Encrypted traffic classification is a crucial part of privacy-preserving research. With the great success of artificial intelligence technology in fields such as image recognition and natural language processing, how to classify encrypted traffic based on AI technology has become an attractive topic in information security. With good generalization ability and high training accuracy, pre-training-based encrypted traffic classification methods have become the first option. The accuracy of this type of method depends highly on the fine-tuning model. However, it is a challenge for existing fine-tuned models to effectively integrate the representation of packet and byte features extracted via pre-training. A novel fine-tuning model, LAMBERT, is proposed in this article. By introducing an attention mechanism to capture the relationship between BiGRU and byte sequences, LAMBERT not only effectively improves the sequence loss phenomenon of BiGRU but also improves the processing performance of encrypted stream classification. LAMBERT can quickly and accurately classify multiple types of encrypted traffic. The experimental results show that our model performs well on datasets with uneven sample distribution, no pre-training, and large sample classification. LAMBERT was tested on four datasets, namely, ISCX-VPN-Service, ISCX-VPN-APP, USTC-TFC and CSTNET-TLS 1.3, and the F1 scores reached 99.15%, 99.52%, 99.30%, and 97.41%, respectively.
4 December 2023 in Environmental Science and Pollution Research
While packet encryption technology provides user privacy protection, it also brings challenges to traffic classification. Traditional traffic classification technologies based on packet fields or payload data content are becoming powerless in the face of encrypted traffic. Although traffic encryption technology can hide data content, it cannot change the inherent characteristics of application traffic. Therefore, how to classify encrypted traffic by mining traffic features has become a topic of widespread concern. An enormous amount of research shows that machine learning algorithms achieve good results in encrypted traffic classification. However, there is still room for improvement. For example, deep learning methods require a large amount of labeled training data, and it is difficult for pre-trained models to capture the associated features of long sequences of packet. This paper propose a Bidirectional encrypted traffic classification model based on BERT and BiLSTM(Bi-ETC), by inserting [CLS]Token at the connection of Token sequence between BERT and BiLSTM, strengthens BiLSTM's focus on packet-level features while maintaining BilSTM's capture of Token sequence context. Experimental results show that the Bi-ETC model is superior to all proposed classification methods on the ISCX VPN dataset, reaching 99.43%in F1 score and 99.7%in accuracy.