A metadata-based video-on-demand traffic control over the network supporting bandwidth renegotiations

In this paper, we study an effective video-on-demand traffic control algorithm using the metadata over the network supporting bandwidth-renegotiations. The proposed algorithm includes bandwidth smoothing. bandwidth-burstiness estimation and rate adaptation algorithms. The proposed video-on-demand server has not only video database but also metadata database that includes coding information of the compressed video in video database and the traffic burst characteristics with respect to control parameters of the bandwidth smoothing algorithm. Thus, we can predict the traffic properties accurately with a low computational complexity by using the stored metadata, and then determine the efficient bandwidth renegotiating variables such as the renegotiating instants and the required bandwidth in terms of network utilization and video-on-demand service quality. In addition, we present a rate adaptation algorithm that pursues an effective trade-off between spatial and temporal qualities of the decoded video to improve the perceptual video quality when the bandwidth request is rejected.