Key takeaways
- SCADA is real-time process visibility and supervision, MES is execution and recording of production on the floor, ERP is the planning and financial record of the business.
- Choose a single source system for each type of data; other systems copy that data rather than recreate it.
- Writing PLC counters straight to ERP is usually not enough; order context and duplicate checking must be set up somewhere.
- An integration layer does not replace these three systems; it makes the data flow between them defined and traceable.
In a manufacturing plant the same event shows up in three systems in three forms: a counter value in SCADA, a production-order movement in MES, a material document in ERP. Integrations built without settling which one is the source of what end up with the same data contradicting itself in three places. This article explains the roles and the boundary between them.
Three systems, three questions
- SCADA: “What is happening in the process right now?” Real-time values, alarms, trends and operator intervention.
- MES: “How did this production order actually run on the floor?” Which order was produced on which line, with which material, in what time; what was the quality result.
- ERP: “What happened for the business and what should happen?” Planning, purchasing, inventory, cost and the financial record.
This split runs parallel to the split in the ISA-95 levels: SCADA sits at level 2, MES at level 3 and ERP at level 4. Each system has its own time scale and outage tolerance: a SCADA screen must refresh within seconds; an ERP document recorded minutes or hours later may be fine for the business.
Which system owns which data?
Choosing a source system for each type of data is one of the most important steps of integration design. A typical distribution is:
- Real-time process values and alarms: PLC and SCADA. A historian or SQL log can be used for history.
- Production-order status, batches, operation history and traceability: MES; or, if there is none, whichever system takes on that job.
- Material master data, order header, inventory and financial documents: ERP.
This distribution does not have to be the same in every plant; what matters is that the decision is written down and that every system treats data it does not own as a copy. A copy that is not updated goes stale; that is why the question of which system tells the truth matters.
Flows: down and up
Downwards: ERP opens a production order; MES splits it into lines and operations; SCADA and PLC receive the recipe or target values. Upwards: PLC counters and states go to SCADA and MES, and from MES the production confirmation and material consumption go to ERP.
The most critical step of the upward flow is the confirmation: a production event on the floor is turned into a transaction (for example a production confirmation) that fits the ERP’s business logic. Which interface to use and how to prevent a duplicate confirmation must be settled up front; the guide on production integration with SAP covers these decisions.
Common mistakes
- PLC straight to ERP. Sending the counter to ERP raw produces wrong postings without order context, unit conversion and reset handling.
- Two owners of the same data. If both SCADA and MES calculate downtime, reports will disagree. Decide which one is the source.
- Losing time. If the time the record was written is used instead of the event’s source time as data flows upward, shift and production reports shift after a replay following an outage.
- Not designing outage behaviour. If what happens to data when the destination goes into maintenance is not decided up front, loss or duplicate records appear (Store & Forward, idempotency).
- Burying integration inside one of the systems. If logic is spread across SCADA scripts or ERP customisations it becomes hard to monitor and change.
Where the integration layer sits
An integration layer does not replace these systems; it makes the data flow between them visible, defined and traceable. In this role OPC Router reads PLC and OPC UA sources, transforms the data and transfers it to destinations such as SQL, MQTT, REST or SAP. Which system is source and which is consumer, and the transformation in each flow (the data contract), we write down together in the project.
The result is not one big system but systems that each do their own job with clearly defined boundaries between them. To try this structure with a pilot see the pilot plan guide and the integration planner.
Frequently asked questions
What is the difference between SCADA and MES?
SCADA manages real-time process visibility and supervision; MES manages the execution and recording of production orders. One answers “what is happening in the process now”, the other “how did this order run”.
Can SCADA connect directly to ERP without an MES?
Technically yes; but the jobs an MES takes on (order context, batch, confirmation logic, duplicate checking) must be done elsewhere. That work can be built in an integration layer; the scope is defined for each project.
Does OPC Router replace an MES?
No. OPC Router is an integration platform; it does not provide an MES’s business logic such as planning, quality and batch management. It builds the data flow between shop-floor systems and MES, ERP and databases.