Line-side edge
One MaestroHub binary on an industrial PC next to the line reads the machines, keeps the plant's Unified Namespace and feeds MES, ERP and the cloud. This is the smallest deployment and the usual first one.
- Scenario
- 1 of 5
- Shape
- Tier 1 · one machine per plant
- Applies to
- 3.0
Choose it when you have one plant on one network, no Kubernetes skills on site, and Windows or Linux industrial PCs.
- MaestroHub
- Reads machines
- Data
- Buffered on disk
- Management
- Your tooling
What you install
| Level | Component | How it runs | Recommended hardware |
|---|---|---|---|
| L3 | MaestroHub, Tier 1 | One binary as a native service (systemd or a Windows service) | 8 vCPU, 16 GB RAM, 200 GB SSD |
When something fails
| What fails | What happens |
|---|---|
| The MaestroHub process | The service manager restarts it after 5 seconds. Buffered writes survive the restart, because the buffers are files on disk. |
| The uplink to ERP or the cloud | Reading the machines and serving the floor carry on. Writes that fail wait in their Store & Forward buffers. When the 50 GiB disk pool or a buffer's limit is reached, the buffer's drop policy applies (by default, drop the oldest). |
| The PC | Reading stops until the PC is replaced and its configuration restored. Tier 1 has no standby. |
Limits. There is no high availability: one machine is one point of failure. Each further plant is a separate deployment; scenarios 4 and 5 show how to run many.
Previous: Deployment scenarios · Next: Edge plus DMZ hub · How to read the drawings: Deployment scenarios