Firewall for Data Plane for Cyber-Attacks Detection and Prevention by Deep Learning in SDN

As the networking landscape shifts from conventional networking to automated networking, software defined networking will play an increasingly important role. The new network design will prioritise data protection. The purpose of this paper is to use a variety of deep learning approaches to categorise traffic as either benign or malicious using the attributes provided in the dataset. When a dataset has been preprocessed, classification is performed to sort the traffic in one of the classes. By using the Stacked Deep Auto-Encoder technique described in the study, we achieved a 99.75% accuracy rate. With this goal accomplished, deep learning methods were successfully used to classify network traffic.