| Multicast distribution of video is an important service for many existing and future Internet applications. Today's Internet lacks support for quality of service (QoS) provision, which makes the transmission of real-time video challenging. In addition, the heterogeneity of the Internet's transmission capacities and end-systems makes it extremely difficult to agree on acceptable traffic characterization among multiple receivers of the same video stream.; The thesis presents our work in scalable and adaptive video multicast to meet the above challenges. The objective of our work is to explore the feasibility and techniques for video multicast over large scale heterogeneous networks with end to end control. Such techniques can be applied both in best-effort networks, such as the Internet, and in networks with QoS provision to accommodate variations and inaccuracies in resource reservation.; To address the fairness problem of video multicast in a heterogeneous environment, we explore both layered video multicast and replicated stream video multicast. In layered video multicast, the video source compresses video into several layers and multicasts each layer to a separate multicast group. The receivers dynamically adapt to the appropriate video layers according to their processing power and network conditions. We develop a novel scheme to support Hierarchical Rate Control (HRC), which is incorporated into our Layered Video Multicast with Retransmissions (LVMR) protocol. In replicated stream video multicast, we explore the Destination Set Grouping (DSG) approach in which a video source transmits a small number of video streams carrying the same video at different quality levels. Each receiver chooses the appropriate stream to receive. This scheme is shown to improve fairness significantly at a small bandwidth cost. Our work also proposes solutions to improve the bandwidth efficiency of the DSG protocol.; To further improve the video transmission quality over best-effort networks, we develop retransmission-based error control mechanisms, utilizing smart retransmission and local recovery. A novel scheme, adaptive playback, is propose to dynamically adapt playback point of video frames in time of congestion to allow more successful retransmission for lost packets. To achieve fair resource sharing among competing sessions, we proposed multi-session video rate control schemes to enhance current layered video multicast rate control protocols, such that improvement in fair resource sharing can be achieved both among multiple video multicast sessions, and between video sessions and TCP sessions. |
