DEV Community

Cover image for Clean Core Integration for SAP S/4HANA Public Cloud: A Checklist for Engineering Teams
APPSeCONNECT
APPSeCONNECT

Posted on

Clean Core Integration for SAP S/4HANA Public Cloud: A Checklist for Engineering Teams

TL;DR: On SAP S/4HANA Public Cloud, you cannot modify the core. That moves every custom rule, sync job, and workaround into your extension and integration layer. Design that layer first, or every upgrade becomes a fire drill.

Why does clean core change the integration job?

Public Cloud is SaaS. SAP ships two major upgrades a year and controls the core. Anything you build has to survive those upgrades without manual fixes. That rules out old on-premise habits: custom tables, direct database reads, and modified standard code.

Where should custom logic live in SAP S/4HANA Public Cloud?

Three options, in order of preference:

Key user extensibility for fields, simple logic, and forms inside the tenant.
Developer extensibility with ABAP Cloud, using released objects only.
Side-by-side extensions on SAP BTP or external services, calling released APIs.

Rule of thumb: if it depends on something SAP has not released, expect it to break on an upgrade.

What breaks first in a midsize SAP stack?

The edges. A typical midsize business runs SAP alongside Shopify or Amazon, a CRM, a WMS, and a shipping tool. Orders, inventory, customers, and invoices must agree across all of them. Point-to-point scripts carried over from the old ERP are usually the first thing to fail.

The clean core integration checklist

Audit customizations. Tag each one “differentiating” or “habit.” Only differentiating logic earns an extension.
Map every system touching ERP data. Record direction, frequency, and system of record per entity.

Build on released APIs only. Check SAP Business Accelerator Hub before writing code.
Clean master data before migration. Duplicate customers and bad item records survive go-live.

Keep the ERP as the system of record. Sync outward from it, not around it.

Set sync timing per entity. Orders and inventory usually need near real time. Product and price data can often batch.
Make flows observable. Retries, error queues, and alerts matter more than throughput on day one.
Regression-test integrations every release. Put integration tests into each SAP upgrade cycle, not just go-live.

The bottom line

Clean core works like a contract. SAP keeps the core stable, and you keep your logic outside it. Teams that respect that contract get quiet upgrades. Teams that fight it rebuild their old workarounds in the cloud. A clean integration layer is also the groundwork for AI and automation, which only work on connected, trusted data.

What has tripped up your team on clean core? Share it in the comments.

Still deciding whether Public Cloud fits your business? Our buyer’s guide covers fit, cost, timeline, and how it compares with NetSuite, Business Central, and Acumatica: appseconnect.com

Top comments (0)