We’re designing the integration architecture for our Honeywell MES 2023.1 cloud deployment and debating between real-time API integration vs scheduled batch imports for work order management.
Our ERP pushes about 200-300 work orders daily. The team is split - developers prefer REST API for real-time updates, but operations wants batch imports for better control and API error handling predictability. We’re also considering a hybrid integration approach.
What are the practical tradeoffs? How does batch import scalability compare to API performance in cloud environments? Anyone running hybrid models successfully?
We use real-time API integration and it works well for our volume (similar to yours). The key is implementing proper error handling with retry logic and dead-letter queues. API gives us immediate visibility when orders are created in ERP. Batch imports create delays that frustrated our production planners.
The validation checkpoint is a good point. Could we achieve that with API integration by implementing a staging area in MES where orders land first and require approval before release? That might give us real-time benefits with operational control. Has anyone implemented that pattern?
From an operations perspective, batch imports provide much better control. We can review and validate orders before they hit the shop floor. With API integration, bad data from ERP flows directly into MES, and we’ve had issues where incorrect routing or BOM data caused production chaos. Batch gives us a validation checkpoint.
We run a hybrid model - API for standard orders, batch for bulk imports and complex orders requiring validation. The hybrid integration works well because you can route different order types through appropriate channels. Standard repeating orders go via API for speed. New product orders or engineering changes go through batch validation. This requires more complex integration logic but provides flexibility.