Design And Implementation Of SDN Controller Optimization System Based On Openstack

With the increase in the number of cloud computing tenants and kinds of network service,the traditional cloud computing network has been difficult to meet the network requirements which are changing dynamically.Therefore,the introduction of SDN in the cloud platform to improve network performance has become a common solution.If system architecture engineers can select a SDN controller which is suitable for their own network requirements when deploying the network service of cloud platform,it would greatly improve the overall performance of the cloud platform.In order to help system architecture engineers to select a suitable solution from numerous controllers,this thesis designs an SDN controller optimization system based on OpenStack.The system focuses on performance test of four open source SDN controller(ONOS,Ryu,OVN and OpenDaylight).The test results will be generated into a visual test report,which could help users to complete the selection of controller schemes.Users can customize the test schemes according to their specific network requirements.The test scheme includes four options: controller,the type of tenant network,test scenario and test indicator.In order to test the network performance of OpenStack cloud platform after integrated SDN from different angles and levels,the test indicator is divided into two parts: control plane and data plane.This thesis firstly analyzes the current research of SDN controller and OpenStack,and then briefly introduces the relevant technical theories and architectures.After that,the design and implementation of system are introduced in detail.The system is divided into three parts from top to bottom: access management layer,test task layer and infrastructure layer.Flask and Bootstrap framework are used to implement the access management layer,and ECharts framework is used to visualize the test results.The test task layer is divided into control plane and data plane.The control plane test is mainly completed by invoking OpenStack CLI to send the creation request with time stamp.The data plane test is completed by using Ping and Iperf test tools,after logging in the cloud host from OpenStack node which is as a jump host.The test infrastructure layer uses the Python module Paramiko to remotely log in the OpenStack node for automated deployment of the integrated environment and test environment.Finally,the test of the optimization system are carried out.The test results show that the system has achieved the expected design goals and implemented SDN controller optimization based on the network requirements from users.The test report provided by the system has certain reference significance and application value when users are deploying the network service of private cloud.