Research On Routing And Spectrum Assignment Problem In Elastic Optical Networks

Internet traffic is facing an exponential growth due to the advent of newbandwidth-intensive services, which poses a great pressure on traditional WavelengthDivision Multiplexing (WDM) optical networks. Compared to the WDM optical network,a Coherent Optical Orthogonal division Multiplexing (CO-OFDM) elastic optical networkhas a finer granularity of channel spacing and can allocate spectrum more flexibly, whichtherefore makes it a promising solution for the next-generation optical transport network.In this thesis, we look into the problems on how to enhance the spectrum efficiency ofthe elastic optical networks. First, for the routing and spectrum assignment (RSA) problem,we develop an Integer Linear Programming (ILP) model and several efficient heuristicalgorithms. The results show that the heuristic algorithms are efficient to perform close tothe ILP model in terms of required spectrum resources. Second, we apply the sub-bandvirtual concatenation (VCAT) technique to the elastic optical network. Analytical modelsfor estimating the lightpath blocking performance are developed and simulations based ondynamic lightpath demands are conducted to evaluate the benefit of the VCAT technique.The results indicate that the VCAT technique can help significantly reduce lightpathblocking probability compared to the case without applying the technique. Finally, we alsoconsider the electronic traffic grooming for IP over elastic optical networks. We develop amixed ILP (MILP) model to achieve optimal traffic grooming. Numerical results show thattraffic grooming can significantly improve the throughput for the elastic optical network.