QoS - LiquidStream: Scalable monitoring and bandwidth control in P2P live streaming

There is a lot of research work that has recently focused on P2P live streaming. As a result a lot of algorithms and system architectures have been proposed that achieve high upload bandwidth utilization of the participating peers, low delays during stream diffusion while systems are stable under dynamic network conditions and peer behavior. However, these systems make the assumption that the average upload bandwidth capability of the participating peers is stable and greater than the bandwidth the system requires for the successful delivery of the video stream, an assumption that makes these system impractical. This paper relaxes this assumption by proposing a monitoring mechanism for the total upload bandwidth capabilities of the participating peers, which is scalable, accurate, dynamic and with low overhead. Exploiting this monitoring mechanism, we also present and evaluate a method that allows the accurate and timely estimation of the required (minimal) amount of bandwidth that an external set of resources has to contribute in order to maintain uninterrupted stream delivery at the specified quality, thus making P2P systems practical for use.