Skip to main content

17.1 BW/4HANA to Datasphere Mental Model

1. Why this phase matters​

A BW/4HANA developer already knows many concepts that remain valuable:

  • InfoObjects,
  • ADSOs,
  • transformations,
  • DTPs,
  • process chains,
  • CompositeProviders,
  • queries,
  • master data,
  • semantic layers.

The mistake is assuming every BW object has a one-to-one Datasphere replacement.

Instead, identify the architectural capability first.

2. Capability mapping​

BW capabilityModern architecture question
AcquisitionHow should source data enter the platform?
PersistenceWhere should durable data live?
TransformationWhere should business logic execute?
Semantic modelingWhere should reusable meaning be defined?
SchedulingHow are dependencies orchestrated?
MonitoringHow are failures detected and recovered?
ConsumptionWhich semantic interface serves consumers?

3. Three-platform mental model​

BW/4HANA
   |
   | existing enterprise warehouse
   v
Datasphere
   |
   | cloud data integration + semantic layer
   v
BW bridge
   |
   | managed BW capabilities for selected scenarios
   v
Modern SAP data architecture

This is not a required topology. It is a way to reason about responsibilities.

4. When BW/4HANA remains relevant​

Keep BW/4HANA when:

  • the organization has significant existing investment,
  • critical processes depend on BW-specific capabilities,
  • migration risk is high,
  • contractual or regulatory requirements require the existing platform,
  • a staged modernization is preferable.

5. When BW bridge becomes relevant​

Consider BW bridge when:

  • BW-style extraction and modeling are needed in the cloud,
  • existing BW knowledge and content provide migration leverage,
  • SAP application integration requires supported BW capabilities,
  • the organization wants to modernize without immediately rewriting every BW process.

6. When native Datasphere is preferable​

Prefer native Datasphere patterns when:

  • the use case fits Datasphere modeling and integration capabilities,
  • new cloud-native data products are being built,
  • BW-specific runtime semantics are unnecessary,
  • simplification is more valuable than preserving historical implementation.

Architect exercise​

Take a 500-object BW landscape and classify each workload:

  • retain in BW/4HANA,
  • move to BW bridge,
  • redesign natively in Datasphere,
  • retire.

Define objective criteria rather than platform preference.