P2P streaming with layered video

Peer-to-peer (P2P) video streaming has become a very popular application in the Internet to distribute video content to a large number of Internet users. However, critical problems remain open in system design for P2P video streaming. The main objective of this thesis is to enhance P2P video streaming by using advanced video coding techniques, mainly layered video.For P2P live video streaming, we investigate how to use layered video in chunk-based and substream-based mesh-pull P2P networks. Two distributed protocols, chunk trading and substream trading, are designed to provide built-in incentives and demand/resource adaptation in P2P live streaming. In chunk trading, when the system has sufficient resource supply, every peer will receive excellent video quality at the full video rate when the system's resource supply cannot support the full video rate for every peer, the received video rate/quality of a peer is commensurate with its upload contribution to the system. This provides incentives for peers to re-distribute and discourages free-riders. In this manner, the system also adapts the system demand to the system supply. Additionally, the system can guarantee a certain QoS for the received video. Substream trading refines chunk trading with a reduced transmission delay and an improved system throughput. Both systems are evaluated by trace-driven simulations. A prototype of chunk trading has been implemented and evaluated in the Internet.We investigate how to apply multi-stream video coding, including multiple description coding (MDC) and layered coding, on P2P video-on-demand. Conventional MDC simplifies the system design by producing equally important descriptions, but it has a lower coding efficiency and bandwidth utilization. Conventional layered coding has a higher video coding efficiency, but it needs sophisticated layer caching and scheduling schemes to provide unequal protection to different layers. We propose a multi-stream coding and transmission scheme, Redundancy Free Multiple Description (RFMD) Coding and Transmission, based on layered coding. Like MDC, with RFMD, all substreams have equal importance. Thus, video quality gracefully degrades as substreams are lost, independently of which particular substreams are lost. Like layered coding, only the source bits are collectively transmitted by the supplying peers, increasing the bandwidth utilization in the system. Compared with conventional layered coding and MD-FEC, RFMD achieves the maximum system performance in terms of the received video quality, with the minimum consumed storage in peers.Synchronous trading can be used to provide incentives for P2P file sharing (e.g., tit-for-tat in BitTorrent) and P2P live video streaming (e.g., chunk/substream trading). However, synchronous trading becomes less appropriate if the peers do not have the common interest, such as in P2P VoD. We propose a networked asynchronous bilateral trading (NABT) framework to provide incentives for general P2P applications. NABT exploits the built-in reputation in social network. We show that the distributed NABT strategy can potentially achieve a high trading efficiency that is comparable with a centralized global currency system.