A Unified Namespace for a multi-line plant
PLCs, MES and the historian are linked point to point, and every new system means another integration.
- Data in
- PLCs (OPC UA) · MES (SQL) · Historian
- What it does
- A model per asset type and an instance per machine; pipelines publish the result to the namespace on the embedded broker.
- Outcome
- MES, SCADA, analytics and cloud all subscribe to one topic tree.
Clean data for analytics and AI
Data scientists spend most of their time decoding tag lists and fixing units.
- Data in
- Vibration and temperature tags · Work-order context
- What it does
- Conditions with deadband and aggregates, time-aligns, validates against the model and delivers to Snowflake, Databricks or GoodData.
- Outcome
- Models start from validated, contextualised payloads instead of raw tags, and only valuable data goes to the cloud.
Electronic batch records and traceability
Batch information is scattered across MES, PLCs and quality systems and compiled by hand for reports.
- Data in
- MES · PLC process values · Quality system
- What it does
- A “Batch” model merges these sources into one object; a pipeline runs on the event when the batch completes.
- Outcome
- Reporting reads from one place and the manual compilation step disappears.
Governed access for AI agents
An agent needs machine data, but you do not want to give it direct access to PLCs or SCADA.
- Data in
- Connected systems · Modeled instances
- What it does
- The Industrial MCP Server exposes selected pipelines as named tools; the agent can call only those.
- Outcome
- Agents use scoped, documented tools instead of raw credentials.
An equipment maker wants to collect data securely from machines in the field and offer cloud services.
- Data in
- Field machine controllers
- What it does
- The hub runs on or next to the machine, and one model produces the same structure at every customer site.
- Outcome
- A consistent data feed for services such as remote monitoring and digital twins.
