Receiver-driven Layered Multicast-英文文献.pdfVIP

Receiver-driven Layered Multicast Steven McCanne Van Jacobson University of California, Berkeley and Network Research Group Lawrence Berkeley National Laboratory Lawrence Berkeley National Laboratory mccanne@ van@ Martin Vetterli University of California, Berkeley martin@ Abstract 1 Introduction State of the art, real-time, rate-adaptive, multimedia applications The Internets heterogeneity and scale make multipoint com- adjust their transmission rate to match the available network ca- munication design a difficult problem. For real-time mul- pacity. Unfortunately, this source-based rate-adaptation performs timedia, we would like to “broadcast” a live signal from poorly in a heterogeneous multicast environment because there is any particular sender to an arbitrarily large set of receivers no single target rate — the conflicting bandwidth requirements of along paths with potentially high variability in bandwidth. all receivers cannot be simultaneously satisfied with one transmis- The simplest solution to this problem is to distribute a uni- sion rate. If the burden of rate-adaption is moved from the source form representation of the signal to all interested receivers to the receivers, heterogeneity is accommodated. One approach to using IP Multicast [8]. Unfortunately, this is suboptimal — receiver-driven adaptation