ANALYSIS OF MODERN LOAD BALANCING METHODS IN SDN NETWORKS

Abstract

This article analyses modern approaches to load balancing in software-defined networks (SDN). Various balancing methods aimed at achieving an even distribution of traffic between servers in order to ensure greater network performance and reliability are investigated.
The balancing methods at the transport level are analysed. In particular, approaches such as Round Robin, Least Connections and Weighted Round Robin are considered, which contribute to the efficient load distribution. Next, we consider application-level methods, such as 
the use of HTTP proxies, which allows for more intelligent traffic distribution depending on the characteristics of applications and users. The article also analyses the use of balancing methods at the SDN controller level, which centrally manage traffic distribution in programmable networks.
Particular attention is paid to the prospects for improving the load balancing strategy. The importance of using machine learning methods to optimise the balancing process is highlighted. The application of these methods will allow the system to adapt to changing network 
and load conditions, which will improve the efficiency and optimality of traffic distribution. The possibility of taking into account the specific requirements of users and applications to achieve the best load balancing is also considered. It is noted that the transition to dynamic load management can contribute to the optimal use of network resources and prevent congestion. In addition, the analysis takes into account the possibility of using context and additional information about applications to better understand their behaviour and needs. This can lead to more intelligent and context-aware balancing strategies that improve traffic distribution and quality of service.
Despite the progress made in the field of load balancing in software-defined networks (SDNs), there are important challenges and promising areas for further research. The development of machine learning and dynamic control methods and integration with the latest technologies can help make the balancing strategy even more efficient and reliable for various scenarios.