The Portal is where you configure things. The Gateway is what actually touches your data. They never share credentials.
The Gateway polls the Portal for pipeline definitions and schedules, then connects directly to Fusion and your SQL Server using credentials that never leave your network.
Build the SQL in the Portal's query editor. Run it against a small, hardcoded row limit to confirm it returns what you expect — nothing touches your production tenant yet.
Pull your SQL Server table's schema through the Gateway, then map each query column to a target field, with type conversion and any value transformations applied.
Pick from Insert Only, Full Replace, Soft Deactivate & Merge, Delete-Match & Insert, or write Custom SQL for the merge step — the wizard generates the rest.
Publishing creates a real, named BI Publisher report in your Fusion environment's dedicated custom folder — a standard SQL report, not a workaround.
Set a recurrence, and the Gateway picks up the schedule on its next sync. Runs happen locally, in your time zone, with retry and row-level error handling built in.
Every run is logged — success, partial success with skipped rows, or failure with a clear reason. Unschedule or detach a pipeline any time; you can always re-publish it later.
SQL Server is the common target, but not the only one. The Gateway can also be configured to load into other databases, or to export each run straight to plain spreadsheet files — the pipeline, mapping, and schedule you build stay exactly the same.