Streaming video using cooperative networking

With the increase of bandwidth in computer networks, video streaming is expected to become a common practice in various networking environments. The Internet was originally designed to provide best-effort service to transport data without any guarantee on the end-to-end performance. It does not make any promise about the variation of the packet delay jitter, bandwidth, and packet loss, either. It is also more challenging for supporting video streaming services in wireless networks that constantly suffer from quality degradation caused by interference, channel fading, mobility, etc.;The main objective of this dissertation is to improve the overall video streaming performance in various networking environments, such as IP-multicast in wired network and wireless mesh networks (WMNs), using cooperation among participants including clients and routers. We investigate a number of key challenging issues associated with video streaming, i.e. the reliability issue in IP-multicast, the high throughput routing metric design issue, the server/peer selection issue, the admission control issue and peer cooperation issue in WMNs. We explore solutions to the above issues using a cooperative networking approach, which includes constructing overlay Peer-to-Peer (P2P) retransmission networks and exploring hybrid architecture of content distribution networks (CDN) and P2P networks.;The first part of this dissertation addresses the reliability-related issues in IP-multicast. We propose a novel overlay P2P retransmission architecture to exploit path diversity. An approach that leverages both disjoint path finding and periodic selective probing to take into account peer's recent packet loss probability, retransmission delay and recent retransmission performance is proposed to effectively construct an efficient and dynamic overlay peer retransmission network.;The rest of this dissertation focuses on addressing some important challenges in video streaming over WMN. We first investigate high throughput routing metric for video streaming over WMNs. We propose a radio and bandwidth aware routing metric for video streaming for WMNs. Our experimental results using a testbed show that the enhanced routing protocol has better performance than the traditional minimal hop count based routing protocol in terms of packet loss and network jitter.;Next we explore a Unified Peer-to-Peer And Cache framework (UPAC) for high quality video on demand services over infrastructure multi-hop WMNs. UPAC is a hybrid approach combining the CDN and P2P approach.;The research is divided into the following steps: (1) We first study the general cooperative mesh content server selection problems in WMNs. A main feature of the UPAC architecture is to deploy mesh content servers in the WMNs so as to balance the load of the network and improve the content availability. We propose a routing metric based cross-layer server selection scheme. Simulation results demonstrate that the proposed scheme performs significantly better than other common server selection schemes used in CDN. (2) We then propose a novel admission control algorithm with per-flow routing in WMNs to select the mesh content server and transmission path. We formulate the admission control problem with the interference constraint and investigate the optimal solution using a centralized algorithm. Because the optimal solution is NP-hard, we propose a heuristics approach with admission control and per-flow routing. We demonstrate that the proposed heuristics solution can achieve better overall video streaming quality than the approaches without admission control. (3) Finally, a BitTorrent-like P2P approach for UPAC is proposed to further explore the peer/server cooperation in WMNs. In order to meet the stringent requirements on delay and packet priority for video streaming in WMNs, we propose to use routing layer information to select peers and manages peer/data list using an urgent data first strategy combined with the rarest first strategy. Simulation results show that, with the proposed P2P cooperation support, the network load is further balanced, and the network capacity can be utilized much more efficiently.