Traffic and network engineering in emerging generation IP networks: A bandwidth on demand model

This paper assesses the performance of a network management scheme where network engineering (NE) is used to complement traffic engineering (TE) in a multi-layer setting where a data network is layered above an optical network. We present a TE strategy which is based on a multi-constraint optimization model consisting of finding bandwidth-guaranteed IP tunnels subject to contention avoidance minimization and bandwidth usage maximization constraints. The TE model is complemented by a NE model which uses a bandwidth trading mechanism to rapidly re-size and re-optimize the established tunnels (LSPs/lambda SPs) under Quality of Service (QoS) mismatches between the traffic carried by the tunnels and the resources available for carrying the traffic. The resulting TE+NE strategy can be used to achieve bandwidth on demand (BoD) in emerging generation IP networks using a (G)MPLS-like integrated architecture in a cost effective way. We evaluate the performance of this hybrid strategy when routing, re-routing and re-sizing the tunnels carrying the traffic offered to a 23-node test network.