Two hybrid multicast algorithms for optimizing resources in a distributed VoD system

Among many different multicast approaches, batching and patching are two commonly used policies. In this paper, we propose two hybrid multicast algorithms inspired by the maximum factored queue length (MFQL) batching scheme - used to decide which video queue will be serviced with a multicast session - and by a threshold-based patching scheme - applied to control the partial streams transmission before a threshold is reached during an ongoing multicast session. The novelty is that our algorithms have been designed to efficiently handle the requests in a distributed VoD system while optimizing the usage of the sever/network resources. Precisely, one key issue in the optimization of these resources is the computation of the threshold. We show in this paper how that threshold is derived to deliver a service. In addition, we conduct some simulation experiments that provides us some insightful information about the impact that our threshold based multicast algorithms have in the average waiting time in a distributed VoD system.