This section describes the environment that must be in place for the clustering to function as a whole—that is, for the clustering backend, the application server, and the connecting network share. On the hardware side, the most important requirements are an NVIDIA GPU with sufficient VRAM, as well as enough CPU cores, RAM, and disk space for the image and catalog data; the share should be approximately five times the size of the catalog. On the software side, a Linux environment with Docker version 23 or higher, NVIDIA drivers, and the NVIDIA Container Toolkit is required; Windows is supported via Docker Desktop or WSL. The share containing the catalog data must be accessible to both servers and the container starts; certificate and model mounts are required only for HTTPS or individual presets. Internet access is not required, as the image is delivered as an offline tarball .