Lesson 6 of 6
Integration with S/4HANA and what to check
10 min
Listen to this lesson
IBP plans only as well as its data. Master data (products, locations, customers, sources of supply), history and stock come from S/4HANA; the agreed plan goes back. SAP offers two main integration routes, and most projects use both.
| Route | What it moves | How it runs |
|---|---|---|
| SAP Cloud Integration for data services (CI-DS) | Time-series data: master data, history, stock and the released plan back to S/4HANA | Scheduled batch tasks; a Data Services Agent runs in Contoso's network next to S/4HANA |
| Real-time integration (RTI) | Order-based planning: master data and orders such as sales orders, purchase orders, requisitions | Changes are captured in S/4HANA, queued with bgRFC and pushed to IBP through the Cloud Connector |
| APIs and files | External data such as customer forecasts or market data | OData services, file uploads |
Integration routes for IBP; check SAP's integration guide for your release.
At Contoso, Theo's integration team builds the CI-DS tasks. The scope is time-series only in phase one: product, location, customer and location-product master data, monthly billing history, stock on hand, and, outbound, consensus demand as planned independent requirements for the S/4HANA MRP run. RTI comes with order-based planning in a later phase. SAP's material describes RTI as the primary route for order-based planning going forward, using bgRFC for transactionally consistent transfer.
Master data for planning
IBP master data is derived from S/4HANA objects, and every derivation is a mapping rule that can be wrong. The table shows Contoso's mapping for phase one.
| IBP master data type | S/4HANA source | Mapping rule to test |
|---|---|---|
| Product | Material master (basic data) | Only finished goods and key components in scope; base unit of measure carried over |
| Location | Plants 1710, 1720, 1810 | Location type: plant or distribution center |
| Customer | Business partner in customer role | Grouped into Great Lakes, other fleet and aftermarket |
| Location product | Material / plant (MRP views) | Only combinations with an MRP type other than ND |
| Transportation lane | Special procurement keys and sourcing | Toledo supplied from Detroit and Chicago |
| Resource | Work centers such as ASSY-10 | Capacity in pieces per month derived from shifts |
Illustrative course data: phase-one master data mapping.
The rule "only material / plant combinations with an MRP type other than ND" matters at Contoso: spare parts in 171S are excluded from MRP after the conversion, so they must not appear as planning objects in IBP either. A count test catches this; a screen test does not.
The outbound plan: what S/4HANA receives
Contoso sends consensus demand to S/4HANA as planned independent requirements (PIRs) for CMC-ALT-100 and CMC-ALT-240 in plants 1710 and 1720, requirements type per the planning strategy agreed in sap-pp-202, in an agreed PIR version. MRP then plans components such as CMC-RTR-200 from those PIRs. Check PIRs in MD63 (display) and the effect in MD04 or the Monitor Material Coverage app.
Two integration moments carry the most risk. The initial load before the first parallel run brings 36 months of billing history and all master data; it is large, runs once per test tenant, and must be reconciled in full. The monthly cycle then runs delta loads of master data, one month of history, and one release. Test both: a green monthly cycle on top of a wrong initial load still produces a wrong forecast.
What to verify in a planning run
A planning run has no single expected value the way a posting does, so verification combines reconciliation, controlled data and plausibility. Contoso's checklist for each cycle:
| Check | How | Evidence |
|---|---|---|
| Master data complete | Count location-products in IBP against plant-material combinations in scope in S/4HANA | Two counts and a difference list |
| History reconciled | Monthly totals per plant in IBP equal the agreed S/4HANA billing report | Totals table signed by Sunita |
| Forecast job correct | Controlled data set: forecast and MAPE equal a spreadsheet calculation | Spreadsheet and job log |
| Supply plausible | Capacity load for ASSY-10 matches hand calculation within tolerance | Planning view snapshot |
| Release complete | CI-DS task record count equals PIR items; totals per material equal consensus | Task log, MD63 output |
| Release correct | Sample of PIRs: material, plant, date, quantity, version | Sample sheet with five to ten items |
Contoso's planning run checks (illustrative).
- Do record counts match between the CI-DS log and S/4HANA?Yes → compare totals · No → read the task log errors first
- Do totals per material and plant match consensus?Yes → sample items · No → check unit of measure and bucket mapping
- Do sampled PIRs have the right dates and version?Yes → run MRP and check components · No → check the period-to-date mapping
What would you do?
The CI-DS task log shows 2,400 records sent, but MD63 shows 2,310 PIR items for the same version. What should Theo do first?
- CI-DS moves time-series data in batches through an on-premise agent; RTI pushes order-based data via bgRFC.
- At Contoso, consensus demand returns to S/4HANA as planned independent requirements.
- Verify a planning run with counts, totals, controlled data and samples.
- Watch bucket-to-date, unit and location-to-plant mapping in the release.
Check your understanding · not graded
Lesson check
1. Which route moves time-series data between S/4HANA and IBP?
2. What does RTI use to transfer order-based data consistently?
3. Where do you display the planned independent requirements created by the release?
4. Every released month lands on the first working day. What kind of defect is this?