Skip to main content

18.12 Enterprise Multi-Cloud Architecture

1. Reference architecture​

 SAP APPLICATIONS       CLOUD WAREHOUSES       OBJECT STORAGE
       |                       |                      |
       +-----------+-----------+----------------------+
                   |
             integration layer
                   |
        +----------+----------+
        |                     |
   federation              replication
        |                     |
        +----------+----------+
                   |
              DATASPHERE
                   |
       +-----------+-----------+
       |                       |
 enterprise semantics     data products
       |                       |
       +-----------+-----------+
                   |
             SAC / BI / APIs

2. Platform ownership​

Define one owner for each layer:

LayerExample responsibility
Sourceoperational correctness
Integrationmovement and reliability
Datasphereharmonization and semantics
Governanceclassification and policy
Consumptionbusiness experience

3. Multi-cloud principles​

Principle 1 — Do not copy by default​

Copy data only when the copy provides measurable value.

Principle 2 — Keep semantics close to ownership​

Business definitions should have an accountable domain owner.

Principle 3 — Separate raw from governed​

A raw landing area is not an enterprise semantic product.

Principle 4 — Design failure domains​

Every external source can fail independently.

Principle 5 — Measure freshness​

"Near real time" is not an architecture requirement until translated into a measurable SLA.

4. Architecture trade-offs​

Centralized Datasphere​

Pros:

  • common semantics,
  • governance,
  • reusable data products.

Cons:

  • central dependency,
  • possible bottleneck,
  • cross-domain ownership complexity.

Federated domains​

Pros:

  • domain autonomy,
  • local ownership.

Cons:

  • duplicated semantics,
  • governance complexity,
  • inconsistent definitions.

A hybrid model is often appropriate.

5. Architect capstone​

Design an enterprise platform containing:

  • S/4HANA,
  • BW/4HANA,
  • HANA Cloud,
  • Redshift,
  • Oracle,
  • SQL Server,
  • S3,
  • SaaS APIs,
  • Datasphere,
  • SAC.

Your architecture must define:

  • source ownership,
  • integration pattern,
  • persistence,
  • semantic ownership,
  • security,
  • monitoring,
  • recovery,
  • cost controls,
  • retirement strategy.