IFS Cloud and IBM Sterling integration

What moves between them

Data flows from IBM Sterling into IFS Cloud. When a trading partner sends an EDI 850 (purchase order) or 810 (invoice) to Sterling, ml-connector polls Sterling's mailbox at a configured interval (every 5 to 30 minutes), detects the new message, extracts the EDI payload, and translates it into an IFS PurchaseOrderSet or PostingProposalHeadSet record with the customer-provided company code. Line item quantities and prices map to the corresponding GL account dimensions based on the supplier and item codes in the EDI. Functional 997 acknowledgements are posted back to Sterling as a workflow message. Reference data such as suppliers and GL accounts is read from IFS first to validate that inbound EDI references exist before creating financial records.

How ml-connector handles it

ml-connector stores IFS tenant URL, OAuth credentials, and company code configuration, and stores Sterling instance host, port, and Basic Auth credentials, all encrypted at rest. On each poll cycle, it calls Sterling's mailbox API with the configured authentication, extracts the EDI payload, and parses it using industry-standard X12 or EDIFACT rules. For each PO or invoice line, it validates that the supplier exists in IFS using the EDI supplier identifier, queries IFS GL accounts to find the matching cost center or GL account, and composes an IFS OData POST request to PostingProposalHeadSet or PurchaseOrderSet. Before posting, it reads the ETag value from the target resource (to satisfy IFS ETag requirements for mutations) and includes that ETag in the PATCH or POST operation. If Sterling rate-limits the request (HTTP 429), ml-connector backs off and retries. If IFS returns a 5000-element exceeded error on an expanded query, ml-connector paginates separately and makes multiple requests. After successful creation in IFS, it posts a 997 functional acknowledgement back to Sterling marking the EDI message as processed. A full audit trail records the source EDI, the IFS record created, any validation errors, and the timestamp of each operation.

A real-world example

A mid-sized discrete manufacturer receives purchase orders and invoices from 40 trading partners via EDI through IBM Sterling. Before the integration, the procurement team downloaded EDI files daily from Sterling, manually mapped supplier and GL codes to IFS, and re-entered the data by hand, a process that took 2 to 3 hours daily and introduced typos and duplicate orders. With IFS and Sterling connected via ml-connector, each inbound EDI document lands in IFS with the correct company code and GL account within minutes, the supplier is validated before the PO is created, and the trading partner receives a functional acknowledgement automatically. The procurement team now spends 15 minutes daily reviewing items that failed validation and correcting supplier master data, and the remaining 100+ orders per week flow through without intervention.

What you can do

• •Poll IBM Sterling mailbox for inbound EDI purchase orders and invoices on a schedule you control.
• •Translate EDI 850 (purchase orders) and 810 (invoices) into IFS Cloud PurchaseOrderSet and PostingProposalHeadSet records with the correct company code.
• •Validate supplier and GL account dimensions in IFS before creating financial records, rejecting EDI documents that reference non-existent suppliers.
• •Post EDI 997 functional acknowledgements back to IBM Sterling to confirm message receipt and processing status.
• •Maintain a full audit trail of every EDI message received, IFS record created, and validation error, with replay capability for failed operations.

Questions

How does ml-connector handle the different authentication methods required by IFS and Sterling?

ml-connector stores IFS OAuth 2.0 client credentials and Sterling Basic Auth credentials separately, encrypted at rest. On each poll cycle, it refreshes the IFS OAuth token (lifetime 60 minutes) and calls Sterling's mailbox API with the configured username and password over HTTPS. Both credentials are scoped to the customer's tenant or instance, ensuring isolated access.

What happens if an EDI purchase order references a supplier that does not exist in IFS?

ml-connector validates the supplier against IFS SupplierSet before creating the PurchaseOrderSet record. If the supplier is not found, the EDI document is marked as failed, logged to the audit trail, and the procurement team is notified via alert. The EDI payload is retained so the operator can correct the supplier master data in IFS and replay the order.

Does ml-connector handle the IFS ETag concurrency requirement?

Yes. Before posting or patching an IFS financial record, ml-connector reads the record's @odata.etag value and includes the ETag header in the mutation request. If the ETag has changed due to a concurrent update, IFS returns an HTTP 412 error; ml-connector retries the read and update cycle to ensure the record is current before proceeding.