Adaptive population-based multi-objective optimization in SDN controllers for cost optimization

Abstract

In Wireless Sensor Networks, Software Defined Networks (SDN) provide a logically centralized control plane as a potential means of streamlining network management (WSNs). The employment of several SDN controllers to build a physically distributed SDN is a common tactic to boost speed, expand scalability, and offer fault tolerance. However, the deployment of many controllers results in increased synchronization and deployment expenses. Therefore, selecting the optimal location for SDN controllers to improve WSN performance is a research issue. In this paper, the multi-objective optimization problem known as the controller placement problem (CPP) is initially formulated. Cost, time, and reliability are just a few of the restraints that are taken into consideration in this regard. In addition, a new Adaptive Population-Based Cuckoo Optimization (APB-CO) for optimal controller placement is implemented. In the end, APB-CO performs experiments to validate the efficacy by analyzing Sync (7.5), Coverage (47), Controller Cost (4.8), and Fitness (0.6983) for the 100th node variation at network 1. The proposed model obtained the controller cost as 34.4, compared to the existing method such as Simulated Annealing (44.3) and Greedy Approach (42.6).