A scalable cluster-rendering architecture for immersive virtual environments

Complex virtual environments often require computational resources exceeding the capabilities of a single machine. Furthermore immersive visualization can exploit multiple displays fostering the needing of computational power. We hereby present a system, called XVR Network Renderer, allowing rendering load to be distributed throughout a cluster of workstations operating concurrently. The proposed solution consists in a set of software modules structured as a single master multiple-slaves architecture. The master software intercepts all the graphical commands performed by an OpenGL application, without any modification of the source code. The commands are then streamed and executed individually by each slave client. The Network Renderer can be seen as a virtual OpenGL context with high capabilities. The system can be configured to work with a wide range of complex visualization setups, like CAVEs, automatically handling stereoscopy, performing perspective corrections and managing projection-related common problems. Any number of displays can be simultaneously managed by the cluster.