1. Impact of Mathematical Optimization on OSPF Routing in WDM Optical Networks
- Author
-
Himanshi Saini and Amit Kumar Garg
- Subjects
Routing protocol ,Random graph ,Computer science ,business.industry ,Ant colony optimization algorithms ,Open Shortest Path First ,Evolutionary algorithm ,Communications system ,business ,Multiplexing ,Bottleneck ,Computer network - Abstract
High-speed communication systems require robust routing protocol. The selection and optimization of the routing protocol in Wavelength Division Multiplexed (WDM) optical networks decide utilization of bandwidth capacity offered by these networks. When Open Shortest Path First (OSPF) protocol is used for routing, link weight selection is one of the main concerns in order to optimize the routing. In this paper, OSPF link weight selection is performed through Ant Colony Optimization (ACO) and Evolutionary Algorithm (EA). The novelty of the work lies in the implementation of optimization techniques in offline network mode on networks of variable densities with an objective is to reduce utilization of bottleneck link, i.e., to reduce congestion. As a result, knowledge of optimization specific to type of network can be retained in online networking mode. ACO and EA are implemented on random network which is composed of 4 nodes and 5 links (4n5e), standard National Science Foundation NETwork (NSFNET), and standard COST 239 networks in order to test the optimization on networks of different densities. Maximum End-to-End (E2E) latency and bottleneck Link Utilization (LU) obtained after application of optimization techniques is compared for networks considered. It is observed that EA optimization has a better optimized denser network (COST 239) and ACO has optimized NSF and 4n5e networks better than EA optimization.
- Published
- 2018