What’s Next#

Now that you understand the OSMO deployment architecture, you’re ready to begin deploying OSMO.

Ready to Begin?#

Select the deployment model that fits your needs and environment.

Quickstart Deployment

Run the complete OSMO control plane, compute plane, and a GPU workflow on a local NVIDIA GPU workstation.

Self-contained Deployment

Run the control plane, compute plane, and required dependencies in one non-cloud Kubernetes cluster for edge and local environments.

Single-plane Deployment

Deploy the service and backend operator in the same Kubernetes cluster for testing, development, or evaluation.

Split-plane Infrastructure

Prepare infrastructure for control and compute planes deployed on separate clusters.

Use the following table to compare dependencies, target environments, and backend scale.

Deployment	Dependencies	Environment	Backends	Recommended for
Quickstart Deployment	Included (non-production)	Local workstation	1	Evaluation and development
Self-contained Deployment	Included	Edge or on-premises	1	Production when the complete stack must remain local
Single-plane Deployment	External	Edge, on-premises, or cloud	1	Production with separately managed dependencies
Split-plane Deployment	External	Cloud, on-premises, or hybrid	Many	Production across multiple compute clusters