Encrypted Network Traffic Analysis

The traditional network monitoring techniques such as deep packet inspecting models do not work for encrypted communications. Machine learning (ML) and deep learning (DL) techniques are designed in the literature to address this issue. In the previous chapter, we have discussed securing TCP/IP across network and transport and application layers through encryption, and we have discussed securing non-TCP/IP communications, together with discussions on the network traffic analysis and challenges involved in it. In this chapter, we discuss the methodology followed for Encrypted Network Traffic Analysis (ENTA), tools available to collect network traffic, different datasets, feature selection and extraction, different techniques for ENTA and performance measures.