In today’s increasingly interconnected world, cybersecurity has become more critical than ever, with the rise of sophisticated threats and a growing number of bad actors targeting vulnerable networks.

Software Defined Networks (SDN), with their decoupling of control and forwarding planes, present unique security challenges by introducing novel attack vectors that traditional defenses may not address. Although some SDN-specific datasets exist, they have notable limitations, and many Intrusion Detection Systems (IDS) still rely on conventional network datasets, which overlook SDN’s distinctive architecture. This research focuses on creating a custom SDN intrusion dataset, captured from a hybrid physical/virtual tested, to bridge this gap. The resulting dataset will facilitate the training of state-of-the-art machine learning models capable of identifying both traditional and SDN-specific threats, thereby significantly enhancing network security.