Your Edge, Your Apps: Hosting SCADA, Soft PLC and Python Apps on Factorian

The industrial edge is turning into an app platform. Controllers, gateways and edge computers are no longer closed boxes with a fixed feature list. They are Linux machines that run containers, and the value of a plant increasingly depends on which apps can run on them, and how easily those apps can be provisioned, updated and trusted.

Factorian approaches this from the data side. A Factorian Edge node or cluster is a process historian and a managed container platform in one: it stores and contextualizes your time-series data, and it hosts the applications that need that data, whether they come from Factorian, a partner or your own team. This post looks at the trend, what a Factorian cluster can host, and how partners can build on it.

The Edge Is Becoming an App Platform

The major automation vendors have reached the same conclusion: open, containerized edge platforms let customers add capability without replacing hardware.

• Bosch Rexroth ctrlX: ctrlX OS is a Linux-based, hardware-independent operating system with an app store where Rexroth and its partners publish apps, from motion and vision to Node-RED and databases. Since 2022 it has also been available on third-party hardware.
• Siemens Industrial Edge: apps are packaged and run as containers, and Industrial Edge Management deploys and manages them across edge devices. Partners offer tested third-party apps through its marketplace.
• Rockwell Automation: the OptixEdge appliance runs customers’ own Docker containers alongside FactoryTalk Optix on a secured, Linux-based operating system, so collection, processing and visualization sit on one device.

The pattern is consistent: Linux, containers, an app store or management console, and a partner ecosystem. What differs is the starting point.

Descriptions summarize each vendor’s public product information.

Inside a Factorian Edge Node

A Factorian node runs a Linux host, Factorian container orchestration and the Factorian core services: API, Pub/Sub, Historian and Core. Everything else, shown as dashed boxes, is optional and yours. Three nodes join into a cluster that shares one zero-touch orchestration layer, with all data replicated across them in near real time.

1. Hosting Third-Party Containerized Apps

Because the cluster is a container platform, anything packaged as a container can run next to the historian.

• SCADA: run a containerized SCADA or visualization application on the same cluster that archives its data, with no separate server to maintain.
• Soft PLC: host a soft PLC for supervisory or non-critical control workloads close to the process.
• Python apps: build self-service data apps and Streamlit dashboards for your engineers.
• ML models: train and run machine learning models at the shop floor, where the data is, without moving it to a public cloud.
• Dashboards and marketplace apps: add Grafana dashboards and partner apps without touching the core services.

2. Provisioning That Feels Like Plugging in a Device

• Order and plug in: Factorian clusters arrive as managed, pre-built hardware with orchestration already in place.
• Scale by adding nodes: connect another node to add vCPU, RAM and storage, and the orchestration layer takes care of the rest.
• Right-sized hardware: nodes start at 16+ GB RAM and 2+ TB of storage. Smaller clusters use rugged, fanless DIN-rail units for the plant floor, larger sites use rack hardware, and you can also use your own bare metal VMs.
• Fleet-wide control: run an autonomous cluster at every site, and update and manage applications centrally.

3. The Historian as the Data Foundation

Hosting apps only pays off if they get good data. Every app on a Factorian cluster sits beside a lossless, contextualized time-series historian, where values keep their source timestamps and live inside an asset, batch and phase hierarchy.

• Open interfaces: apps read data through REST, MQTT, Python and MCP, so nothing depends on a proprietary format.
• Context included: recipe, batch, phases, events and control limits arrive with the values, so an ML model sees batches and phases rather than anonymous tags.
• No data movement: hosting the app beside the data avoids latency and cloud egress fees.
• AI-ready: the MCP server lets AI agents and chat assistants query the same data your apps use.

4. A Platform Partners Can Build On

Customers rarely want one vendor for every application. Factorian is built so partners can develop on the platform and host their apps on it.

• Software vendors: package your SCADA, analytics, quality or digital-twin application as a container and deploy it to Factorian clusters wherever your customers run them.
• Integrators and implementation partners: deliver connectivity, historian and applications on one managed stack, with fewer moving parts and a shorter path to validation.
• OEMs and machine builders: ship machines with a Factorian Edge node in the panel, so each machine arrives data-ready and able to run your own apps.

If you build industrial software or machines, talk to us about partnering.

5. Resilient and Compliant by Design

• Autonomous operation: each cluster keeps acquiring, storing and serving data if the link to the corporate network drops.
• High availability: nodes replicate all data in near real time, and if one node fails the others carry on with no manual failover.
• Built for regulated sites: 21 CFR Part 11 compliant, with an immutable audit trail and read-only enterprise interfaces.

The Bottom Line