An optimized recognition algorithm for SSL VPN protocol encrypted traffic

Abstract

With the widespread use of Virtual Private Networks (VPNs), the identification of Secure Sockets Layer (SSL) VPN encrypted traffic has become an important issue. This paper first introduced SSL VPN encrypted traffic and analyzed the flow of its handshake protocol. Then, an improved fingerprint recognition algorithm was designed to identify SSL streams. Capsule Neural Network (CapsNet), an optimized convolutional neural network, was used to recognize SSL VPN. An experimental analysis was carried out on the ISCXVPN2016 dataset. It was found that the recognition accuracy of the proposed method reached up to 99.98% for SSL streams, and the convergence speed was high; the recognition precision reached 98.16%, and the recall rate reached 99.98% for SSL VPNs, both of which were better than the algorithms such as random forest (RF) and C4.5. The experimental results verify the effectiveness of the optimized recognition algorithm for SSL VPN recognition and make some contributions to its application in practice.