Dynamic asynchronous buffer management to improve data continuity in P2P live streaming

In data-driven mesh-based P2P live streaming system, the playback buffer management affects the audiences' quality of experience (QoE) significantly. Due to the instability and heterogeneity of P2P networks, the buffer is not always ready for playback. It is not only associated with the media timestamp, but also with the data availability in its neighbors. So how to manage the buffer synchronization position is very critical for provision of playback continuity. In this paper, we introduce a Buffer Map Matrix (BMM) data structure to describe the chunks availability of each peer and its neighbors. Furthermore we propose a dynamic asynchronous buffer management algorithm, MaxCount, in which the peer synchronizes its buffer window to the position where the maximum amount of chunk could be received. The position is predicted according to the BMM table and neighbors' upload capability. The window shifts to the position where the max amount of chunks in the buffer is available in the neighbors. We carried out extensive experiments with NS-2 simulation. The experiment results show that the MaxCount algorithm has better performances than existing approaches in playback continuity and adaptability in network churn, especially when the aggregated upload bandwidth of the peers is limited.