Abstract
Software-Defined Networking (SDN) has emerged as a transformative technology that significantly enhances network management and control capabilities, especially in the complex and varied landscape of Internet of Things (IoT) networks. As IoT devices proliferate, the demand for more scalable, flexible, and efficient network topologies becomes imperative. This research explores the comparative performance of centralized and distributed Software-Defined Networking (SDN) architectures in the context of Internet of Things (IoT) networks. Leveraging tools such as mininet-wifi for network emulation, Floodlight as the SDN controller, Iperf for performance metrics, and MQTT protocol for IoT device communication, we systematically analyzed three different network topologies: bus, mesh, and tree. Each experiment was subjected to a consistent traffic load of 50Mb to assess network robustness. Distributed networks utilized dual-controller configurations managed through Docker containers, whereas centralized networks employed a single-controller setup. Performance metrics such as latency, throughput, jitter, and controller resource consumption (CPU and memory) were meticulously recorded. The results indicate nuanced differences between the centralized and distributed approaches, with the distributed configuration showing generally improved fault tolerance and recovery capabilities, as well as reduced controller CPU usage. The study underscores the potential of distributed SDN controllers in enhancing the resilience and redundancy of IoT networks, suggesting a scalable solution for future IoT infrastructure development. The findings offer crucial insights into the design of robust, efficient, and scalable SDN environments for the ever-growing IoT landscape.
DOI: 10.61416/ceai.v26i3.9164