NAP: An Agent-Based Scheme on Reducing Churn-Induced Delays for P2P Live Streaming

Peer-to-peer (P2P) multimedia streaming provides a scalable solution for IPTV. However, delays from channel switch and streaming recovery are typically in the scale of 10-60 seconds, which have hindered the extensive commercial deployment of P2P systems. We call these two types of delays, churn-induced delays. Obtaining assurances on churn-induced delays in dynamic and heterogeneous network environments is a challenge. In this paper, we devise a simple, yet efficient agent-based P2P streaming scheme, called NAP, which reduces churn-induced delays. We first formulate the problems of minimizing channel-switching delay and streaming recovery delay. We then present the detailed methodology of NAP. In addition, we develop a queuing model for the P2P streaming scenario and analyze the properties of NAP based on this model. Our numerical study reveals the effectiveness of NAP, and shows that NAP significantly reduces churn-induced delays, especially channel-switching delays.