Research On Routing And Spectrum Allocation Models And Algorithms Of Elastic Optical Networks

With the rapid development of the Internet,new business forms continue to emerge.In the face of increasing network data traffic,traditional wavelength division multiplexing networks have become increasingly difficult to meet the ever-increasing bandwidth demand due to their inherent technical limitations.Therefore,the flexible optical network technology is proposed on this basis.The flexible optical network adopts a more fine-grained spectrum allocation method,which can improve the spectrum resource utilization rate of the network and can meet the bandwidth demand of large capacity.In response to the need for network expansion and facilitation,network function virtualization technology(NFV)has emerged,which enables the network functions carried by original physical devices to implement virtual network functions(VNF)in a software-defined manner and integrate them into the network,Multiple VNFs are deployed in the data center according to a certain order in the service function chain(SFC)(that is,the dependency of the VNF)for business data flow processing,which increases the flexibility and portability of the network.However,how to effectively deploy data centers and VNFs in the network has become an important issue affecting network resource consumption.The main research aspects of this thesis are:(1)the data center load balancing problem that considers the VNF dependency relationship,and(2)the virtual network function mapping problem that considers the VNF dependency relationship.In order to solve these two problems,the main contributions of this thesis are:1.Research the data center load balancing problem considering the VNF dependency.The dependency of VNF has an important influence on the resource allocation in the network,and it also plays a key role in solving the RSA problem.Therefore,this thesis first mathematically expresses the dependency of VNF and considers the data center of VNF dependency.The load balancing problem is established a global optimization model with constraints whose objectives are to minimize the standard deviation of the data center load and the consumption of spectrum resources.In order to solve the model,firstly design the processing algorithm for the dependence relationship,and design the crossover and mutation operators,and on this basis,design an effective algorithm based on genetic algorithm for solving the model.In order to verify the accuracy of the model and algorithm,simulation experiments are performed on different topological networks with other algorithms.The results show that the model established in this thesis is accurate and the algorithm is feasible.2.Research the virtual network function mapping problem considering the VNF dependency.Due to the existence of VNF dependencies,the deployment of data centers and VNFs is more challenging.In order to minimize data center resource consumption,VNF resource consumption,and spectrum resource consumption,this thesis establishes a method based on the minimizing consumption of these three resources.The target constrained optimization model.In order to effectively solve this model,we designed the data center determination algorithm and the VNF deployment algorithm,and combined it with the genetic algorithm to solve the model.In order to verify the accuracy of the model and algorithm,compared with other algorithms on unused topological networks,the simulation results show that the algorithm in this thesis has a better effect on solving this model than other algorithms.