Architectures
An architecture is the high-level shape of a stack: how many gateways, in what roles, on what network layout. The slugs mirror Ignition's documented system architectures — Basic, Scale Out, Hub and Spoke. You pick one at the start of the wizard, or name it directly with ignition-stack create <name> --arch <slug>. Every architecture resolves to the same kind of project; they differ only in the gateways and topology they set up. Architectures are editable presets, not forks: the wizard's composer lets you attach services per gateway from one as a starting point.
The architectures
| Architecture | Gateways | Network split | Edge default | Use it for |
|---|---|---|---|---|
| basic | 1 | off | none | The default starter stack: one gateway + a database |
| scale-out | 2 (frontend + backend) | on | none | A two-tier frontend/backend demo |
| hub-and-spoke | 1 hub + N spokes | off | spokes | A central hub aggregating many edge sites |
Every architecture defaults its database to Postgres and offers the optional reverse-proxy scaffold.
Choosing an architecture
- Start with basic if you want one gateway and a working historian. It is the lightest stack and the baseline everything else extends.
- Choose scale-out to show a frontend/backend split across the gateway network, the canonical multi-gateway field shape.
- Choose hub-and-spoke to show one hub aggregating many Edge spokes. Watch the RAM advisory as the spoke count grows.
Configuration record
The architecture you pick is saved in the stack's configuration record. Pass that record back to ignition-stack create <name> -f to recreate or clone the stack — the positional name overrides the recorded one, so the same file can produce a differently-named project.