Load Balancing Algorithm in SDN Multicontroller Network
Alexander Zhetov, Valeria Kaijalainen, Vladislav Mogilatov, Artem Volkov, Ammar Muthanna · 2023
In the field of networking and services 2030, software-defined networking is an effective tool for managing network resources, which allows separate control and data transmission. However, with the increasing number of devices and applications running on SDN networks and using the server interface to communicate with the network, there is a challenge to efficiently distribute the load among the controllers. The purpose of the work is to determine the method of load distribution on controllers in a software-defined network on the server interface to reduce the load on the central processing unit (CPU) of network elements. Three load balancing methods are presented to solve the problem: proxy server, clustering, and orchestration. The CPU utilization value of the network elements was used to evaluate the effectiveness of the proposed methods. While modeling the network loading processes, the clustering method gives a positive result - requests sent to one controller do not affect other controllers, or their impact does not exceed the error, the load on processors of other controllers does not rise above 20%, and the background load is 5-20%. The chosen method is an optimal solution for planning, construction of software-defined networks that does not require additional tools for implementation.