Research on the Application of Transformer in Encrypted Traffic Recognition

Encrypted traffic recognition is a crucial research area in network security. However, traditional recognition methods cannot meet practical needs due to the concealment and diversity of encrypted traffic. The emergence of deep learning techniques provides new ideas for encrypted traffic recognition. As a neural network model based on the attention mechanism, transformer exhibits excellent sequence modeling ability and has achieved successful applications in various domains. In this paper, we introduce the current research status of applying transformer in the field of encrypted traffic recognition, analyze the technical development line, propose the framework of encrypted traffic recognition built upon hybrid learning, and look forward to the development direction of transformer in the domain of encrypted traffic recognition, so as to provide reference and reference for further improving the accuracy and practicality of encrypted traffic recognition.