Skip to main content

16.1 SAP Integration Mental Model

1. The SAP data landscape​

An enterprise SAP landscape commonly contains:

  • SAP S/4HANA Cloud,
  • SAP S/4HANA on-premise,
  • SAP ECC,
  • SAP BW or BW/4HANA,
  • SAP HANA,
  • SAP Datasphere,
  • SAP Business Data Cloud,
  • SAP applications exposing APIs and business data.

Datasphere should not be treated as a generic database connector.

SAP source systems expose different integration contracts.

2. Four questions​

For every SAP integration, determine:

  1. Business object — What business concept is required?
  2. Source contract — CDS, ODP, table, API, or another supported interface?
  3. Delivery mode — federation, initial load, or initial plus delta?
  4. Target purpose — operational access, persistence, semantic modeling, or analytics?

3. Integration hierarchy​

Prefer a business-semantic interface over direct physical-table access when an appropriate supported interface exists.

A typical hierarchy is:

Business-semantic CDS / supported API
              |
              v
       ODP / extraction contract
              |
              v
       Supported table replication
              |
              v
       Direct physical access

This is a decision framework, not a universal SAP rule. The actual supported mechanism depends on source release and object.

4. Federation versus replication​

Federation​

Query source data remotely.

Prefer when:

  • freshness must be immediate,
  • source workload is acceptable,
  • the source contract supports the access pattern,
  • persistence is unnecessary.

Replication​

Persist data in Datasphere.

Prefer when:

  • analytics repeatedly scans the data,
  • source workload must be isolated,
  • transformations require persistence,
  • history is required,
  • consumers need predictable performance.

5. Extraction versus semantic modeling​

Do not confuse:

Extraction contract

with:

Analytics semantic model.

A CDS view may provide a business-oriented source interface. Datasphere may still need additional harmonization, semantic modeling, dimensions, measures, and consumption models.

6. SAP-specific architect principle​

The strongest design usually preserves semantic intent as data crosses the landscape:

S/4HANA business semantics
          |
          v
Extraction contract
          |
          v
Datasphere harmonization
          |
          v
Enterprise semantic model
          |
          v
Consumer

Architect exercise​

For Procurement analytics, compare:

  • direct S/4HANA table access,
  • extraction-enabled CDS,
  • ODP DataSource,
  • replication flow,
  • remote federation.

Select a preferred pattern and defend it using semantics, latency, source impact, delta behavior, and operations.