Request-peer selection for load-balancing in P2P live streaming systems

Abstract

Unlike peer-to-peer (P2P) file sharing, P2P live streaming systems have to meet real-time playback constraints, which makes it very challenging yet crucial to maximize the peer uplink bandwidth utilization so as to deliver content pieces in time. In general, this is achieved by adopting tailor-made piece selection and request-peer selection algorithms. The design philosophy is to regulate the network traffic and to balance the load among peers. In this paper, we propose a new request-peer selection algorithm. In particular, a peer in the network estimates the service response time (SRT) between itself and each neighboring peer. An SRT is measured from when a data piece request is sent until the requested piece arrives. When a peer makes a piece request, the neighbor with smaller SRT and fewer data pieces would be favored among potential providers. This is because smaller SRT implies excess serving capacity and fewer data pieces suggests less piece requests received. We evaluate the performance of our request-peer selection algorithm through extensive packet level simulations. Our simulation results show that the traffic load in the network is better balanced in the sense that the difference of the normalized number of data packets uploaded by each peer is getting smaller and the number of repeated piece requests generated by each peer (due to request failure) is significantly reduced. We also found that the load of streaming server is reduced, and the overall quality of service, measured by playback continuity, startup delay etc, is improved as well.