作者: Jérôme François , Isabelle Chrisment , Thibault Cholez , Wazen M. Shbair
DOI:
关键词: Identification (information) 、 Handshake 、 Certificate 、 Network management 、 Computer network 、 Quality of service 、 Encryption 、 Network packet 、 Computer science 、 Session (computer science)
摘要: Traffic monitoring is essential for network management tasks that ensure security and QoS. However, the continuous increase of HTTPS traffic undermines effectiveness current service-level can only rely on unreliable parameters from TLS handshake (X.509 certificate, SNI) or must decrypt traffic. We propose a new machine learning-based method to identify services without decryption. By extracting statistical features packets small number application data packets, we very early in session. Extensive experiments performed over significant open dataset show our offers good accuracy prototype implementation confirms identification satisfied.