Controller Placement Algorithm In Software Defined Networking

As a new network architecture,Software Defined Networking(SDN)is the mainstream development direction of the future network.To further improve the scalability and reliability,network architecture composed of multiple distributed controllers is widely used in SDN.However,the SDN controller placement becomes a pressing problem in the distributed controllers architecture.In this thesis,the controller placement algorithms in SDN are studied and the main contents are as follows:The concept,the structure and the key technologies of SDN are introduced.The current research works on controller placement algorithms in SDN are summarized.Moreover,the difficulties and challenges of controller placement are analyzed.The existing controller placement algorithms fail to comprehensively consider the differences of SDN switch flow,control link characteristics and controller capacity.A joint controller placement and capacity matching algorithm based on control plane delay optimization is proposed.Considering the constraints of switch delay,controller capacity,and the association between controller and switch,joint controller placement and capacity matching problem is formulated as an optimization problem which achieves the control plane delay minimization.Since the optimization problem is a nonlinear integer programming problem,it is difficult to solve directly.Then,the optimization problem is transformed into two subproblems,i.e.,controller-switch association subproblem and controller capacity matching subproblem.By applying improved K-means algorithm and Dijkstra algorithm to solve the controller-switch association subproblem,and using the Kuhn-Munkres algorithm to solve the controller capacity matching subproblem,the corresponding optimal controller deployment and capacity matching strategy can be obtained.For the network scenario with multiple controller failures resulting in severely limited control plane performance and increased energy consumption due to backup controllers.An elastic controller placement algorithm based on energy consumption optimization is proposed in this thesis.Particularly,based on the switch flow characteristics,the energy consumption of the switch is evaluated in detail.In this thesis,the total energy consumption of the network is defined as the energy consumption of the switch,the master controllers and the backup controllers.The elastic controller placement problem is formulated as an optimization problem which achieves the energy consumption minimization.Since the optimization problem is a complex and nonlinear integer programming problem,it is difficult to solve directly.This thesis applies simulated annealing algorithm to solve the optimization problem,so as to determine the placement strategy of the master controllers and corresponding backup controllers in the network.