Scenario Overview - On Cloud Model Asset Publishing

This scenario guides Platform Operators through preparing runtime assets after cloud-resource onboarding, then assembling them into deployable model assets for downstream deployment and calling.

Applicable Roles

• Platform Operator

Scenario Goals

• Runtime images can be pulled in the target cloud environment.
• Inference frameworks contain the correct service ports and startup commands.
• Model assets can be selected by End Users during quick deployment.

Scenario Flow

Main path: Register a runtime image -> Build an inference framework -> Create a model asset -> Verify user deployment availability

Stage	Key Result
1. Register the image	The target cloud environment can pull a controlled runtime-image version
2. Build the framework	Image, startup command, port, and framework type are compatible
3. Create the model asset	Meta-model, deployment point, cloud model, compute plan, and output protocol are linked
4. Verify availability	End Users see the model and available plan in quick deployment

Before You Start

• Cloud resources and authorization are ready.
• The image address, framework type, model information, and output API rules are prepared.

Recommended Reading Order

1. Maintain runtime images.
2. Maintain inference frameworks.
3. Assemble a deployable model asset in the model library.

Document Index

Document	Description
Runtime Images	Maintain repository address, tag, framework type, and built-in state
Inference Frameworks	Bind images, ports, and startup commands
Model Assets	Combine meta-model, deployment point, compute plan, and output configuration

Related Scenarios

• On Cloud Resource Access (upstream): provides the cloud platform, account, resource pool, and authorization required here; meta-model data comes from Publish Models Preconfiguration
• On Cloud Model Deployment and Calling (downstream): creates deployment instances from the model-library assets prepared here
• My Deployments User Manual (downstream): reviews state, events, and monitoring for deployed models
• Platform Governance and Access Control (cross-cutting): constrains model-asset visibility and calling scope by tenant and tenant

Completion Checklist

Purpose: These are the scenario exit criteria. Use them to confirm that the result is observable and reviewable before continuing. If any check fails, return to the relevant feature guide and follow its troubleshooting section.

Check	Pass Criteria
1	The runtime image can be pulled in the target cloud environment, with the correct framework type and version.
2	Framework startup command, port, and output protocol match the model.
3	The model asset links its meta-model, cloud deployment point, compute plan, and output configuration.
4	A End User can see the model and available plan in quick deployment.

Troubleshooting

Symptom	Check First
Image or framework cannot be selected	Asset state, framework type, region, and association
Model is missing in user deployment	Publication state, deployment scope, authorization, and tenant visibility
Deployed API is unavailable	Startup command, port, output configuration, and health state