Improving Scalability of VoD Systems by Optimal Exploitation of Storage and Multicast

Today, video-on-demand (VoD) systems are challenged by a growing number of users, growing sizes of libraries, and increasing video streaming rates. Therefore, scalability and the bandwidth efficiency of VoD systems have become important considerations. In this paper, we propose a scalable and bandwidth efficient delivery scheme for VoD systems, which optimally exploits the storage and multicast capabilities to reduce the consumption of server capacity resources. Our proposed scheme, which we call prepopulation assisted batching with multicast patching (PAB-MP), facilitates video multicast from the server by strategic preplacement of initial segments of videos at the end-users´ devices. Using an analytical approach, we show how the parameters for the PAB-MP scheme can be selected to achieve optimal performance. Moreover, we propose methods to replicate and place the initial video segments (IVSs) across the end-users´ devices such that load arising from IVSs is evenly distributed among user-nodes. Using simulations, we show that our proposed scheme effectively reduces the load on the server especially under high load, while imposing less burden on individual user nodes. Simulation results indicate that our proposed PAB-MP scheme is significantly more scalable than the other popular approaches that exploit multicast and storage.