Energy-efficient routing and wavelength assignment in translucent optical networks

We study the energy-efficient routing and wavelength assignment (RWA) problem in translucent WDM networks, taking into account transmission limits of optical signals and the consequent need for signal regeneration. The problem is formulated through an integer linear programming model, and a heuristic algorithm called least regeneration first (LRF) is proposed to solve it in large networks. We first show that taking into account physical layer impairments and the consequent possible regeneration is mandatory to achieve satisfactory energy efficiency in RWA algorithms. In this respect, results show that LRF is much more efficient than "impairment unaware" RWA algorithms. Then, we assess the energy efficiency achievable with different modulation formats using the proposed LRF algorithm in single-line-rate networks. Finally, the case of mixed-line-rate networks is also considered, and a heuristic method is proposed to efficiently decide the set of connections, at different bit rates, to be assigned to each traffic demand, with the goal of minimizing the total consumed power. Results show that this method is able to achieve power saving up to 25%.