An intelligent way for optimal controller placements in software-defined–IoT networks for smart cities

Abstract

Software Defined Networks (SDNs) is modern network paradigm that separately manages the control logic-plane and hardware data-plane. The programmable network SDN centrally manages different smart devices connected in the networks. These smart devices are facilitated by Internet of Things (IoT) and collect real time data for network devices. One of the emerging applications of IoT network is Smart City which purpose is to manage the city without human interventions. The rapid growths of IoT devices like actuators and sensors in smart city are generating large amount of data and increasing the traffic in network. To maintain the quality of services (QoS) of smart city IoT networks, the SDN characteristic is one way to provide better services to network users. Traditional IoT network having the challenges like, access delays, security and reliability issues. The SDN enabled IoT networks are capable to overcome these challenges and improves the network performance. One single controller is not sufficient to manage all IoT devices in smart city networks. So, multiple controllers are required to handle Software Defined-IoT Networks (SDN-IoT). Different metaheuristic algorithms are implemented to overcome controller placements problems (CPP) but as the number of controllers increases the performance of approaches degrades due to premature conversions. To better balance the exploration and exploitation features an intelligent Hybrid Differential Evolution and Whale Optimization (DEWO) Algorithm is proposed that optimizes the controller placements in SDN-IoT enabled smart city networks. The algorithm searches the optimal locations of controllers and balance the switch loads in respect of minimum latency. The algorithm also minimizes the link failure and evaluate the minimum end to end delay. The proposed hybrid algorithm is evaluated up to 40 controllers with other state of art method and able to achieve better results in comparison with other metaheuristic algorithms.