Skip to main content

Target Architecture

1. Enterprise layers​

Design the target across:

Sources → Integration → Persistence/Preparation → Semantic Layer → Analytical Consumption → Governance/Operations

Do not treat these as mandatory physical layers. They are logical responsibilities that may be implemented by different Datasphere capabilities.

2. Space strategy​

Use spaces as explicit security and ownership boundaries. Datasphere documentation describes spaces as secure areas where acquisition, preparation, and modeling occur; data is not available outside a space unless it is shared or exposed.

Design:

  • domain spaces
  • shared/consumption boundaries where justified
  • administration boundaries
  • development/test/production separation

Avoid creating spaces merely because a diagram looks cleaner.

3. Integration strategy​

For each source decide:

federate | replicate | transform | persist | publish

Record the reason and operational consequences.

4. Semantic architecture​

Separate:

  • reusable business entities
  • domain semantics
  • analytical models
  • consumer-specific presentation

The goal is semantic reuse without creating a single monolithic enterprise model.

5. Target-state diagram​

Your final diagram should show:

  • systems
  • domains
  • spaces
  • data movement
  • semantic boundaries
  • security boundaries
  • consumers
  • operational control plane