Skip to main content

🧩 Data Warehouse (DWH)

✅ Overview​

Centralized data store integrating organization-wide data and storing it optimized for analysis use.

✅ Problems Addressed​

  • Data is siloed per business system.
  • Complex JOINs and pre-processing are required for every analysis.
  • Reporting and analysis become personalized and reproducibility is low.

Data Warehouse provides a centralized infrastructure centered on "Integration", "Normalization", and "History Management" for these.

✅ Basic Philosophy & Rules​

  • Format data with ETL (Extract → Transform → Load) and store in DWH.
  • Schema design is optimized for analysis like Star / Snowflake schema.
  • History is also managed by retaining time-series data (SCD etc.).

✅ Suitable Applications​

  • Integrated reports, BI dashboards.
  • Centralized management of company-wide KPIs.
  • Definitive data store for audit and regulatory compliance.

❌ Unsuitable Cases​

  • Exploratory analysis wanting to handle raw data flexibly (Data Lake is more suitable).
  • Cases where real-time nature is important like streaming.

✅ History (Genealogy / Parent Styles)​

  • Standard approach continuing since 1990s.
  • Multiple schools exist like Kimball / Inmon methods.

✅ Related Styles​

✅ Representative Frameworks​

  • Amazon Redshift
    Pioneer of cloud DWH. Provides large-scale analysis processing.

  • Google BigQuery
    Serverless DWH realizing scale and query acceleration.

  • Snowflake
    Characterized by virtual warehouse structure and compute separation architecture.

  • Teradata / Oracle Exadata
    Traditional on-premise DWH, infrastructure for high-performance analysis processing.

✅ Design Patterns Supporting This Style​

  • Template Method
    Unifies ETL procedures (Extract → Transform → Load).

  • Strategy
    Switching optimization strategies (Index / Partition).

  • Iterator
    Used when processing massive data sequentially.

  • Facade
    Integration layer (BI tools and metadata management) hides internal complexity.

✅ Summary​

DWH is the royal road for standard reports and definitive data analysis,
and is still widely used as a stable business analysis platform.