Extended Reality Network Traffic & Quality of Experience

The XR Traffic\u2013QoE Dataset provides a comprehensive collection of network-traffic measurements and perceptual QoE annotations spanning Virtual, Augmented, and Mixed Reality applications. Data is captured using an end-to-end XR experimental platform built with the Oculus Quest 2 headset, a traffic-shaping router, and a cloud-rendering device running Virtual Desktop Streamer. For this release, the dataset is provided exclusively in processed comma-separated value (.csv) format, enabling seamless integration into machine-learning pipelines; the raw packet-capture (.pcap) files are already available through the IEEE Dataport repository (https:\/\/ieee-dataport.org\/documents\/metaverse-network-traffic-classification-and-prediction). The dataset includes diverse XR applications\u2014VR gaming, VR video, VR chat\/VoIP, AR, and MR\u2014recorded under five bandwidth regimes (15, 30, 60, 120 Mbps, and adaptive) and evaluated at 60, 90, and 120 Hz display frame rates. Each session is time-aligned with a four-level perceptual QoE scale aligned with ITU-T P.812\/G.1035: Cybersickness (1), Discomfort (2), Tolerable (3), and Comfort (4). This enables holistic analysis of network behavior, system dynamics, and perceptual outcomes, supporting both XR traffic management and QoE prediction research.