Evaluation of performance and scalability in SDN networks using sFlow

Abstract

Software-Defined Networking (SDN) separates the data plane from the control plane, enabling centralized and flexible network management. Traditional networks face significant challenges in terms of scalability and flexibility; therefore, implementing efficient traffic monitoring techniques is crucial. Using a protocol like sFlow within an SDN network helps optimize performance and resource management, facilitating the analysis of network scalability and efficiency. In this work, we investigate the integration of the sFlow protocol into an SDN architecture to collect real-time network status data and evaluate its scalability. A custom topology was created using the Mininet simulation environment and the Ryu controller, with the sFlow protocol and sFlow-RT deployed to capture traffic metrics, including packet loss, jitter, throughput, and total traffic. The results enable an analysis of how network scalability is affected as the number of devices increases. As more devices connect, network capacity begins to decline, especially when over thirty devices are connected, leading to higher packet loss and reception failures. This behavior heavily depends on how efficiently the controller manages the network.