Glossary designed
This is the first pass. The glossary should become the shared language layer for users, developers, operators, product, and architecture reviewers.
| Term | Meaning |
|---|---|
| AI Cloud | Product and platform name for the shared GPU, app, IAM, billing, runtime, evidence, and developer platform. |
| Allocation | A GPU capacity reservation that moves through requested, provisioning, active, releasing, and terminal states. |
| App instance | A project-owned app runtime resource created from a catalog app and version. |
| App SDK | Developer-facing product surface for building apps on AI Cloud. |
| Catalog | Product surface for discovering GPU SKUs, apps, and versions. |
| Control plane | AI Cloud services that own identity, lifecycle, billing, audit, and orchestration. |
| Evidence | Durable proof artifact or command result attached to a task, release, or readiness decision. |
| GPUaaS | The GPU capacity/runtime product domain inside AI Cloud. Keep this term for resource families, legacy repo paths, and GPU-specific internals; use AI Cloud for the product brand. |
| Manifest | App/version metadata package describing runtime expectations, config, placement, artifacts, and capabilities. |
| Platform foundation | Shared-services and production-readiness architecture track. |
| Release ring | Ordered rollout segment for validating changes before broad production exposure. |
| Tenant | Organization-level ownership scope above projects. |