Scalability of Wi-Fi Performance in Virtual Reality Scenarios

The adoption of Virtual Reality (VR) applications in Wi-Fi networks intensifies each year. VR applications impose strict Quality of Service (QoS) requirements, necessitating low latency and high throughput. Meeting VR QoS requirements in Wi-Fi networks is especially challenging due to unpredictable channel fading and interference. This paper presents a comprehensive scalability study of Wi-Fi performance in multi-user VR scenarios. We investigate whether simply increasing Access Point (AP) capabilities, specifically through Multi-User MIMO (MU-MIMO), is sufficient to support dense VR deployments. To this end, we developed a high-fidelity simulation framework in ns-3 to estimate the network capacity when serving VR traffic. Our analysis meticulously evaluates the impact of critical factors, including the number of antennas at the AP and STAs, MU-MIMO scheduling algorithms, channel sounding period, and different channel conditions. The results reveal a critical finding: Scalability is not linear. In particular, doubling AP antennas from 8 to 16 yields only a 35% gain in capacity under typical conditions, not the 100% linear scaling one might expect. We identify and analyze the key bottlenecks that prevent performance from scaling indefinitely with an increased number of AP antennas, providing crucial insights for the design of next-generation Wi-Fi systems aimed at supporting the Metaverse and future immersive VR applications.