RCPFH: Reliable controller placement in software-defined networks using fuzzy systems and a modified walrus optimization algorithm

Abstract

Software-Defined Networking (SDN) is a novel network architecture that separates the control plane from the data plane, enabling centralized and programmable management of network resources. One of the key challenges in SDN is determining the optimal number and locations of controllers, called the Controller Placement Problem (CPP), to ensure balanced load distribution, minimal latency, and high network reliability. This paper introduces a novel three-phase approach called Reliable Controller Placement using Fuzzy Logic and Metaheuristic Algorithms (RCPFH), which efficiently optimizes controller placement in SDN environments. In the first phase, the approach employs a fuzzy logic system guided by Levy Flight parameters to estimate the optimal number of controllers by evaluating critical factors such as energy consumption, congestion levels, and load variance across the network. The second phase utilizes a Modified Walrus Optimization Algorithm to identify the most suitable controller positions, considering path reliability, processing capacity, and propagation delay. Finally, in the third phase, backup controllers are selected to enhance system reliability in the event of controller failure. The proposed RCPFH framework is evaluated using four real-world network topologies from the ZOO Topology dataset. Comparative experiments with state-of-the-art approaches demonstrate significant performance improvements: up to a 38 % reduction in energy consumption, an 11 % decrease in load variance, a 36 % increase in network availability, a 17 % reduction in average latency, and a 15 % decrease in link failure rate. These results validate the effectiveness of RCPFH in optimizing SDN performance while maintaining robustness and operational efficiency.