Data storage and retrieval for video-on-demand servers

Abstract

Video on demand (VOD) is a system that provides a service to users to browse and watch videos within a computer network. One of the most challenging aspects in such system is to have a VOD server with a capability of retrieving and transmitting different blocks of videos simultaneously. A part from this the VOD server should also capable of producing high number of simultaneous streams with a low average waiting time. In this paper we present a new data storage scheme employed by the VOD server to eliminate latency and hence producing maximum number of simultaneous streams. In this scheme the disk is divided into regions and blocks of video are scattered within the regions based on the residual theory rule. We also developed a stream caching technique for reducing the average waiting time of the system.