Research On Spectral Allocation And Defragmentation In Elastic Optical Networks

The rapid devlelopment and various new kinds of internet service have been increasing the total network traffic continuously.However,traditional wavelength division multiplexing(WDM)network applies fixed grid allocation which is no longer suitable for the requirements of rapidly developing network requirements.The orthogonal frequency division multiplexing(OFDM)based Elastic optical network(EON)can deeply divide spectrum resources into fine-granularity frequency slot(FS).Therefore,EON is able to increase network efficiency and reduce blocking probability.In this thesis,aiming on spectral allocation and defragmentation problem in EON,we carry out a indepth and comprehensive research using the method combining with theoretical analysis and numerical simulation.Firstlt,this thesis gives a brief introduction including basic conceptions,constructions and key techniques of EON.Meanwhile,several typical routing and spectral allocating algorithms of EON are demonstrated.Moreover,the spectral allocating strategies and an improved priority-of-service based allocating algorithm are proposed.In this algorithm,arriving traffic will firstly be sorted by priority according to the spectral state information,and then network will do routing and spectral allocating in ascending order of priority.Theoretical analysis and numerical results prove that this algorithm is effective to reduce the fragmentation level and blocking probility in network,and raise network efficiency.Additionally,a defragmentation algorithm based on expansion of the maximal unoccupied spectrum blocks is proposed.This method begins with choosing several unoccupied blocks combing using spectral state information.After that,moving the traffic on both sides of the chosen blocks to two ends of spectrum,and thus the chosen blocks have been expanded.Theoretical analysis and numerical results prove that this algorithm can clearly release the problem of fragmentation.