Mapping of Arbitrary Traffic Demand and Network Topology on a Mesh of Rings Network

Abstract

Abstract An efficient algorithm is presented to map an arbitrary traffic pattern and network topology on a mesh of rings network. The network is described by a two-layer model. The first layer defines the ring configuration and the second the connection allocation. Rings are constructed systematically from faces. It is assumed that a ring provides a fullmeshed network with a certain capacity between every node pair. Simulated annealing (SA) is used to optimize the mapping to best fit the traffic demands. For that, an objective function with certain parameters is introduced. It is possible to incorporate boundary conditions, e.g. the maximum number of nodes in a ring in this function. The algorithm is tested on the basis of the COST 239 Pan-European case study as an example.