Abstract
This study investigates the process that manages load on software-defined networks (SDN) controllers in the fog layer of the mobile high-density Internet of Things. The work considers SDN core controllers that handle OpenFlow switches located on mobile fog devices. The task addressed is to improve the efficiency of the ecosystem by reducing the load imbalance in SDN controllers and network latency. The integration of software-defined networks allows for flexible centralized traffic management, but SDN controllers become a bottleneck and are subject to load imbalance. This study proposes an architecture for the mobile high-density Internet of Things ecosystem focused on the use of SDN controllers in the fog layer. A feature of the architecture is a three-level hierarchical principle of building the SDN core. Based on this architecture, a mathematical model of the process of dynamic load balancing between SDN controllers was built. The model has made it possible to formalize the requirements for objective functions when optimizing the load balance of SDN controllers and minimizing network latency. The use of the model enabled devising a method for dynamic load redistribution on SDN core controllers. The main difference of the method is the use of a modified dung beetle optimization (DBO) algorithm with dynamic adjustment of the fitness function weights. The application of the DBO algorithm has made it possible to reduce the load dispersion in controllers by up to 30% compared to the classical genetic algorithm. The final solution is generated approximately 40% faster while the average query processing delay is reduced by 20%