Share-E: Exploiting peer-to-peer streaming in periodic broadcasting of videos for heterogeneous clients

Abstract

There has been much work on multicast approaches for scalable systems. Periodic broadcasting (PB) is the most effective approach to building large-scale video-streaming services. But, one weakness of PB is the assumption of a homogeneous environment where all clients have the same resources such as reception bandwidth and storage. This assumption about clients is not reasonable in practical situations. In this paper, we propose a PB technique that allows low-bandwidth clients to be serviced equally with high-bandwidth clients in terms of service latencies by using peer-to-peer streaming between clients. In the proposed system, the server builds multicast trees of clients so that low-bandwidth clients are serviced from other clients with sufficient upload bandwidth instead of from the server. Our proposed system can provide all clients with equal QoS such as video quality and service latency, regardless of their resource capability. Using extensive simulation, we have demonstrated our system has better performance than existing PB protocols.