GLR, an OpenGL render server facility

In particular, off-loaded rendering using GLR permits expensive high-end graphics hardware to be shared and amortized across a group of users on a local area network. Hybrid GLR programs for applications like computer-aided design and scientific visualization can use local OpenGL rendering for dynamic, interactive 3D graphics, then use GLR to render higher-quality static scenes (enabling higher tessellation, high-quality texturing, antialiasing, etc.) with sub-second latency.

GLR's approach is novel because access to the rendering hardware is virtualized using render intervals allowing guaranteed access to the frame buffer resources for a client-requested duration. GLR is implemented as a client library and a specialized X window manager that schedules render intervals. Unlike a traditional window manager, the GLR window manager policy controls the mapping and unmapping of windows to effect the render interval policy. GLR demonstrates that the X Window System model is adaptable for uses other than the traditional window system.