Research On Resource Mapping Algorithms For Virtual Networks In Elastic Optical Networks

With the rapid development of data centers,cloud computing,Internet of Things(Io T)and other technologies,the rigid problems of traditional network architectures are becoming more serious.As the network function virtualization(NFV)technology could abstract one physical network into several independent virtual networks to share its physical resources,this high performance-cost ratio technology makes it possible to serve the connection requests according to the diverse needs of users.The NFV technology has become an effective way to solve the existing problems of the current networks due to the realization of the dynamic loading of network functions and the on-demand deployment of network resources,which accelerates the network technology innovations and their applications.Using the spectrum resources with fine granularity,elastic optical networks(EONs)can allocate bandwidth flexibly for connection requests with higher spectrum utilization,and meet the diverse needs of operators with less change to the existing network architectures,and thus EONs are considered to be the potential physical infrastructure for the NFV technology.Therefore,it is of great theoretical significance and practical value to carry out the research of virtual network resources mapping in EONs.The first chapter outlines EONs,NFV and the virtual network mapping.It introduces the research status of EONs,virtual optical networks(VONs)and virtual network function service chainings(VNF-SCs).The EONs structure and the principle of spectrum resource allocation are presented in the second chapter.With the definition of the mapping model of the underlying physical network and the upper virtual network,this chapter describes the principle of the virtual network mapping,summarizes the virtual network mapping algorithm according to different classification standards,and makes an analysis on advantages and disadvantages of various mapping strategies.This chapter also introduces the basic principle of genetic algorithm(GA),and makes an analysis on the characteristics of its application to large-scale optimization problems.Aimed at sharing physical resource in EONs efficiently for multiple VONs,the third chapter sets up an integer linear programming(ILP)model of VONs mapping in EONs,and proposes a novel algorithm of VONs mapping with load balancing(VMLB).In VMLB algorithm,the node computing resource,the node closeness centrality and the available link bandwidth of the physical network are comprehensively considered to complete the virtual nodes mapping.Besides,both the link distance and the occupied link bandwidth are combined to implement the virtual links mapping.Compared with the existing algorithm,simulation results show that the proposed algorithm could reduce spectrum resource usage and effectively improve the uniform performance of the network load of user services.The results have been published in the academic journal entitled Journal of Xidian University.To efficiently reduce the physical resources usage of VNF-SCs mapping in inter-datacenter elastic optical networks(DC-EONs),the fourth chapter sets up an ILP model and proposes a bi-population co-evolution algorithm(BPCEA)that is based on GA and used to solve the model for the first time.In two stages of the virtual nodes mapping and the virtual links mapping,the proposed algorithm sets two populations and makes the two populations evolve jointly with diverse convergence strategies and evolutionary strategies to search the global optimal solution.To improve the convergence efficiency,the Greedy Search(GS)strategy is designed to provide initial populations for BPCEA.Compared with the existing algorithms,simulation results show that BPCEA can greatly reduce the amount of physical resources used.The results are to be submitted to the academic journal entitled IEEE Communications Letters.On the basis of summarizing the content of this thesis,the fifth chapter presents the following three future study issues: the underlying physical network architecture,the optimization objectives of resource allocation and their optimization algorithms.