Network Traffic in the Metaverse: The Case of Social VR

The Metaverse connects our physical reality with virtual worlds. Social VR platforms facilitate the creation of such virtual worlds, enabling activities such as interactive teaching, conferences, and community gatherings. These activities can be performed in mixed mode, with some participants physically present in the same location. In this paper, we evaluate the feasibility of such mixed-mode events by studying three leading social VR platforms. We uncover the network traffic patterns generated by these platforms, which affect the user experience when multiple users share the same network. We explore the traffic patterns to show that model loading creates a significant overhead and impacts user quality of experience. When the number of simultaneously connected users increases, some operations lead to network congestion that degrades or even interrupts service for most users. From these observations, we derive concrete action points to develop social VR platforms that enable mixed-mode events over the same network.