A Case Study of Load Sharing Based on Popularity in Distributed VoD Systems

In our research, we consider a distributed video-on-demand (VoD) system in which only the most popular videos are replicated in all the servers, whereas the rest of them are distributed through the system following some allocation scheme. In this paper, we present an algorithm to efficiently share the load in such a system and an analytical model that captures the performance of this algorithm, which we validate through simulations. One novelty in our work is that our analytical model lets us relate popularity and partial replication of some of the videos and to predict the user waiting time. We exploit such relationships to assist the system designer to select the size of the servers and network, the optimal number of servers to maintain short waiting time and to predict when the network encounters bottleneck