Skip to content

Snowflake Data Cloud Framework

Data Architecture Layers

  • Layers: mainline layers where data flows from lower order layer to the next
    • Sources: SoRs
    • Landing: Cloud storage
    • Raw: ingested into database for processing
      • analysts and data scientists can be allowed access for exploration
    • Integrated/Refined: conformed, after business transformations have been applied
    • Presentation/Curated: Semantic, secured, optimized for consumption
    • Shared: business entities make data available for consumption by other business entities or other corporation for monetization
  • other areas that complement the mainline layers
    • common: contain common assets such as UDFs that are common across all business entities
    • workspace: dedicated work-areas for analysts and data scientists to persist intermediate results
  • BP: split layers into different databases (v/s schema) if replication of a particular is required (e.g. cross-geo data access)
  • Service architecture: how responsibilities are divided
    • Platform as a Service: the central team provisions databases and access, business entities are responsible for managing end-to-end data layers
    • Data as a Service: Central team brings the data from all source systems into the Raw layer, business entities are responsible for integrated, through presentation (access or semantic) layer
    • Analytics as a Service: Central team is responsible for all layers including presentation
  • Data security:
    • Functional role: assigned to users, a collection of access roles
    • Access role: group privileges to databases