Integer Linear Programming Models for Media Streams Planning

Advanced collaborative environments frequently need to transfer highly demanding multimedia data streams with minimum possible latency. Since bandwidths of the streams are close to capacities of currently available links, routing of data transfers therefore requires planning with respect to link capacities and latency optimization. This paper describes integer linear programming techniques for optimal solution of the multimedia streams planning problem. Several methods for cycle avoidance in transmission graphs and network flows formulation are proposed. Performance of the methods was evaluated to identify their efficiency for real-time planning and to compare it against an earlier constraint programming approach applied in an application middleware called CoUniverse. According to the results, network flows formulation appears to be the most promising.