A Novel Scheme for Controller Selection in Software-Defined Internet-of-Things (SD-IoT)

The software-defined networking (SDN) standard decouples the data and control planes. SDN is used in the Internet of Things (IoT) due to its programmability, central view and deployment of innovative protocols, and is known as SD-IoT. However, in SD-IoT, controller selection has never been studied. Controllers control the network and react to dynamic changes in SD-IoT. As sensors communicate frequently with the controller in SD-IoT, there is a degradation in performance with scalability and an increase in flow requests. Hence, the controller performance and selection are critical for SD-IoT. However, one controller's support for certain functions is high while another's is poor. There are various SD-IoT controllers, and choosing the best one might be a multi-criteria choice. An analytical network decision making process- (ANDP) based technique is employed here to identify feature-based optimal controllers in SD-IoT. The experimental analysis quantifies the high-weight controller from the feature-based comparison. An ANDP-based feature-based controller selection strategy is suggested, which selects the controller with the best feature set first, before comparing performance. This paper's main contribution is to evaluate the ANDP for SD-IoT controller selection based on its features and performance validation in the SD-IoT environment. The simulation results suggest that the proposed controller outperforms the controller selected with previous schemes. Choosing an optimal controller in SD-IoT reduces the delay in both normal and heavy traffic scenarios. The suggested controller also increases throughput while using the central processing unit (CPU) efficiently and reduces the recovery latency in case of failures in the network.