video on demand service in heterogeneous networks that integrates wireless and wirelined networks by using the network aware concepts, including the integration of a heuristic multicast algorithm (DDVMA) and service oriented architecture (SOA), is proposed. Meanwhile, a peer-to-peer hot VoD (P2HVoD) architecture is presented to efficiently use client’s cache. The experimental results showed that the defection probability, average latency and delay jitter variation were effectively decreased in the proposed architecture and the feasiblity and practicability of the proposed work is thus verified.
|
|---|
