Research On Switch Migration And Real-time Routing Updates In Software-defined Networking

Software-defined network(SDN)is a new type of network structure composed of a variety of network technologies.It consists of a data plane at the bottom,a control plane at the middle level,and an application plane at the top level.It is flexible and agile.Automatically optimize the utilization of network resources,especially suitable for applications with constantly changing traffic requirements in the network.The difference between software-defined networks and traditional networks is that SDN separates the control plane and data plane of network devices.Therefore,according to real-time dynamic network requirements,users can adjust the entire network structure or expand network functions through simple programming operations.The basic equipment at the bottom layer can always play a role,saving a lot of economic overhead without the need to replace new equipment.At the same time,this will greatly shorten the iterative cycle of the network architecture.Nowadays,with the advent of 5G networks,the network has flooded with massive amounts of data,and the agile characteristics of software-defined networks can better meet the different needs of different applications in the 5G era.Therefore,software-defined network technology has become one of the key technologies of 5G networks.Nowadays,SDN is mainly used in two network environments,namely the data center and the core network.Based on this research background,this article mainly discusses two issues:(1)Software-defined network supports smart city perception,which is called software-defined Io T(SD-Io T)network.Aiming at the problem of the need to deploy and implement load balancing of multiple controllers in SD-Io T networks in order to improve network performance,this article first introduces the current research status of switch migration in SD-Io T,and describes the existing migration strategies,and then proposes Switch Migration with the Minimal Global Migration Cost(MMM)strategy under SD-Io T network,while considering the migration cost and the impact of migration on the load of the current domain and other control domains,select the switch that needs to be migrated in the high load domain.In addition,the Lyapunov optimization method is adopted to ensure that the controller obtains the best choice for the migration switch under the condition of long-term stable load.The experimental results prove that compared with the existing related strategies,the MMM strategy always has the smallest migration cost and the total network load,and at the same time it can better ensure the load balance between the controllers.(2)Aiming at the requirement of low delay and real-time data transmission in data center network environment,the controller needs to constantly update forwarding rules to improve network performance.This paper first introduces the existing routing update strategies,and then proposes data flow selection and real-time routing update strategies based on Fuzzy theory(Fuzzy Quantified Based Flow Selection and Route Update Scheme,FANS)in data center.Firstly,based on fuzzy theory and link available capacity,FANS decides the data stream to be updated,and then further mines the data stream that can be updated synchronously according to the limit of link capacity and switch storage capacity,so as to better achieve the purpose of real-time routing update.The experimental results show that FANS can meet the requirements of load balancing while greatly reducing the routing update delay.