Oracle JD Edwards and Adobe Commerce integration

What moves between them

The main flow runs from Adobe Commerce into Oracle JD Edwards. ml-connector reads completed Commerce orders, invoices, and credit memos and posts them into JD Edwards as supplier vouchers or GL journal entries through the matching orchestration, mapped to the right JD Edwards accounts and business units. Customer and item master records flow the same direction so each posted transaction references an address book and item record that already exists in the ERP. Reference data such as accounts and business units is aligned so every journal line lands on a valid JD Edwards dimension. JD Edwards data service tables are read-only, so financial entries are only created through orchestrations, never by direct table writes.

How ml-connector handles it

ml-connector stores both credential sets encrypted. On the Adobe Commerce side it signs each request with OAuth 1.0a on PaaS or refreshes the IMS bearer token before its roughly 24-hour expiry on the Cloud Service. On the JD Edwards side it accepts the full AIS Server URL per customer, since there is no shared base address, posts the username and password to obtain a session token, sends that token on the jde-AIS-Auth header, and re-authenticates automatically on HTTP 444 when the token is invalidated by a server restart. Because JD Edwards has no outbound webhooks, it polls Commerce orders and invoices on a schedule, or consumes the Commerce HMAC SHA256 webhook where the storefront has it enabled, then writes into JD Edwards through orchestrations the customer has imported into Orchestrator Studio. Accounts and business units are mapped first, so every posted line references a valid JD Edwards account. Neither system supports an idempotency key, so ml-connector deduplicates on the Commerce increment_id and dedups its own jobs before posting. It also expects the JD Edwards CNC admin to add its egress IPs to the AIS allowedHosts list, since otherwise the server returns HTTP 405.

A real-world example

A mid-sized industrial parts distributor with about 300 employees sells through an Adobe Commerce B2B storefront and runs Oracle JD Edwards EnterpriseOne for finance and distribution. Before the integration, the accounting team exported a daily order and invoice report from Commerce and re-keyed the sales and refund totals into JD Edwards journal entries, which delayed revenue recognition and caused mismatches between storefront sales and the ledger. With Oracle JD Edwards and Adobe Commerce connected, each completed order and invoice posts into JD Edwards automatically against the correct account and business unit, and customer and item records stay aligned. The accounting close starts with sales already booked, and the manual re-keying step is gone.

What you can do

• •Post completed Adobe Commerce orders and invoices into Oracle JD Edwards as vouchers or GL journal entries through the matching orchestration.
• •Push Commerce credit memos into JD Edwards so refunds are reflected in the ledger.
• •Keep customer and item master records aligned between Adobe Commerce and JD Edwards.
• •Bridge the JD Edwards AIS session token to Adobe Commerce OAuth 1.0a or IMS OAuth 2.0 credentials.
• •Poll Commerce on a schedule with deduplication on the Commerce increment_id and a full audit trail on every record.

Questions

Which direction does data move between Oracle JD Edwards and Adobe Commerce?

The main flow is Adobe Commerce into Oracle JD Edwards. Completed orders, invoices, and credit memos move from Commerce into JD Edwards as accounting entries, and customer and item records are aligned across both systems. JD Edwards data service tables are read-only, so ml-connector creates entries through orchestrations rather than writing tables directly.

How does the integration handle JD Edwards being on-premises with no fixed URL?

ml-connector accepts the full AIS Server URL, environment, and role per customer, since JD Edwards has no shared public base address. It posts the JDE username and password to obtain a session token, sends that token on the jde-AIS-Auth header, and re-authenticates on HTTP 444 when an AIS Server restart invalidates the token. The customer's CNC admin must also add the connector egress IPs to the AIS allowedHosts list to avoid HTTP 405 rejections.

Does this integration use webhooks or polling?

It uses both, depending on the storefront. JD Edwards has no native outbound webhooks, so ml-connector polls Commerce orders and invoices on a schedule using searchCriteria paging. Where Adobe Commerce has signed HMAC SHA256 webhooks enabled, ml-connector can also consume order and invoice events as they fire, then post into JD Edwards through the configured orchestrations.