Confirm what the machine already exposes
Collect the controller model, communication modules, firmware and existing engineering documentation. Confirm protocol availability and license requirements with the equipment owner. Avoid assuming that an Ethernet socket implies a supported data interface. Identify which connection is already used by an HMI or engineering workstation.
Choose a narrow read-only scope
Select a small tag set that supports one operational question. Start with a conservative polling interval approved by the automation team. Measure communication errors, response times and any observed impact on existing clients. Read-only access can still consume controller resources, so the pilot needs an operational stop condition.
Validate a gateway as a complete dependency
A gateway adds configuration, power, networking and a maintenance responsibility. Record where it buffers data, how it recovers from restart and who owns updates. Test a loss of the upstream system without changing the controller program. Ensure a downstream application failure cannot force an unintended machine action.
Decide using recorded evidence
Document the useful tags, verified data types, observed load and recovery results. Compare the effort of maintaining the interface with alternative approaches. A successful pilot justifies a controlled next step, not immediate plant-wide replication. Keep the original configuration and a tested removal plan until the new path is accepted.
Put it into practice
Use the related tool to check your assumptions, then bring the results to your project discussion.