Secure SDN Traffic based on Machine Learning Classifier

Nowadays, the majority of human activities are carried out utilizing a variety of services or applications that rely on the local and Internet connectivity services provided by private or public networks. With the developments in Machine Learning and Software Defined Networking, traffic classification has become an essential study subject. As a consequence of the segregation of control and data planes, Software Defined Networks have some security flaws. To cope with malicious code in SDN, certain operational security techniques have been devised. In this paper, a machine learning model, supervised, was utilized to identify normal and malicious traffic flows. While, normal traffic were generated using Internet traffic generator, malicious traffic were accomplish by Scapy and Python. The main network features of the OpenFlow flow table such as Packets count, bytes counts, packet rates, byte rate for forward and revers flows, were extracted. The combination of good ML classifier and dataset produced the greatest accuracy rate over 99% in DDoS attack detection, according to the results. Further to the main aim, the presented approach could be utilized to classify different traffic flows with the purpose of balance and priorities the important traffic