<![CDATA[The increasing complexity of modern computer networks requires flexible, efficient, and manageable network architectures. Software Defined Networking has emerged as a networking paradigm that separates the control plane from the data plane, enabling centralized and dynamic network management. This study aims to analyze network performance in SDN architecture using a Systematic Literature Review (SLR) approach. The methodology follows the PRISMA framework to ensure a systematic and transparent literature selection process. Relevant studies were collected from major scientific databases and analyzed based on performance parameters, including throughput, latency, jitter, and scalability. The results indicate that SDN implementation improves throughput by 20%–84.6% with an average of 53.2%, reduces latency by 15%–39.3% with an average of 27.4%, reduces jitter by 10%–31.5% with an average of 18.8%, and improves scalability by 18%–42% with an average of 29.6%. Routing optimization and programmable data plane approaches contributed the most to throughput improvement, while machine learning-based approaches were more effective in reducing latency and optimizing network traffic adaptively. However, no single approach can be considered universally optimal, as effectiveness depends on the implementation context. Furthermore, the dominance of simulation-based studies highlights the need for further validation in real-world network environments. This study provides a comprehensive synthesis to support SDN performance optimization strategies and future research directions.]]>
