Research On Routing And Spectrum Allocation For Green Elastic Optical Datacenter Networks

With the rapid development of information technology,such as big data and cloud computing,datacenter network business traffic has shown explosive growth,which has caused great consumption of network spectrum resources and the energy consumption of datacenter networks.How to solve the problem of spectrum resource allocation and energy consumption of datacenter networks is an important research direction in the current communication field.This thesis focuses on the spectrum resource allocation and service energy consumption of the elastic optical datacenter networks,and studies the routing and spectrum allocation and energy saving algorithms in the elastic optical datacenter networks.It mainly includes two aspects: one is a routing and spectrum allocation scheme based on anycast or multicast for datacenter services,the second is a green energy-efficient routing and spectrum allocation scheme based on joint costs.The specific research content is as follows:First of all,in view of the datacenter service routing and spectrum allocation in the elastic optical datacenter networks,this thesis proposes a routing and spectrum allocation scheme based on anycast or multicast for datacenter service.The solution calculates the cost of all frequency slot windows in the alternative path according to the spectrum resource status in the network and the service request bandwidth,selects the least cost route and frequency slot window for the services,and reduces the generation of spectrum fragments.According to simulated annealing algorithm,the routing and spectrum allocation model is established,which selects anycast or multicast-based routing spectrum resource allocation scheme for the service to reduce the connection blocking rate.Simulation results show that the proposed scheme has improved in reducing network spectrum fragmentation and improving resource utilization,and effectively reduces the connection blocking rate.Secondly,in view of the business energy consumption problem in the elastic optical datacenter network,a green energy-efficient routing and spectrum allocation scheme based on joint cost is further proposed.According to the network energy consumption and spectrum resource status,as well as the service request bandwidth and service duration,this solution dynamically updates the energy consumption costs of network nodes and links,and preferentially selects low-power modulation methods to calculate the minimum energy consumption cost alternative paths for each datacenter,and calculate the resource cost of all available frequency slot windows on the alternative path.According to the renewable energy supply of the datacenter,the scheme uses the simulated annealing algorithm to establish a green energy-efficient routing and spectrum allocation model,which can obtain the minimum joint cost of anycast or multiple energyefficient routing and spectrum resource allocation scheme through iterative optimization.The simulation results show that the proposed scheme can effectively reduce the network services energy consumption and blocking rate.