Impact of Graphs and Methodologies on Optical Access Networks Planning

Meeting the increasing capacity demands and availability requirements of modern optical access networks is crucial for network operators. Efficient planning tools are needed to evaluate new architectures over various areas before deciding on their adoption for a real deployment. This work shows the generalization capacity of Gabriel Graphs (GGs) for obtaining graphs with similar features from an input graph. The generated GGs allow to compare ILP formulations and a state-of-the-art Heuristic for planning Unprotected, Type A, and Type B Passive Optical Network (PON) architectures. The results show that the ILP requires, on average, 22% less fiber than the Heuristic. It is shown that the network’s sparseness drives the fiber requirement, as planning PONs with the ILP in sparse GGs use, on average, 12% more fiber than over the original graph. It is also shown that the network operator would use a comparable amount of fiber in planning an unprotected PON with the state-of-the-art Heuristic and a protected PON using Type A or Type B with the ILP.