Graph-model-based dynamic routing and spectrum assignment in elastic optical networks

In recent years, the concept of the elastic optical network (EON) was proposed and very quickly became an attractive solution for flexible channel bandwidth utilization. Analogous to the routing and wavelength assignment (RWA) problem in traditional wavelength-routed networks, routing and spectrum assignment (RSA) is the most basic and critical resource management issue in EONs. In this work, we exploit and modify the layered graph model, which is a well-known RWA model, to design two RSA heuristic algorithms named LG-FF and LG-SP. Additionally, we also derive an analytical model to estimate the number of required layered graphs. Numerical results demonstrate that LG-SP can achieve the same blocking performance level with significant time complexity reduction as compared to the near-optimal solution. Whether we care about blocking performance or aim at saving computational time, the layered graph is an important representation model of resource utilization status to efficiently solve the RSA problem.