Rate control algorithms for video storage on disk based video servers

The emerging video-on-demand servers are equipped with large capacity and high speed disks, but the disk seeking time can be significant (in the case when several users access different data simultaneously). Studies show different approaches to reduce the disk seek time in order to increase the number of users that can be served simultaneously. Most of them provide strategies for disk storage (disk placement) of the video data. Since most video data is encoded with variable bit rate (VBR), using fixed-rate disk placement strategies may result in reduced quality (PSNR). We analyze the impact of disk placement strategies on the performance of the VBR video stream. We then show that the performance can be improved by using rate control techniques that are aware of the disk placement constraints, so that the VBR video source stream can be optimized for those constraints. This optimization (which can be performed off-line) results is quality improvements of 0.5 dB to 1.5 dB in our experiments.