A QoS-based joint scheduling and caching algorithm for multimedia objects

With the development of the broadband Internet, multimedia services have been widely deployed and contributed to a significant amount of today's Internet traffic. Like normal web objects (e.g., HTML pages and images), media objects can benefit from proxy caching; yet their unique features such as huge size and high bandwidth demand imply that conventional proxy caching strategies have to be substantially revised. Moreover, in the current Internet, clients are highly heterogeneous; it is necessary to differentiate their Quality-of-Service (QoS) requirements in streaming. However, the presence of an intermediate proxy in a streaming system poses great challenges to designers. This paper proposes a novel QoS-based algorithm for media streaming with proxy caching. We employ layered coding and transmission, and jointly consider the problems of caching and scheduling to improve the QoS for the clients. We derive general and effective solutions to the problems and evaluate their performance under various configurations. The results demonstrate that the proposed algorithm can accommodate diverse QoS demands from the clients, and yet satisfy stringent resource limits.