Providing scalable on-demand video services for heterogeneous receivers

To provide scalable video-on-demand services, precious system bandwidth must be shared among video requests. Although many efficient bandwidth-sharing techniques have been proposed, they are all designed for homogeneous receivers, i.e., the clients are assumed to have the same receiving bandwidth. For networks with client heterogeneity, these techniques either cannot work or have to compromise their performance. In this paper, we address this problem and propose an efficient solution allowing each client to use its whole receiving bandwidth to download data. To serve a client, the server first selects a set of serving channels for the client according to its actual receiving bandwidth. The video data needed for this client are then scheduled and dynamically adjusted for delivery over these channels. We evaluate the performance of the new technique using simulation and compare it with an existing scheme. Our study shows that by effectively leveraging client heterogeneity, the new technique achieves significantly smaller service latency.