Name: Michael Sidorov Degree: PhD Supervisor: Prof. Ofer Hadar Subject: Video QoE Prediction in Encrypted Data Traffic

Name: Michael Sidorov Degree: PhD Supervisor: Prof. Ofer Hadar Subject: Video QoE Prediction in Encrypted Data Traffic Abstract: Video traffic has become the dominant form of data traversing the Internet. Estimating users' video quality of experience (QoE) is challenging, as it depends on many subjective and poorly defined factors such as mood, personal preferences, and individual expectations. When content providers (CPs) address the QoE estimation problem, they typically have access to both the transmitted and the received data, enabling direct comparison and evaluation. Internet service providers (ISPs), on the other hand, traditionally rely on packet sniffing tools (e.g., Wireshark) and deep packet inspection (DPI) techniques to analyze traffic. However, the widespread adoption of the end-to-end encryption for privacy protection has significantly limited the effectiveness of DPI-based approaches. In this seminar, we present several models developed to address this challenge for two widely used video conferencing platforms, Zoom and WhatsApp. These models enable ISPs to estimate users' QoE using encrypted traffic only, without access to payload data.