Abstract

Network operators face the challenge of solving routing and wavelength dimensioning problems while aiming to simultaneously minimize the network cost and ensure that the network performance meets the level established in the Service Level Agreement (SLA). We call this problem the routing and wavelength dimensioning (R&WD) problem. Finding the optimal solution to this problem is hard. Thus, in practice, heuristics are used to obtain a good solution with low computational complexity. The strategy used most to solve the R&WD problem has been to separate the routing from the dimensioning problem. The routing part is tipically solved in two phases: first, the information of every node pair is routed using the shortest path; second, the set of shortest paths that balance the traffic across the network is selected. Later, according to the routing obtained, in order to meet the SLA, the number of wavelengths per link is dimensioned using an appropriate method. In this work we apply an approach that differs from the common practice described above in the following ways: (a) it simultaneously solves the R&WD problems, (b) it does not necessarily use the shortest path routing algorithm, and (c) the load balancing strategy is not used. In spite of these great differences with the common approach, the algorithm proposed here yields significative better results than the current approach. (C) 2014 Elsevier Inc. All rights reserved.