In today’s class, we will discuss the paper, Requet: Real-Time QoE Detection for Encrypted YouTube Traffic.
First, we will go through the list of questions from your reading assignment. More specifically, we will learn about various stages of video streaming sessions and related QoE metrics. We will discuss the motivation and key insights from the paper before talking about its data-generation techniques, choice of machine learning model, and results.
The learning objective of today’s class is to understand the broader research area of QoE estimation for encrypted services and how to apply machine learning tools over network data for this task. Though in this lecture, we focus on QoE estimation for video streaming services, we will discuss how these learnings can be generalized to estimate Quality of experiences for other services.
Note: Swaroop and Sabrina will blog the discussion summary for this class.
In next class, we will discuss the paper, Neural Adaptive Video Streaming with Pensieve. Please fill out the reading assignment form before the class.
Note: Sanjay and Roman will write the blog summarizing the next class’ discussions.