Adaptive Multicast of Multi-Layered Real-Time Video

Tatsuya Suda
Department of Information and Computer Science
University of California, Irvine

and

Networking Research Program, NSF

Monday, May 11, 1998 
11am-12:15n
 Interschool Lab, 7th floor, Schapiro CEPSR Bldg.

Host: Yechiam Yemini

Abstract

In an era of proliferating multimedia applications, support for video transmission is rapidly becoming a basic requirement of network architectures. Furthermore, since most video applications (e.g., teleconferencing, video broadcast, and video surveillance) are inherently multicast in nature, support for point-to-point video communication is not sufficient; a novel video multicast service is required. However, multicast video transport is severely complicated by variation in the amount of bandwidth available throughout the network.

This talk will describe Dr. Suda's recent systems research efforts, which have been directed toward the development of efficient multicast service for real-time video. Dr. Suda proposed an adaptive, multicast-capable, real-time video service that operates in high speed networks in the face of heterogeneous and time-varying bandwidth constraints. In this novel service, raw video is encoded into multiple layers, each with a different network discarding priority. The layer with the highest priority contains the most important aspects of the video stream, while layers with lower priorities contain refinement data for those receivers with the additional bandwidth to receive it. The service attempts to optimize the utilization of the network and the quality of video at each of the receivers by dynamically adjusting the transmission rates of each video layer in response to network congestion feedback.

Through empirical study and simulations, the proposed video service is shown to scale and efficiently utilize network resources.

This talk will also briefly summarize Dr. Suda's recent research accomplishment.



Luis Gravano
gravano@cs.columbia.edu