| With the popularity and development of Internet,internet system are becoming more complex and difficult to manage.Software-Defined Networking(SDN)is a new type of network architecture that is expected to change the existing network dilemma,which has attracted extensive research attention.The core idea of the software-defined network is to adopt a framework in which the control plane and the data plane are separated to realize flexible control of network data forwarding.Different from the traditional network,the control plane is the core of the SDN network.The optimal deployment of controllers involves network delay,load and network security.Similar to the traditional network,the SDN network also faces the network failure in the event of emergency,or network robustness problems with link failures.Therefore,it is of great significance to study the optimal deployment of controllers to improve the robustness of SDN networks.This paper focuses on the optimal controller deployment considering network delay and load balancing in k-links failure scenario of SDN network,referred to as Robust Control Placement for k-links failure(k-RCP).In order to solve the k-RCP problem,we first propose a heuristic method with a lower time complexity to solve the k-RCP problem.To further improve the performance of solving the k-RCP problem,a robust controller based on the primal-dual algorithm is proposed.The main research work of this paper includes the following aspects:(1)The optimal controller deployment problem considering network delay and load balancing in k-links failure scenario is proposed for the first time.Integer linear programming(ILP)is established for the k-RCP problem.By optimizing the configuration of controller resources,the model minimizes the cost of the controller and maximizes the transmission efficiency of control information while satisfying the required controller coverage proportion of SDN network in the event of any k-links failure;(2)Two metric for controller robustness and controller information transmission efficiency is proposed.Cpr index measures controller robustness in case of k-links failure.The higher the value of Cpr,the more robust the controller deployment results are.TE index measures the information transmission efficiency.The higher TE value,the easier control information is transmitted.(3)A heuristic algorithm for finite controller reconstruction based on GN algorithm is proposed.The algorithm takes into account the controller coverage proportion and transmission efficiency,and significantly reduces the time complexity of Branch and Bound algorithm to solve k-RCP problem.Simulation experiments show that the heuristic algorithm can significantly reduce complexity of solving the k-RCP problem and effectively improve the survivability of SDN network.(4)In order to further improve the performance of heuristic algorithm,this paper combines remote node extraction,clustering and primal-dual algorithm to propose a k-RCP approximation algorithm.Firstly,the k-RCP approximation algorithm modifies the instance,extracts outliers node in SDN and simulates the worst scenario of k-links failure,then relaxes the integer linear programming model established in this paper into linear programming,and then obtains its dual programming model.Secondly,using the primal-dual approximate algorithm gets the controller deployment strategy.Simulation experiments show that the k-RCP algorithm can effectively realize the optimal deployment of SDN network controllers with considering both network delay and load balancing,and can significantly reduce the time complexity of controller deployment.In order to effectively improve the availability of SDN network in face of links failure,this paper proposes a heuristic algorithm for finite controller resource reconstruction and a k-RCP approximation algorithm.After algorithm analysis and comparison experiments,the heuristic algorithm proposed in this paper can effectively reduce the time complexity for k-RCP problem.On this basis,the k-RCP approximation algorithm further improves the performance of heuristic algorithm to achieve better SDN controller deployment capability considering robustness,efficiency,and load balancing. |
PDF Full Text Request