Comprehensive Analysis of Software-Defined Networking: Evaluating Performance Across Diverse Topologies and Investigating Topology Discovery Protocols

Abstract

: Software-defined networking (SDN) represents an innovative approach to network architecture that enhances control, simplifies complexity, and improves operational efficiencies. This study evaluates the performance metrics of SDN frameworks using the Mininet simulator on virtual machines hosted on a Windows platform. The research objectives include assessing system performance across various predefined network topologies, investigating the impact of switch quantities on network performance, measuring CPU consumption, evaluating RAM demands under different network loads, and analyzing latency in packet transmission. Methods involved creating and testing different network topologies, including basic, hybrid, and custom, with the Mininet simulator. Performance metrics such as CPU and RAM usage, latency, and bandwidth were measured and analyzed. The study also examined the performance and extendibility of the OpenFlow Data Path (OFDP) protocol using the POX controller. Results indicate that balanced tree topologies consume the most CPU and RAM, while linear topologies are more efficient. Random topologies offer adaptability but face connection reliability issues. The POX controller and OFDP protocol effectively manage SDN network scalability. This research aims to analyze performance in a manner consistent with numerous previous studies, underscoring the importance of performance metrics and the scale of the network in determining the efficiency and reliability of SDN implementations. By benchmarking various topologies and protocols, the research offers a valuable reference for both academia and industry, promoting the development of more efficient SDN solutions. Understanding these performance metrics helps network administrators make informed decisions about implementing SDN frameworks to improve network performance and reliability.