Investigations of Optimized Optical Network Performance Under Different Traffic Models

Abstract

Network planning for existing as well as future high-speed networks is important for extracting best performance of networks. Network planning requires reliable traffic model which can serve constant as well as variable ingress traffic. In this paper, network performance is investigated with uniform and population-distance traffic models. Open Shortest Path First (OSPF) protocol is used for routing and link weight determination is a crucial task for this routing. Link weights in networks of different densities are optimized with an objective of minimizing network congestion. Ant Colony Optimization (ACO), Genetic Algorithm (GA) and Simulated Annealing (SAN) optimization techniques are applied to examine link weights of National Science Foundation NETwork (NSFNET) and standard COST 239 networks. The novelty of the work lies in investigations of network performance with different optimization techniques, traffic models and density. The outcome of this work can assist in optimizing overall network planning. Maximum latency and congestion of both networks are compared for each optimization and traffic model. It is observed that population-distance traffic modeling has reduced network congestion for both the networks but this traffic model has increased maximum latency of NSFNET. Performance of COST 239 network which is denser than NSFNET, has improved with population-distance traffic model w.r.t. congestion and latency.