Regeneration and grooming in MPLS over WDM networks

• Autor(a) principal: Nogueira, Bruna
• Data de Publicação: 2019
• Outros Autores: Mendes, Ricardo, Martins, Lúcia, Gomes, Teresa, Girão-Silva, Rita, Santos, João, Cinkler, Tibor
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10316/87681; https://doi.org/10.1364/JOCN.11.000465
• topic: Shortest Path; Regeneration; Traffic Grooming; Transponder; ILP; Heuristic; WDM

This work addresses the Grooming, Routing and Wavelength Assignment with Regeneration (GRWAR) problem for meshed networks, in static traffic scenarios. It focuses primarily on the Grooming and Routing with Regeneration (GRR) problem. Physical impairments were considered in the model, imposing a limit in lightpath lengths without regeneration. This creates the opportunity of strategically placing transponders for both grooming and regeneration. The GRR problem is tackled lexicographically, which means that for a given network topology and a traffic matrix, the goal is to route and groom connection requests in a way that minimizes the number of transponders and, for that number of transponders, throughout minimum length routes. To solve this problem, an Integer Linear Programming (ILP) formulation was developed considering undirected networks and explicitly including the relationship between the different network layers. The results of the ILP model are used to validate a heuristic suited to be included in a Software Defined Network (SDN) platform in a multilayer core network. The GRWAR problem is addressed after a solution for the GRR problem is obtained, using a first-fit approach to assign wavelengths to the lightpaths obtained by the ILP and the heuristic. The wavelength assignment problem is less critical, due to the large number of wavelengths per fiber. The ILP and the heuristic were compared for small networks, with the heuristic providing good results in very short time, proving to be efficient and better suited for larger problems, as is confirmed through a set of proposed bounds.