Optimal stream-scheduling scheme in video-on-demand

Due to the high bandwidth and the long-lived nature of digital video, server and network bandwidths are proving to be bottleneck in video-on-demand. Therefore many researches are presented on efficient utilization of bandwidth resources. Stream-Scheduling mechanisms are an important way of efficient utilization of bandwidth resources. Periodic broadcast have been used to stream a set of popular videos from a remote server to a large number of clients. In periodic broadcast, the initial segments are repeated more frequently than later ones. This has the benefit of reducing both the client startup delay and the transmission bandwidth requirements. But the inherent average-delay is the half of the initial smallest segment size. In order to reduce the inherent delay, the initial smallest segment size must be decreased, and thus resulting in more bandwidth requirements. In order to solve the confliction of service delay and network bandwidth in video-on-demand, an optimal stream-scheduling scheme was proposed that is called Patching-combined Segment Broadcast scheme (PSB). A theoretic analysis of optimizing channel allocation for the popular videos and service delay were given with Lagrange multiplier method and G/G/m queuing model. Performance analysis shows that PSB scheme can achieve up to 50% bandwidth resource reduction compared to other exiting schemes under general tolerable delay. On the other hand, the delay is decreased and Quality of Service is improved under a bandwidth usage.