Optimization of traffic flows in multiservice telecomunications networks

This dissertation investigates routing optimization in IP telecommunication networks, under normal working conditions as well as under failure conditions. The main objectives of the present optimization procedure are the minimization of the maximum link utilization in the network and to provide a configuration that guarantees a 100% survivability degree. Traditionally two different steps are used to achieve this goal. The first one aims to solve the well known “General Routing Problem (GRP)” in order to find the optimal routing network configuration and, successively, a set of “optimal” backup paths is found in order to guarantee network survivability. Furthermore, traditional survivable techniques assume that the planning tasks are performed in a network control center while restoration schemes are implemented distributively in network nodes. In this dissertation innovative linear programming models are presented that, making use of the Multi Protocol Label Switching – Traffic Engineering (MPLS-TE) techniques and IS-IS/OSPF IP routing protocol, melt routing and survivability requirements. The models are extremely flexible, thus it is possible to improve the objective function in order to fit itself to newer applications and/or traffic typologies. The models presented in this dissertation help Internet Service Providers to optimize their network resources and to guarantee connectivity in case of failure, while still be able to offer a good quality of service.