Real-Time vs Batch: Choosing SAP B1 Sync
Real-time stream or hourly batch pull? We explain how to pick the right refresh rate for SAP B1 data from latency, system load and decision needs.

How fresh should data be? 'Real-time' sounds good, but it is not always needed and not always cheap.
Two models
- Batch: data is pulled at set intervals (hourly, daily). It is simple and predictable and strains the source less; the data can be as stale as the pull interval.
- Real-time / streaming: a change is passed on as soon as it happens. It gives fresh data; setup, monitoring and error handling are harder.
Questions to decide
- Must the decision made with this data happen within minutes, or the next day? A monthly close report does not need second-level data.
- Can the source system take the extra load? Querying a live ERP database often can slow daily work.
- What happens when something goes wrong? In a batch pull it is corrected on the next run; in a stream the risk of silent loss is higher.
- Does fresh data change the number? Documents corrected during the day and cancellations do not settle before the close.
Where does latency come from in SAP B1?
- Pull interval: a document created after the last pull appears on the next run.
- Header-line consistency: a document's header and lines must arrive consistently in the same run.
- Retroactive changes: values such as cost can change afterwards; that is why the history must be reprocessable.
- Warehouse processing time: the transformation time from the pull until it shows in a report.
Where Floemdata stands
Floemdata Analytics pulls from SAP B1 periodically: the setup wizard can install an optional hourly automatic run, and the pipeline runs as a scheduled task. The progress marker (watermark) advances only after the warehouse confirms the load; a failed load retries the same window on the next run. Real-time streaming is not a verified product feature and is not claimed here. The shared-folder model that combines the data of several installations in one panel scans the folder every 60 seconds, according to the documentation.
To get started, write down for each report 'how old can this number be at most?'; for most reports the answer turns out to be 'one day'. Questions: info@floemdata.com.