Load Balancing in SDN-Enabled WSNs Toward 6G IoE: Partial Cluster Migration Approach

Abstract

The vision for the sixth-generation (6G) network involves the integration of communication and sensing capabilities in internet of everything (IoE), towards enabling broader interconnection in the devices of distributed wireless sensor networks (WSN). Moreover, the merging of SDN policies in 6G IoE-based WSNs i.e. SDN-enable WSN improves the network’s reliability and scalability via integration of sensing and communication (ISAC). It consists of multiple controllers to deploy the control services closer to the data plane for a speedy response through control messages. However, controller placement and load balancing are the major challenges in SDN-enabled WSNs due to the dynamic nature of data plane devices. To address the controller placement problem, an optimal number of controllers is identified using the articulation point method. Furthermore, a nature-inspired cheetah optimization algorithm is proposed for the efficient placement of controllers by considering the latency and synchronization overhead. Moreover, a load-sharing based control node migration (LS-CNM) method is proposed to address the challenges of controller load balancing dynamically. The LS-CNM identifies the overloaded controller and corresponding assistant controller with low utilization. Then, a suitable control node is chosen for partial migration in accordance with the load of the assistant controller. Subsequently, LS-CNM ensures dynamic load balancing by considering threshold loads, intelligent assistant controller selection, and real-time monitoring for effective partial load migration. The proposed LS-CNM scheme is executed on the open network operating system (ONOS) controller and the whole network is simulated in ns-3 simulator. The simulation results of the proposed LS-CNM outperform the state of the art in terms of frequency of controller overload, load variation of each controller, round trip time, and average delay.