Oracle E-Business Suite and SAP Ariba integration

What moves between them

The main flow runs from SAP Ariba into Oracle E-Business Suite. ml-connector submits Ariba reporting jobs for approved purchase orders and invoice headers, downloads the paginated results, and posts them into the EBS purchasing and payables open interface tables for the right operating unit, where the standard import concurrent programs pick them up. Supplier records read from the Ariba Supplier Data API are imported into the EBS supplier interface so vendor master stays current. Supplier registration and qualification updates can flow back from EBS into Ariba through the Supplier Data API, which is the one place Ariba accepts writes over REST. EBS general ledger postings are produced by those imports inside Oracle, so no financial entries are written back into Ariba.

How ml-connector handles it

ml-connector stores both credential sets encrypted. On the Ariba side it fetches an OAuth2 client credentials token, caches it, refreshes it before the one-hour expiry, sends the apiKey header on every call, and appends the realm query parameter. On the EBS side it accepts the full instance URL per customer, obtains a session token, attaches the mandatory operating unit id and responsibility context to every request, and re-authenticates when a call returns 401. Because neither system pushes events for this use case, both sides are polled: Ariba reporting jobs are filtered by an updated date window, capped at the one-year span Ariba enforces, while EBS interface views are filtered on last update date. Ariba purchase orders and invoice headers map to the EBS purchasing and payables open interface columns, and suppliers map by a stable natural key. Because EBS writes are asynchronous, a POST that inserts an interface row is not a finished record; ml-connector confirms the import either by triggering the concurrent program where it is deployed or by polling the base table. Ariba returns HTTP 429 with a remaining-calls header when rate limits are hit, so the connector watches that header and backs off, while EBS has no rate limiter and is kept to low concurrency so a bulk sync does not slow the shared instance. The interface SOURCE column is set to a fixed connector identifier so a re-run does not double-import. Every record carries a full audit trail and can be replayed if a downstream call fails.

A real-world example

A mid-sized manufacturer with about 1,500 employees runs Oracle E-Business Suite R12.2 for financials, payables, and purchasing, and uses SAP Ariba for sourcing and procurement across several buying categories. Before the integration, procurement staff approved purchase orders and processed supplier invoices in Ariba, then a finance clerk re-keyed those purchase orders and invoice headers into Oracle by hand each week and reconciled supplier lists between the two systems by spreadsheet. With Oracle E-Business Suite and SAP Ariba connected, approved purchase orders and invoice headers flow from Ariba into the Oracle open interface tables and import on the normal concurrent program runs, and supplier records stay aligned. The weekly re-keying disappears and month-end close starts with procurement activity already in the ledger.

What you can do

• •Import approved SAP Ariba purchase orders into the Oracle E-Business Suite purchasing open interface for the correct operating unit.
• •Bring SAP Ariba invoice header data into the Oracle payables open interface so Payables Open Interface Import processes it.
• •Keep supplier master data aligned between Oracle E-Business Suite and the SAP Ariba Supplier Data API.
• •Bridge Ariba OAuth2 plus apiKey authentication to the Oracle EBS session token and required operating unit context.
• •Poll both systems on the schedule you set, with retries, duplicate guarding, and a full audit trail on every record.

Questions

Which direction does data move between Oracle E-Business Suite and SAP Ariba?

The main flow is SAP Ariba into Oracle E-Business Suite. Approved purchase orders, invoice headers, and supplier records read from Ariba are posted into the matching Oracle open interface tables for import. Supplier registration and qualification updates can flow back into Ariba through the Supplier Data API, which is the one place Ariba accepts writes over REST, while general ledger entries stay inside Oracle.

Why does the integration poll both systems instead of using webhooks?

Neither system offers a self-service webhook for this use case. SAP Ariba exposes bulk procurement data through async reporting jobs that you submit, poll, and download, and Oracle E-Business Suite has no SaaS-style event push, so its interface views are read on a schedule. ml-connector polls both sides on the cadence you choose, filtering Ariba jobs by an updated date window and EBS views by last update date.

How does ml-connector confirm an invoice or purchase order actually landed in Oracle?

Writes to Oracle E-Business Suite are two-step. A REST post inserts a row into an open interface table, and a concurrent program such as Payables Open Interface Import then validates and imports it. ml-connector treats the insert as incomplete until import is confirmed, either by triggering that concurrent program where it is deployed or by polling the base table, and it sets the interface SOURCE column to a fixed identifier so a re-run does not double-import.