Flow-based anomaly intrusion detection using machine learning model with software defined networking for OpenFlow network

Abstract

Moving towards recent technologies, Software Defined Networking (SDN) produces a promising network framework to combine the overall network management system with network programming. It gives a more effective tracking system towards the data center. By centralized system and symmetric controller, it prevents security cracks from creating new threats during OpenFlow packet transmission with vulnerabilities. It creates more interest to the researchers to work towards Flow-based SDN for the priority-driven algorithm in anomaly intruder detection. In this paper, we made a study towards a priority-based model using SDN to control the flow of data packets over the network, gives assurance to the bandwidth enforcement, and reallocation is made through virtual circuits. The network behavior of the system is continuously monitored through the machine learning model for normal and abnormal traffic data transmission to detect anomaly intruders. Flow-based machine learning (ML) model with SDN act as an intelligent system to limits the throughput virtually through the flow of reserved bandwidth and make use of extra bandwidth, which presents more than the utilization bandwidth for priority-based applications with minimal cost while compared with the traditional methods. The proposed work also compared with the schemes available at the network to produce outcomes with fast routing and the fault tolerance of existing networks to overcome the gap open at the security of the SDN architecture to detect and identify vulnerabilities.