An efficient traffic engineering based on multi-topology routing for future Internet

There is no doubt that the future Internet traffic will be dominated by the multimedia services. However, their strict quality of service requirements, as well as bursty nature, are not well suited for the best-effort delivery of the Internet infrastructure. In this paper, we propose a novel multi-topology routing based traffic engineering approach. The scheme can provide simple, yet efficient, solution with the near-optimal network performance even under unpredicted traffic spikes. While taking into account the flow's delay restriction in routing, the proposed scheme can also provide link failure resiliency. First, based on a proposed algorithm, fully edge-disjoint logical views of a network are extracted in a way that the delay of the longest path is upper bounded. Then, by using the master-slave optimisation problem, the proposed scheme selects the longest acceptable path for each traffic type. This can guarantee that the shortest paths are always available and can be used by the most legitimate traffic in the network. We prove that finding the multiple disjointed logical topology is NP-hard. Therefore, we present heuristic algorithms to handle the problem. Using extensive evaluations based on real and arbitrary networks and traffic matrices, we show that our scheme can achieve efficient resource utilisation with regards to flows' delay requirement.