Research On Spectrum Allocation Algorithms In Space Division Multiplexing Elastic Optical Networks

With the rapid development of emerging technologies such as virtual reality and artificial intelligence,modern society has an increasing demand for high-speed and large-capacity data transmission such as high-definition video and large data files.The elastic optical networks(EONs)have several advantages such as high flexibility and good spectrum utilization which makes it as the promising solution.The combination of space division multiplexing(SDM)technology and elastic optical network can greatly increase the transmission capacity and well meet the growing business needs,but it also brings some new problems.The inter-core crosstalk(ICXT)in multi-core optical fibers can limit the transmission distance of services and reduce the transmission quality of services,while excessive spectrum fragmentation can also cause subsequent service blocking,affecting the overall performance of the network.It is widely accepted by academic as a convenient and efficient strategy to solve these problems at the network layer.This thesis mainly studies the spectrum allocation algorithm to solve the problems of inter-core crosstalk and spectrum fragmentation in EONs with theoretical analysis and numerical simulations.The thesis first introduces key issues in EONs,including spectrum resource allocation,and delves into the issues of inter core crosstalk and spectrum fragmentation in space division multiplexing elastic optical networks.This thesis proposes a spectrum allocation algorithm based on threshold to reduce crosstalk in multi-core optical fibers.The algorithm combines core grouping and core priority strategy.The idea is to change the priority of the intermediate core through the dynamic frequency slot(FS)threshold.When the traffic load is high,the services with small frequency slot are allocated to the intermediate core in priority,and the protection frequency between different services is used to reduce inter-core crosstalk.The simulation results in two typical network topologies,NSFNET and USNET,show that under the same crosstalk generation,our algorithm can withstand higher network loads compared to other algorithms,and is also superior to traditional algorithms in terms of blocking rate.In order to solve the problem of spectrum fragmentation in the network,this paper proposes a spectrum allocation algorithm based on improved spectrum partition(Im-SP).In the traditional spectrum partition algorithm,the dedicated partitions are blocked,and the spectrum fragmentation of other dedicated partitions will be serious when the services are reallocated.The improved algorithm in this paper uses two different frequency slot allocation Patterns to optimize the above situation,which will not affect the spectrum fragmentation degree of other dedicated partitions.The simulation results in two typical network topologies,NSFNET and USNET,show that when there are five types of traffic rates: 40 Gbps,100Gps,200 Gbps,400Gbps,and 800 Gbps,the Im-SP algorithm proposed in this thesis improves the performance of bandwidth blocking probability and spectrum utilization rate by 10.7% and 7.5%,respectively,compared to the ZB algorithm and FASA-SP-FLF algorithm in the NSFNET network under high load of 1600 Erl.In the USNET network,the Im-SP algorithm proposed in this thesis improves the performance of bandwidth blocking probability and spectrum utilization rate by 8.5% and 5.8%,respectively,compared to the ZB algorithm and FASA-SP-FLF algorithm,at a high load of 2000 Erl.