MegaDrop: A Cooperative Video-on-Demand System in a Peer-to-Peer Environment

In a Peer-to- Peer(P2P) environm ent, which can im prove scalability and lessen bandwidth waste, the load of the Video- on-demand (VoD) services can be distributed to reduce the burden on centralized video servers and maximize the bandwidth utilization. Conventional P2P techniques for VoD services only consider data between active peers in the same VoD session. The inactive peersthat leave the session but still hold some of the media content in their local storage are ignored. In this article, we design and im plement a fully decentralized VoD service via P2P techniques, which is ref erred to as the MegaDrop system. The MegaDrop system not only takes active peersinto consideration but also provides mechanisms for discovering inactive peersthat contain desired m ediaobjects. Evaluation results of the M egaDrop system show that our architecture performs more efficiently when the more inactive peersinvolving to provide media blocks.