Intelligent Intrusion Detection System For SDN (Software Defined Networking)

Software-Defined Networking (SDN) offers centralized and programmable control.It does this by decoupling the control plane from the data plane.That setup simplifies management.It also improves scalability and traffic optimization.Still, this centralization turns the SDN controller into a single high value target.As a result, the controller becomes vulnerable to Denial of Service (DoS) and Distributed DoS (DDoS) attacks.Conventional Intrusion Detection Systems (IDS) depend on fixed rules and thresholds.Those systems often fail to tell legitimate high volume traffic apart from malicious floods.That leads to poor adaptability in dynamic network environments.In this work, we design an intelligent IDS for SDN.It uses Machine Learning (ML) integrated tightly with the Ryu controller.The system relies on Mininet to generate both benign and attack traffic.That includes ICMP pings, TCP/UDP flows, and hping3 based DoS scenarios.From these flows, we collect features like packet count, byte count, and duration.Those features train a Random Forest classifier.The classifier identifies malicious flows accurately.During operation, the IDS works in real time.It continuously collects flow statistics from switches.It classifies each flow.Then it automatically installs drop rules for traffic seen as malicious.A Flask based web dashboard provides live visualization of traffic volume and security alerts to the user.Experimental evaluation shows the proposed system reaches more than 98% detection accuracy.