Ouva platform is end-to-end containerized. The server component that distributes data and allows configuration of the system can be run on any Linux, Windows, IBM mainframe, MacOS system.
The processing services that heavily utilize Machine Learning models to analyze incoming video streams can be run on any CUDA-capable GPU cluster (on-premise, enterprise-hosted or cloud with NVIDIA A6000 recommended). As Ouva is built on official CUDA containers, a single image requires approximately 2GB of storage.
Ouva is Kubernetes compatible for automated deployment, scaling and management of the entire platform, so you can test in one patient room and scale to hundreds in no time.