Coping with flash crowd in multi-channel live video streaming for peer-to-peer networks

Abstract

The current multi-channel P2P video streaming architectures incur performance problems including: (i) A low Quality of Service (QoS) in unpopular channels with few viewers; (ii) The sudden drastic increase in the number of requests for a video at the video release time which is referred to as the flash crowd phenomenon. The flash crowd phenomenon in live streaming poses significant challenges in system design. In a P2P live streaming system, a newcomer expects to watch a live video immediately. This paper presents a novel framework for multichannel P2P live video streaming that provides de-centralized mechanisms for handling flash crowds that includes incentive mechanism, load balancing mechanisms, and cross-channel help among the peers for live video streaming in multi-channel P2P systems. Our simulation results demonstrate that the quality of unpopular channels is improved. Moreover, for flash-crowds the proposed method improves the quality of video by reducing the playback delay, distortion, and reducing the redundant traffic.