Encrypted Flow Intelligence: A Literature Review of AI Models for Traffic-Based Threat Detection

Encrypted Flow Intelligence: A Literature Review of AI Models for Traffic-Based Threat Detection

Prof.Sahana Sharma M                                    Charitha K R                                                      Deeksha B Poojary

Dayananda Sagar Academy                          Dayananda Sagar Academy                                Dayananda Sagar Academy

of Technology and Management,                  of Technology and Management,                            of Technology and Management,

Bangalore.                                                  Bangalore.                                                            Bangalore.

Soujanya V                                                      Pragati Sanjay Hubballi

Dayananda Sagar Academy                                    Dayananda Sagar Academy

of Technology and Management,                               of Technology and Management,

Bangalore.                                                            Bangalo

re.

Abstract

The growth of encrypted online communication has strengthened user privacy but also introduced major obstacles for threat detection systems. Since traditionally intrusion detection relies heavily on inspecting readable data, these systems often fail when faced with encrypted traffic. To address this issue, recent studies have turned toward artificial intelligence, particularly approaches using machine and deep learning, which can infer suspicious behavior without decrypting data. This review consolidates findings from recent literature, evaluating model architectures, training techniques, and detection effectiveness. Emphasis is placed on models that can be deployed in real-world environments while maintaining performance and protecting data confidentiality.

Keywords: Systematic review; encrypted traffic; intrusion detection; deep learning; machine learning; network security.