A SURVEY ON DEEP LEARNING TECHNIQUES FOR DARKNET TRAFFIC MALWARE DETECTION

Abstract

Encrypted data that is sent over Tor or a VPN is referred to as "darknet traffic." In order to identify network activity brought on by a cyberattack, it is crucial to be able to recognise, find, and explain darknet activities. Cyberattacks that pose a serious danger to network security and management can be effectively observed through darknet monitoring and classification. Despite the fact that many surveys were conducted on network traffic classification using machine learning techniques, only a small number of researchers have the means to review their work on classifying network traffic using deep learning techniques. In this article, we introduce a novel idea for research on the classification of malware used in darknet traffic that uses Deep Learning techniques, which will aid researchers in improving their surveys. First, based on the survey, we chose a few techniques, including Deep Neural Network (DNN), Convolution Neural Network (CNN), Recurrent Neural Network (RNN), and AutoEncoder (AE) that have proven to be more effective in recent study. Second, the dataset CIC-DarkNet-2020, which contains a variety of darknet activities including VPN and TOR traffic, is used to develop the models. Finally, after analysing the information, we discovered the best Deep Learning Model for classifying Darknet traffic, which has the potential to enhance the efficiency of malware variant detection using constrained system and network resources. Index Terms – Darknet, DNN, CNN, RNN, AE, Deep Learning, Classifying Darknet.