Research On Key Technologies Of Controller Load Balancing In Software Defined Network

In recent years,Software-Defined Networking(SDN)is an emerging paradigm for network design and management.By providing network programmability and separation of control and data planes,SDN provides significant features such as simplified and centralized management,reduced complexity,and accelerated innovation.With SDN,the control and management of network equipment is performed by centralized software called controllers.It solves the problems of management difficulty and complicated configuration in the traditional network,and has been widely concerned by industry and academia.With the continuous increase of network traffic and the gradual expansion of the topology,a single SDN controller is subject to many limitations such as scalability and reliability when managing the network.Therefore,the industry proposes to implement a flexible SDN control plane through multi-controller multi-domain planning.For a multi-controller SDN network,due to the dynamic changes in network traffic,the unreasonable deployment of multi-controllers,and the static connection relationship between the switch and the controller,the processing capabilities between the various controllers vary widely,which can easily cause overload or Light-loaded controllers create load imbalances that severely degrade SDN network performance.Therefore,in large-scale SDN,how to balance the load of multiple controllers has gradually become a hot issue for scholars in recent years.In terms of current research progress,multi-controller load balancing is mainly implemented in the network in the following two ways: static mode,balanced controller load deployment by optimizing controller deployment,and dynamic method,balanced controller load migration through switch migration.However,the above solutions still have the following problems to be improved:(1)At the controller deployment level,as the network scale increases or the problem becomes more complicated,it takes longer calculation time to obtain the best solution.Especially in virtual network slicing,a large number of network devices joining and leaving the virtual network correspond to changes in resource demand and supply.Moreover,the existing methods cannot determine the appropriate controller location when an emergency request is issued due to the occurrence of certain accidents and emergencies(such as earthquakes);(2)at the switch migration level,moving in and out The choice of domain is easy to fall into a local optimum,without considering the integrity,too many controllers will increase the network overhead.At the same time,the switch migration mechanism does not take into account the dynamic changes of the migration threshold and the secondary migration caused by excessive switch migration.This article relies on the National Natural Science Foundation of China-"Research on the Generalized Robust Control Mechanism in Information and Communication Networks".Taking multi-controller load balancing technology as the research goal.This paper improves and optimizes the current static and dynamic load balancing solutions to improve the performance of load balancing among multiple controllers.The main research work of this article is as follows:1.Aiming at the problem of inability to determine the proper controller position in the controller deployment process,this paper proposes a controller deployment method based on graph theory.In large networks,the controller deployment method can also shorten the length of the control path and keep the controller load low.2.In view of the switch migration level existing in the switch migration process,the choice of moving in and out of the sub-domain is easy to fall into the problem of local optimization.In this paper,the main expenses in the network are integrated,and the optimal migration domain selection algorithm is designed by applying the idea of biological genetics.Obtain optimal migration into and out of the domain;at the same time,analogous to the phenomenon of population migration,apply the lifetime and elimination mechanism to the switch,design the switch competition migration algorithm,and balance the number of subdomain switches.The simulation results show that the selection process of the migration domain is optimized,ensuring a balanced distribution of the controller load.