Hybrid model of traffic characterization for optimization in GMPLS optical networks

This paper presents a hybrid traffic model for GMPLS (generalized multi-protocol label switched) optical networks. The model is focused in multi-service traffic characterized as self-similar and is used to produce more accurate QoS (quality of service) metrics which influence the different hierarchies of the network topology. Also, the model is used to implement an optimization procedure with a genetic algorithm for the establishment of link capacities in different physical mediums. The extensions of the model for the hierarchies of LSPs (label switched paths) in the GMPLS suite are also analyzed.