Architecture overview¶
shaide is built in layers, each an independent Pulumi program.
| Layer | Page |
|---|---|
| Internal OCI registry | Platform services |
| Istio, Gateway API, GAIE | Gateway and routing |
| vLLM + llm-d model stacks | Model serving |
| shaide server, control panel | Application layer |
| MCP datasources | MCP datasources |
| Loki, Grafana, Alloy | Observability |
| The interactive installer | Installer |
Deployment order¶
1. Internal OCI registry infra/cloud-harbor
2. Gateway + Istio infra/gateway-provider
3. Model serving app_serving
4. Application layer app_shaide
5. Optional add-ons app_mcp, monitoring
Ordering is strict: serving depends on the gateway, which depends on the registry.
Traffic¶
See Traffic flow.
Supporting detail¶
- Model deployment flow - definition to running service
- Model storage - weight storage, pull and caching
