Dominating set algorithms for sparse placement of full and limited wavelength converters in WDM optical networks

Abstract

Optical wavelength converters are expensive, and their technology is still evolving. Deploying full conversion capability in all nodes of a large optical network would be prohibitively costly. We present a new algorithm for the sparse placement of full wavelength converters based on the concept of a k-minimum dominating set (k-MDS) of graphs. The k-MDS algorithm is used to select the best set of nodes that will be equipped with full-conversion capability. To allow placement of full wavelength conversion at any arbitrary number of nodes, we introduce a HYBRID algorithm and compare its performance with the simulation-based k-BLK approach. We also extend the k-MDS algorithm to the case of limited conversion capability by using a scalable and cost-effective node-sharing switch design. Compared with full search algorithms previously proposed in the literature our algorithm has better blocking performance, has better time complexity, and avoids the local minimum problem. The performance benefit of our algorithms is demonstrated by simulation.