AI-based Network Flooding Attack Detection in SDN using Multiple Learning Models and Controller

The management of the classic IP networks, which are still commonly used today, has grown challenging. The complexity of the network is increased because IT operators must access the network devices independently and use vendor-specific instructions to create any high-level network policies, such as Quality of Service (QoS) or routing policy. A networking paradigm called Software-Defined Networking (SDN) offers a centralized method of network management and control. SDN offers flexibility and programmability in managing and controlling network infrastructure. In this study, we present a framework for simulating an SDN environment using Mininet and Ryu controller to evaluate network policies and security mechanisms. The framework incorporates modules for simulating network traffic using hping3, collecting traffic data, and applying AI algorithms for detecting network flooding attacks. The collected traffic data is pre-processed and used to train AI models capable of identifying various types of network flooding attacks. The best-performing model is then deployed in the controller for realtime attack detection. By employing this framework, network administrators can enhance their understanding of SDN behavior and develop effective strategies for securing their networks against network flooding attacks.