Off-line reduced complexity layout design for MPLS networks

One of the main reasons for the present market enthusiasm for the MPLS architecture is its capability for implementing evolved traffic engineering (TE) functionalities, as they are required, for example, to support the next generation Internet. MPLS paths (LSPs) can be calculated on-line, as demands arrive, or off-line for a given estimate of the traffic matrix. Off-line calculation has the advantage of enabling a globally optimal network design. From the operational standpoint, the layout design for a large network should consider minimizing the layout complexity, thus reducing the cost of operation. Layout optimization objectives usually aim at minimizing total delay in the network, maximum loaded link, or any other function of link loads. To address the layout complexity issue, we propose an optimization problem whose objective is to minimize the number of required paths, while an end-to-end path delay constraint provides the required quality of service (QoS) guarantees. To solve the resulting multicommodity flow allocation problem, a MINLP (mixed integer non linear program) formulation is used. The results are then compared to the solutions obtained with a commonly used cost function, which does not consider complexity.