Dynamic Routing And Spectrum Assignment Algorithms Based-on Differentiated Degraded-Service In Elastic Optical Networks

With the aggregation and segmentation of spectrum resource,flexible variation of allo-cated resources,efficient accommodation of different data rate,etc.,Elastic Optical Network(EON)based on Orthogonal Frequency Division Multiplexing(OFDM)technology can be well adapted to the future development of large-capacity,high-rate,scalable,dynamic and variable network communication services,and is regarded as the mainstream direction of the future optical transmission network.The Routing and Spectrum Allocation(RSA)is a key issue for EON,which mainly focuses on improving the resource utilization and reducing traf-fic blocking probability.As for it,three effective RSA algorithms are proposed and evaluated by simulation in this paper.In this paper,we introduce the background,development status and key technologies of EON,focus on the RSA problem of EON,classify and summarize the RSA algorithms,and analyze the main ideas of the typical RSA algorithms.Secondly,in order to reduce the impact of the spectrum fragments caused by the leaving of traffics on the occupied spectrum area,we establish a Mixed Integer Linear Programming(MILP)model with a joint objective that minimizes the number of blockd traffics and leav-ing-time difference of traffics between new traffic and adjacent established traffics in the time-spectral domain and time-spatial domain.And then we propose a Fragment-Aware Routing and Spectrum Allocation based on Holding-Time(FARSA-HT)algorithm.The algorithm adopts a spectrum allocation strategy that minimizes the standard deviation of the leaving time between adjacent traffics in the time-spectral domain and time-spatial domain,and always selects a spectrum window with the minimum standard deviation of leaving time for new traffic,which can ensure the leaving-time similarity of traffics between new traffic and established traffics in the time-spectral domain and time-spatial domain,and provide wider,more connected spectrum resources for future traffics.Thirdly,in order to solve the problems of excessive degradation,unfair resource alloca-tion,and the neglect of spectrum and revenue balance in the existing Degraded-Service(DS)algorithms,we establishe a MILP model with a joint objective to minimize both spectrum consumption,the priority and frequency of DS,and proposes a Dynamic Routing and Spec-trum Assignment based-on Differentiated DS and Adaptive Modulation(DRSA-DDAM)al-gorithm.The algorithm designs a DS loss function and DS window selection strategy which combines the service-priority differentiation,the distance-adaptive modulation and DS tech-nology to minimize DS loss and indirect impact.In addition,to alleviate the problem of unfair resource allocation caused by DS,a network revenue function considering the relationship between spectrum and revenue is designed,which achieves the purpose of efficient utilization of spectrum resources and maximization of network revenue.Fourthly,in order to solve the problems of unfair resource allocation,and the neglect of quality of experience(QoE)of low-priority traffics in the existing DS algorithms,we estab-lishes a MILP model with a joint objective that minimizes both the priority and frequency of DS and transmission delay loss,and then propose a Delay-aware Degradation-Recovery Routing and Spectrum Allocation(DDR-RSA)algorithm.The algorithm not only considers the service-priority differentiation,but also designs the DS delay loss function and DS win-dow selection strategy to minimize the loss of DS.Simultaneously,network revenue function considering the relationship between spectrum allocation and revenue is designed,which re-duces the service charge of the degraded users and enhances the stickiness.In addition,the Degradation Recovery(DR)strategy is designed in the DS window selection strategy,which recovers the degraded traffics by utilizing the unoccupied spectrum domain under the premise of ensuring the fixed amount of transmission traffic data,and achieves the purpose of provid-ing optimal DS window for high-priority traffics,reducing the transmission delay loss of the degraded traffics and maximizing the network revenue.Finally,the EON simulation platform is built by Visual C++to simulate and analyze the proposed three algorithms under the dynamic traffic model and to compare them with the ex-isting algorithms in the literatures.The effectiveness of the algorithms is verified.