A Robust Deep Learning-based Approach for Network Traffic Classification using CNNs and RNNs

Abstract

The application of deep learning has become prevalent in the area of network traffic classification. Deep learning has acquired widespread use in network traffic classification, providing a powerful and flexible approach to identify different types of traffic in a network. This research provides a novel deep learning-based technique for network traffic classification. The proposed method leverages both convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to classify network traffic. This study aims to demonstrate the efficiency of deep learning in classifying network traffic and emphasize the potential of using a combination of CNNs and RNNs for this task. The proposed approach first uses CNNs to extract features from the raw network traffic data, and then uses RNNs to classify the extracted features. The proposed approach was evaluated using a recent dataset of network traffic from different sources, such as the CIC-IDS 2022 dataset. The findings suggest that the proposed technique can attain a high classification accuracy rate (98%), and surpasses traditional machine learning-based methods. Additionally, the study demonstrates the ability of the proposed approach to manage extensive and intricate datasets, making it suitable for real-world network traffic classification tasks. Overall, this paper demonstrates the effectiveness of deep learning-based approaches for network traffic classification, and highlights the potential of using a combination of CNNs and RNNs for this task. In future work, the proposed approach could be further improved by incorporating additional deep learning architectures, such as auto encoders and generative models, to extract more meaningful features from the raw network traffic data.