Implementation and performance analysis of collaborative peer-to-peer video-on-demand systems

Abstract

Video-on-Demand (VoD) systems are of high popularity and responsible for generating large volumes of traffic in nowadays communication networks. The infrastructure of the VoD system providers needs to keep up with the increasing number of its user base and therefore it is important to find scalable and smart solutions that can adapt while maintaining a satisfactory performance. Basing VoD systems on peer-to-peer (P2P) technologies could distribute the burden of a network while optimizing, or at least maintaining, the performance during video playbacks. Content replicated and distributed by active users avoids the need for a huge number of dedicated servers. The usage of P2P for VoD also inherits some policies such as peer selection policy or piece selection policy that could be leveraged in a collaborative manner by entities such as Internet System Providers (ISP). ISPs usually try to safeguard networks with traffic shaping or load balancing. Bringing a collaborative interface could allow the ISPs to inform third party applications of certain overlay networks that should be respected, allowing the protection of some links or nodes from certain traffic. Achieving symbiosis between a P2P VoD service and ISPs may bring new opportunities to improve end user experience while guaranteeing the ISP best interests. In this context, this project aims to develop a P2P VoD prototype that allows to study the effectiveness of commonly used peer, piece, and replicas management/selection strategies, as well as the ability to implement new strategies. Additionally, it integrates the development of complementary P2P VoD systems monitoring strategies and collaborative mechanisms between the P2P VoD system and ISPs.