A scalable video-on-demand system using multi-batch buffering techniques

Abstract

A video-on-demand (VoD) system delivers videos on demand over an installed network. Due to the large size of digitized videos, expensive video servers with high I/O capability are needed in order to provide VoD services in metropolitan areas. In addition, there is a great need for efficient networking distribution/interaction schemes so that the video servers can serve as many clients as possible. In particular, because of scalability problems, the classical unicast VoD system is not suitable for large-scale deployments. In this paper, a highly scalable VoD system with a low per-user cost is described and evaluated. We first analyze the performance degradation problems using recently proposed VoD systems, namely batched and centralized-buffer VoD systems that occur during the handling of interactions. Then a new system called the multi-batch buffer (MBB) system, which attempts to solve these problems, is proposed. The proposed system handles a majority of interaction requests by scalable buffering techniques employed in the buffer of the local servers and the set-top boxes (STB). We have performed extensive simulation for the analysis and performance evaluation of our proposed VoD system. The simulation results demonstrate that our VoD system is very scalable and outperforms related state-of-the-art VoD systems.