Implementation And Optimization Of SDN Multi-Controller Architecture Based On IPv6

The Internet has continuously become an indispensable part of people's life.As the rapid development of information technology and the continues to expand in data scale,the limitations of existing network architecture restrict the development of advanced information technology,such as Internet of Things,Big Data and Cloud Computing.A few requirements and methods have been proposed by researchers to improve the existing network architecture.As a new network architecture,Software Defined Network has received great attention from related fields,which is based on the separation of the control and forwarding planes in networks.SDN architecture combined with a novel OpenFlow protocol has provided a new scheme to solve the drawbacks of existing network architecture.The core property of SDN is located in control plane that affect the overall performance of the networks.In a large-scale network architecture,high workload of controller will lead to decline of network service capabilities and even paralysis,which can be solved by deploying multiple controllers.It is necessary for new network architecture to compatible with IPv6 which is new and well-rounded in order to meet the rapid development of network technology.OpenFlow,the famous southbound protocol in SDN,already supports the operation of the IPv6 protocol.But the 128-bit address of IPv6 will affect the Network forwarding performance.We propose using Flow lable of IPv6 protocol to optimize the matching process of flow table of OpenFlow,and the method will improve overall performance of the SDN multi-controller architecture.This paper provide a Mininet-based SDN simulation environment and use some remote RYU controllers to build SDN multi-controller architecture.And introduce Zookeeper system to implements the collaborative work of multiple controllers and the management of network-wide information.Each switch is connected to the controller through a unique IPv6 address,and distinguish primary or secondary between different controllers to avoid conflicts.In order to optimize the flow table,the IPv6 flow label is used to implement the matching of the OpenFlow flow table.The network paralysis coping mechanism of multi-controller architecture is proved by experiments.In the two cases of using IPv6 flow labels or not using IPv6 flow labels,the forwarding performance of the SDN multi-controller architecture is compared by forwarding delay and delay jitter to confirm the availability of SDN multi-controller architecture and the feasibility of related optimization schemes.