Research On Spectrum Allocation And Survivability In Elastic Optical Networks

With the rapid development of cloud computing,internet of things,big data and other network technologies,the traffic demand in the communication network shows an exponential growth,which puts forward new requirements on the data transmission capacity and hardware resource allocation in the backbone optical network.The traditional Wavelength Division Multiplexing optical network technology is not sufficient enough to adapt to this new requirement due to its low frequency utilization and rigid spectrum allocation.In recent years,the elastic optical network technology based on optical Orthogonal Frequency Division Multiplexing has become a promising technology because of its fine-grained spectrum slices,flexible modulation methods and efficient spectrum utilization.However,the spectrum continuity constraint in elastic optical network technology may generate a large number of spectrum fragments in the process of dynamic establishment and release of service connections,and cause serious service congestion.Although the spectrum converter can relax the spectrum continuity constraint,it may lead to the network cost to increase dramatically.In addition,as elastic optical network has a large amount of data carrying capacity,any node or link failure may cause a large number of business interruption and data loss,leading to a serious decline in service quality,or even a huge economic loss.Therefore,the problem of spectrum resource optimization and survivability is of great significance to the future development and application of elastic optical networks.Aiming at the above two challenges,this thesis studies the following research.To address the problem of spectrum allocation and optimization in elastic optical networks,a spectrum allocation method based on hybrid spectrum conversion resource pool is proposed to establish connections for business requests.In this strategy,the longest continuous spectrum segment set is established according to the idle state of the bandwidth on the service bearing route.Then,from the direction of the source node to the destination node,the spectrum conversion relation in the set of adjacent spectrum segments is investigated in turn,and the limited range spectrum converter is adopted in priority to relax the spectrum continuity constraint and calculate all feasible optical paths.Finally,according to the principle of low frequency continuous spectrum segment priority,the unique optical path is determined and the corresponding spectrum and spectrum converter are allocated.The simulation results show that the proposed method can effectively reduce the number of full-range spectrum converters in the network while maintaining a low blocking rate,thus reducing the configuration cost of network hardware resources.In order to survive the services running on backbone elastic optical network from a large-scale disaster,a two-channel service recovery strategy based on link lifetime is proposed in this thesis.This strategy aims at the network survivability problem caused by catastrophic network failure,such as service interruption and severe loss of service time.At the same time,the two-channel routing scheme is adopted,and the threshold of traffic loss is adjusted to restore the service,so as to improve the connectivity of the service.Simulation results show that the proposed scheme has obvious advantages in reducing traffic and duration loss while guaranteeing the service connectivity.