VPN traffic identification method and device based on user behavior model

Through the identification method based on the user behavior model, combined with the BERT and BiLSTM models, user behavior characteristics are extracted and feature fusion is carried out, which solves the problems of high false alarm rate and insufficient generalization ability in the existing VPN encryption traffic recognition technology, and achieves the correct Accurate recognition of VPN encrypted traffic and the objectivity and accuracy of recognition results.

CN119094380APending Publication Date: 2024-12-06EVERSEC BEIJING TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410896271.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-04
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing VPN encryption traffic recognition technology has problems such as high false alarm rate, insufficient generalization ability, relying on manual design features and difficulty in adapting to multiple scenarios, especially in large-scale labeling of data and small sample environments.

Method used

Using the recognition method based on the user behavior model, by building a traffic feature library, user behavior events, paths and retained feature vectors are extracted, combined with the BERT pre-trained model and BiLSTM model, feature fusion and context word embedding are carried out, and user behavior model is constructed for Identify VPN traffic and combine it with statistical analysis module to improve the objectivity and accuracy of the identification.

Benefits of technology

Accurate recognition of general-purpose VPN encrypted traffic is achieved, the generalization and robustness of the model is improved, the false positive rate is reduced, and the objectivity and accuracy of traffic recognition is enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119094380A_ABST
    Figure CN119094380A_ABST
Patent Text Reader

Abstract

The invention provides a VPN traffic identification method and equipment based on a user behavior model, and the method comprises the steps: analyzing user behaviors in network traffic based on a preset traffic feature library, obtaining user behavior monitoring logs, carrying out the grouping of the user behavior monitoring logs according to user IDs, analyzing the behaviors of different users in a specific time window, and carrying out the analysis of the user behaviors in the specific time window. According to the method, feature data related to VPN traffic usage behaviors of users are extracted and calculated, VPN traffic in network traffic is detected by calculating user behavior feature vectors and using a BERT pre-training model and a bidirectional LSTM model, and Session analysis is performed on the overall traffic of the users in combination with a statistical analysis module, so that the objectivity and accuracy of traffic identification are improved.
Need to check novelty before this filing date? Find Prior Art