Oracle Fusion Cloud ERP and SAP Ariba integration

What moves between them

The primary flow moves data from SAP Ariba into Oracle Fusion Cloud ERP. Purchase orders and supplier invoices created or updated in Ariba are polled via async reporting jobs, mapped to Oracle Fusion Cloud ERP cost centers and GL accounts, and posted as procurement records and journal entries. Supplier master records are synced bidirectionally so vendor codes, names, and terms stay aligned. Ariba's reported invoices flow into Oracle Fusion Cloud ERP's AP and accrual accounts, while Oracle Fusion Cloud ERP's journal lines and GL balances remain the source of truth for downstream reporting.

How ml-connector handles it

ml-connector caches separate OAuth2 tokens for each system: Oracle Fusion Cloud ERP's tenant-specific Bearer token and SAP Ariba's Bearer token plus API key, refreshing both before expiry to prevent outages. On the Oracle Fusion Cloud ERP side, the customer provides the pod-specific URL and Identity Domain credentials; ml-connector uses them to authenticate and poll for new invoices and payment records by filtering on LastUpdateDate. On the Ariba side, ml-connector submits async reporting jobs scoped to a date range, polls for job completion via jobId, retrieves paginated results (max 50,000 per call), and maps supplier names and PO line details to Oracle Fusion Cloud ERP dimensions. Because Ariba enforces a maximum one-year window per reporting job, multi-year backfills are submitted as annual chunks. Supplier codes are matched first so invoices land on existing vendors in Oracle Fusion Cloud ERP, and every record carries a full audit trail for reconciliation and replay.

A real-world example

A mid-sized manufacturing company runs Oracle Fusion Cloud ERP for financials and inventory, and uses SAP Ariba for procurement and sourcing across global suppliers. Before the integration, the accounts payable team exported invoices weekly from Ariba, manually matched them to purchase orders in Oracle Fusion Cloud ERP, and entered accrual journals by hand, creating a two-day delay between invoice approval in Ariba and GL posting. With the integration, invoices and POs flow automatically from Ariba into Oracle Fusion Cloud ERP aligned to the correct cost centers and vendors, the AP team receives exceptions only when a match fails, and month-end close no longer requires manual invoice re-entry.

What you can do

• •Sync purchase orders, requisitions, and invoices from SAP Ariba into Oracle Fusion Cloud ERP filtered by update date, with automatic retries and audit trail.
• •Map SAP Ariba supplier codes and names to Oracle Fusion Cloud ERP vendor master records, keeping procurement data consistent across both systems.
• •Post Ariba invoice lines into Oracle Fusion Cloud ERP's AP and accrual GL accounts, aligned to cost centers and ledger dimensions.
• •Bridge OAuth2 authentication for Oracle Fusion Cloud ERP's tenant-specific pod URL and SAP Ariba's dual-credential requirement (bearer token plus API key).
• •Handle Ariba's async reporting job pagination and one-year date-window constraint, automatically splitting multi-year backfills into annual chunks.

Questions

Which direction does data move between Oracle Fusion Cloud ERP and SAP Ariba?

The main flow is from SAP Ariba into Oracle Fusion Cloud ERP. Purchase orders, invoices, and supplier records created or updated in Ariba are polled and posted into Oracle Fusion Cloud ERP's procurement and GL dimensions. Supplier master records can be synced bidirectionally to keep vendor data aligned across both systems.

How does ml-connector handle the authentication difference between the two systems?

Oracle Fusion Cloud ERP uses OAuth2 client credentials against a customer-specific pod URL, issuing Bearer tokens valid for one hour. SAP Ariba requires both an OAuth2 bearer token AND a static API key header on every request. ml-connector caches and refreshes both credential sets independently, presenting the correct headers for each system.

Why must Ariba data be polled in annual chunks for multi-year backfills?

SAP Ariba's async reporting jobs enforce a maximum one-year window between the updatedDateFrom and updatedDateTo query parameters. For historical data spanning multiple years, ml-connector automatically submits separate reporting jobs for each year, polls for completion, and concatenates the results in date order.