Optimising the geographical location of transcoding resources

Ultra-high definition video, sometimes termed 4k or 8k video, is likely to become the format for future video services for both home and cinema media. However, streaming real-time ultra-high definition content consumes a large amount of bandwidth in a network, which can create delays as well as lost packets for other network traffic, in particular on constrained links. Consequently, it is desirable to reduce the traffic along constrained links while providing the same quality of service to clients. By using a combination of application layer multicasting and transcoding it is possible to reduce the bandwidth usage in the network and provide a more efficient content delivery service. However, as transcoding is unlikely to be available at every location, it is required to solve the optimization problem of transcoder placement while taking into account computational and network resource constraints. This paper presents a heuristic solution to this problem. It is shown that this solution is better than a generic problem solver, such as a genetic algorithm, in terms of computational requirements, but without reducing optimization quality.