A rapid increase of the Internet users and traffic at the rate of 31% in years 2011–2016 contributes to emerging of new approaches to the content distribution. Among other approaches, the overlay multicasting seems to be one of the most interesting concepts according to relatively low deployment costs and large scalability. In this paper, the authors formulate a new incremental multicast overlay design problem. In particular, authors assumed that the overlay network is to be upgraded due to an increase of the number of participating users and the need to improve the streaming quality. However, the existing multicast tree structure is assumed to remain fixed. The goal was to minimize the cost of the upgrade, represented in euro/month. To achieve it, for each peer participating in the transmission, a link type offered by one of the ISPs was selected and overlay trees were constructed, rooted at the source of the content. The authors also present a new heuristic algorithm to efficiently solve this problem. According to experiments, the biggest factor influencing the upgrade cost and determining possible streaming quality values that the system can be upgraded to is the initial tree structure.
Live multimedia streaming and on-demand streaming applications (such as Internet radio or Internet TV) have been gaining more popularity in recent years. They require significant amount of bandwidth from media streaming servers and can easily saturate network infrastructure when the number of participant or bit rate of streaming content increases. Overlay multicast is an effective approach to the problem of streaming distribution. It combines flexibility of application layer multicast with efficiency of network layer multicast. Since overlay networks are built on the top of existing infrastructure, the cost of maintenance and deployment of this solution is relatively low compared to traditional Content Distribution Networks (CDN). Based on our previous works, we focus on solving the overlay network design problem to economically distribute content among the participants using overlay multicast. The optimization goal is to minimize the overlay network cost expressed by the cost of access links. Additionally, we assume that the maximum total delay of a streaming tree is upper bounded to provide QoS (Quality of Service) guarantees. We present two approaches to this problem and construct model using Levels and Flow Conservation Constraints. We show how various constraints following from real overlay systems influence the behavior of the distributing system. In numerical experiments we use real ISPs' price lists. To illustrate our approach we present optimal results obtained from the CPLEX solver.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.