A Research On Controller Placement Problem In Software Defined Networking

The rapid growth of data traffic and the diversification of network service requirements have promoted the upgrading of traditional networks.To satisfy the demand for the growing network,researchers have redesigned a new type of network,the Software Defined Networking(Software Defined Networking,SDN).SDN separates the data plane and control plane in the network,making it possible for centraling control and programmable on control plane,making it easier and faster for managing the whole network and developing new functions of network,and making it easier to enlarge the scale of network.Although SDN has many advantages,there are still many problems to be solved to apply SDN to actual network environment.When SDN is applied in a large-scale WAN,the performance of a single SDN controller cannot handle the whole network.To improve the performance of the whole control plane becomes an urgent problem to be solved.Research shows that deploying multiple controllers in the control plane to alleviate the pressure of a single controller is an effective way to solve the problem.Improving the control plane's performance by deploying multiple controllers introduces a new problem,the controller placement problem(CPP).The main goals to be addressed in a controller placement problem are determining the number of controllers required in the network and determining where the controllers should be deployed in the network.Solving this problem helps to apply SDN to actual network environment.This thesis introduces the controller placement problem in SDN.This thesis models the packet processing procedure in SDN as a queue system,by analysing the queue system,this thesis shows how to determine the number of controllers required in a network.On the other hand,this thesis studies the reliability model of controller placement problem,and introduces an algorithm based on particle swarm algorithm(PSO),which effectively optimizes the reliability in controller placement problem.Finally,this thesis introduces a multi-objective optimization algorithm based on genetic algorithm(GA)for controller placement problem,which can optimize multiple objectives of the problem,and the results get good performance,the algorithm can also achieve dynamic deployment.