Performance Evaluation of Software-Defined Networking for Secure and Efficient Computer Communication Networks

Abstract

Software-Defined Networking (SDN) separates network control from packet forwarding and introduces programmable, logically centralized management. This architecture can improve visibility, policy enforcement, and traffic engineering, but it also creates performance and security concerns around controller latency, control-plane saturation, and concentrated attack surfaces. This methodology paper proposes a reproducible framework for evaluating whether SDN can provide secure and efficient computer communication compared with a conventional distributed forwarding baseline. The study uses Mininet, Open vSwitch, OpenFlow 1.3, and a Ryu controller to emulate a multi-switch enterprise network under normal traffic, congestion, and distributed denial-of-service (DDoS) conditions. Performance is assessed using throughput, round-trip time, packet loss, flow-setup delay, and controller resource utilization, while security is measured through attack-detection rate, mitigation time, and legitimate-traffic preservation. A lightweight controller application monitors flow statistics and installs mitigation rules when abnormal source rates and flow growth exceed defined thresholds. The proposed analysis uses repeated trials and confidence intervals to compare conventional networking, basic SDN, and security-enabled SDN. Illustrative results show the expected trade-off: security logic introduces a small control overhead under benign traffic, but centralized detection and rapid rule installation substantially preserve legitimate throughput and reduce packet loss during attacks. The framework is suitable for laboratory-scale evaluation and can be extended to multi-controller, wide-area, or hardware testbeds.