Incentive-Based On-Demand Video Streaming Using a Dual Spatially-Organized Peer-to-Peer Network
Résumé
On-demand video streaming applications are attracting an ever-growing number of users on the Internet. As new terminals, and especially mobile ones, give their users the ability to film and edit videos, it is necessary to provide efficient and scalable solutions that enable users to share their creations with as many users as they want. Traditional client-server-based video streaming solutions can no longer cope with the ever-increasing number of users as they incur highly expensive bandwidth provision cost on the server. This paper addresses these issues for on-demand video streaming by proposing a dual spatially-organized peer-to-peer network. The idea is to distribute video streaming load by allowing real-time cache sharing between peers that are viewing approximately the same portion of the stream. Compared to existing P2P video streaming solutions, our solution (1) provides higher efficiency by connecting users who are watching almost the same part of the video, thus increasing the chance of finding missing video chunks at neighboring users, (2) reduces inter-ISP video traffic by privileging content exchange among peers within the same ISP network, and (3) provides incentives to share content by providing different service levels to users with different contributions to the system. Performance evaluations show the self-scalability of our solution in face of mass arrivals of users requesting the same video stream.