Impact of different link cost metrics on the performance of routing algorithms for traffic-engineered MPLS networks

This paper assesses the influence of different link cost metrics on the overall performance of routing algorithms, which are suitable for realizing traffic engineering in IP networks by means of MPLS (Multiprotocol Label Switching). The performance evaluation is conducted by means of extensive network simulations for a partially meshed network and three distributed routing algorithms, while the average end-to-end delay, total network throughput, and the total packet drop probability are taken as performance measures. It is shown that the link cost metric selection affects the performance of the considered routing algorithms in terms of the typical traffic engineering goals.