New Method Measures Cloud Gaming Experience in Real-Time
Researchers from the University of New South Wales have developed a pioneering method to measure cloud gaming user experience. This novel approach, published in IEEE Access, uses network traffic analysis to identify contextual factors like game title and player activity stage.
The team, comprising Yifan Wang, Minzhao Lyu, and Vijay Sivaraman, has created a system that classifies cloud gaming contexts directly from network traffic. This system, when deployed within an internet service provider's network, provided detailed insights into bandwidth consumption and experience levels across diverse gameplay contexts.
The method employs machine learning techniques to classify traffic based on game type, user activity, and platform. Remarkably, it can identify the specific game being played and the player's activity within the first few seconds of a session. The Random Forest model proved most effective, achieving up to 96.5% accuracy in both game title and gameplay activity pattern classification.
This innovative method offers a real-time assessment of player engagement, classifying player activity into active, passive, and idle stages. It particularly highlights the transition from active gameplay to an idle state as a key indicator of user activity in cloud gaming traffic. This breakthrough enables network providers to optimize services and enhance the cloud gaming experience.
