Optimized resource distribution for interactive TV applications

Though ubiquitous, the full potential of consumer electronic devices in the home, as content creators, remains underutilized due to the limited interaction between the consumers and the existing on-demand application and media services. Although services such as interactive television could change this, the geographic distribution of groups of consumers and the need for on-the-fly media processing that this entails, makes the efficient utilization of resources a complex optimization task requiring mechanisms to simultaneously allocate processing and network resources to groups of users. However, these technologies have not yet been developed, and brute force methods remain prohibitively complex. In order to overcome this problem, this paper proposes heuristic algorithms to both generate end-to-end delay bound multicast trees for individual groups of users and to co-locate multiple multicast trees, such that a minimum group quality metric can be satisfied. The performance of the proposed heuristic solution is evaluated in terms of the serving probability, i.e., the resource utilization efficiency, and computation time of the resource allocation decision making process. Simulation results show that improvements in the serving probability of up to 50%, in comparison with existing generic resource allocation schemes, and several orders of magnitude reduction of the computation time, in comparison to an optimal linear programming solution approach, can be achieved.