JD Edwards Integration

JDE's architectural reality

JD Edwards EnterpriseOne has a distinctive architecture: a normalized table structure dating back to the 1990s, a business logic layer (Business Functions) on top, and progressively newer integration mechanisms layered on as Oracle has invested in modernizing the product. The right integration approach depends on which JDE Tools release the customer is on.

Three integration paths, by JDE release

JDE 9.2 with Tools 9.2.5 or later — Orchestrator

JDE Orchestrator is Oracle's modern integration framework, providing REST-based orchestration of JDE business functions and queries. For customers on current JDE Tools, Orchestrator is the cleanest integration path:

• REST endpoints for querying master data, inventory, and transactions
• Orchestrations to create planned orders, update planning parameters, and write back forecast
• OAuth-style authentication through the Orchestrator Studio

The Horizon implementation team works with the customer's CNC (Configurable Network Computing) administrator to define the required orchestrations during integration setup.

JDE 9.1 or 9.0 with AIS — Application Interface Services

AIS preceded Orchestrator and remains widely deployed. It provides REST/JSON access to JDE forms and queries, suitable for the master data and transaction reads Horizon needs.

Older JDE releases — Direct database access

For customers on older Tools releases without modern integration interfaces, Horizon connects through read-only database access. JDE's table structure is well-documented; the data Horizon needs lives in known locations:

• F4101 (Item Master) for SKU master data
• F4102 (Item Branch) for item-plant relationships
• F41021 (Item Location) for inventory positions
• F4801 (Work Order Master) for production orders
• F4311 (Purchase Order Header) and F4311Z1 for procurement data
• F4211 (Sales Order Header) and F42119 for sales order history
• F3002 (Bill of Material) and F3003 for BOMs
• F3003 and F30L1 for routings

Read-only views or replicated tables are created by the customer's DBA team. This is the path Oracle itself recommends for analytics integration to older JDE releases.

Scheduled file exchange

For customers preferring file-based exchange — common in JDE environments with restricted external network access — Horizon supports scheduled SFTP delivery of flat-file extracts produced by the JDE batch processing framework. Works alongside the API and database paths for environments that need it.

JDE-specific data model handling

Two aspects of JDE require integration-specific handling that generic ERP connectors miss:

• Branch/plant structure — JDE's branch/plant concept is broader than just "warehouse." A branch/plant can represent a profit center, a planning entity, or a physical location. Horizon's mapping treats this explicitly rather than assuming branch = physical location, which produces wrong results for many JDE deployments.
• Category codes — JDE's 30+ category codes per item carry critical planning information (planning family, ABC class, vendor relationships) that's customer-defined. Horizon's integration captures the specific category codes each customer uses for planning, rather than ignoring them or making assumptions.

What Horizon writes back to JDE

• Planned production orders (creating F4801 records or staging records for JDE's MRP to consume)
• Planned purchase orders (F4311 staging)
• Forecast records (F3460 or equivalent) for JDE's planning processes to consume
• Updated planning attributes on the item branch record (safety stock, leadtime, planning code) for JDE's own MRP runs

Mixed-mode and project-based JDE deployments

JDE is heavily used in mixed-mode manufacturers — companies doing make-to-stock, make-to-order, and engineer-to-order from the same operation. Horizon's planning model supports this mix natively. The integration captures order types and planning modes, so each customer order is planned with the appropriate logic rather than forcing everything through a single mode.

Typical project pattern

For a customer running JDE 9.2 with Tools 9.2.5+:

• Week 1-3: CNC team enables Orchestrator and defines the required orchestrations. Master data load and category code mapping.
• Week 3-5: Historical demand load through scheduled Orchestrator extracts.
• Week 5-8: Forecast generation, parallel run alongside existing JDE planning.
• Week 8-10: Cutover. First operational cycle with Horizon publishing back to JDE.

For older JDE releases using direct database access, add 2-3 weeks at the front of the project for the DBA team's view creation work.

Common JDE-specific integration patterns

• Tools release fragmentation — Different customer environments are on very different Tools releases. We confirm the customer's specific Tools release in the first integration conversation and select the integration approach accordingly.
• Customization weight — JDE customers tend to have heavy historical customization, including custom tables and modified business functions. The integration accounts for these explicitly rather than assuming vanilla JDE.
• Branch/plant reconciliation — Multi-entity JDE customers often have branch/plants that mix physical and logical entities. The Horizon load reconciles these explicitly during setup.