Research on Network Encrypted Traffic Detection Technology Based on CNN+LSTM
Yuqing Cui, Aihua Li · 2024
In recent years, network security has received more and more attention in the identification of encrypted traffic, resulting in rapid progress in traffic classification research. Traffic protection measures such as network traffic obfuscation technology are constantly introduced into the Internet, which brings great challenges to efficient traffic classification. This paper proposes a method for classifying network traffic by utilizing a CNN (Convolutional Neural Networks) and LSTM (Long Short-Term Memory) model. The CNN model learns the spatial features of encrypted traffic, while the LSTM model analyzes the sequential features. The output results of both models are fused to classify the traffic. Experimental results show that the proposed method achieves good accuracy, recall rate, and precision when compared to a one-dimensional CNN.