Enhanced Controller Placement Strategies in SDN: A Gravitational Search Algorithm Perspective
Khushi Kalra, Abha Kumari, Akash Yadav, Susham Biswas · 2025
The Controller Placement Problem (CPP) in Software-defined Networking (SDN) is outlined as finding the optimal number of controllers, their placement and identification of the set of switches to be managed by each controller. The CPP is pivotal for ensuring optimal network performance, scalability, and fault tolerance. This paper presents a novel approach to solving the CPP using the Gravitational Search Algorithm (GSA), a heuristic inspired by gravitational laws and mass interactions. Our study explores the effectiveness of GSA in minimizing both switch-to-controller latency and controller-to-controller latency within SDN environments. Simulation results demonstrate that GSA provides efficient solutions for CPP, outperforming traditional placement strategies, including Particle Swarm Optimization (PSO), across several critical metrics. GSA’s gravitational-based search mechanism provides a superior balance between exploration and exploitation, allowing it to escape local optima and achieve more globally optimal controller placements than PSO. Additionally, GSA’s adaptive pull mechanism better distributes controllers, ensuring load balance and minimizing both switch-to-controller and controller-to-controller latency, particularly in large-scale and complex network topologies. These results affirm GSA as a promising alternative for CPP in SDN, offering robust performance and scalability.