16.10 Enterprise SAP Integration Architecture

1. Enterprise reference architecture

                         SAP BUSINESS SYSTEMS
      +----------------+----------------+----------------+
      |                |                |                |
   S/4HANA          BW/4HANA          HANA          Other SAP
      |                |                |                |
      +----------------+----------------+----------------+
                       |
             Supported source contracts
                       |
       +---------------+----------------+
       |               |                |
      CDS             ODP          Replication
       |               |                |
       +---------------+----------------+
                       |
                 SAP Datasphere
                       |
       +---------------+----------------+
       |               |                |
   Acquisition     Harmonization    Semantic
       |               |                |
       +---------------+----------------+
                       |
              Governed data products
                       |
        +--------------+--------------+
        |              |              |
       SAC          BW/BDC        Data Science

2. Domain-aligned integration

Do not create one giant SAP ingestion space for every business domain.

Consider domain boundaries such as:

• Finance,
• Procurement,
• Sales,
• Supply Chain.

Then define:

• source ownership,
• integration ownership,
• semantic ownership,
• product ownership.

3. Canonical data

A canonical model is useful when multiple SAP systems represent the same concept differently.

Example:

S/4 Europe ----S/4 US ----------> Enterprise Customer -> Data Product
S/4 APAC -------/

Do not canonicalize everything.

Canonicalization adds transformation and governance cost.

4. Multi-system S/4HANA

When multiple S/4 systems exist, decide:

• system of record,
• global versus local attributes,
• key harmonization,
• source precedence,
• conflict resolution,
• history strategy.

5. Integration pattern matrix

Requirement	Preferred starting pattern
Immediate source access	Federation
Repeated analytics	Replication
S/4 business semantics	CDS
Delta extraction	Supported CDS/ODP/replication pattern
Legacy extractor	ODP-supported pattern
Table-level replication	Supported SLT/replication pattern
Cross-system harmonization	Persist + model

6. Enterprise operating model

Assign ownership across:

• SAP application team,
• integration team,
• Datasphere team,
• security,
• data governance,
• analytics consumers.

Avoid the anti-pattern:

"Datasphere team owns everything."

They cannot own source semantics or business definitions alone.

7. Architecture decision records

For each major integration, document:

• source,
• contract,
• latency,
• volume,
• delta,
• persistence,
• security,
• support owner,
• rejected alternatives,
• migration path.

8. Architect capstone

Design a global S/4HANA-to-Datasphere architecture for:

• 4 S/4HANA systems,
• 2 SAP BW/4HANA systems,
• 15 business domains,
• 5,000 consumers,
• near-real-time operational analytics,
• daily enterprise reporting,
• historical analytics.

Your architecture must distinguish:

• federation,
• replication,
• CDS,
• ODP,
• semantic modeling,
• data products,
• governance,
• operations.