Heuristic algorithm for maximizing streaming rate in overlay multicast with budget constraint

Abstract

Computers have taken over leadership from the radio and television and became one the most ubiquitous of consumer-electronic appliances. Thanks to the Internet, not only do they incorporate all radio's and television's features like Digital TV and Radio Broadcasting, but they also introduce new facilities. This causes rapid increase of the number of network users, which in turn translates into necessity for high bandwidth. Upgrading physical links in traditional Content Distribution Networks (CDN) is very expensive, therefore many approaches to content distribution have been developed over the last decade. In this paper we focus on overlay multicast due to its relatively low cost of implementation. This is accomplished thanks to the fact that overlay multicast systems are built on the top of the existing network infrastructure. Based on the previous works, we focus on the problem of simultaneous optimization of multicast flows and links' capacity of nodes participating in the overlay multicast. We assume a system with one central source of content, streaming multiple substreams to peers requesting particular amount of them, depending on the quality of the service required. The objective is to maximize the overall streaming rate of the network with additional constraint of budget to be spent on the system deployment. We propose an effective heuristic algorithm to solve the problem and compare it with results delivered by CPLEX solver. Moreover, in numerical experiments we show how various constraints following from real systems influence the behavior of the system.