Domain: System Blueprints

System Blueprints

Platform architecture patterns — the big structural decisions chosen once and lived with for years, from warehouses and lakes to lakehouses, mesh, and streaming.

mindmap
  ((System Blueprints))
    (data warehouse)
    (data lake)
    (lakehouse)
    (data mesh)
    (streaming)
    (open table formats)
    (context and metadata)

01-data-warehouse-architecture

Star/snowflake schema, Kimball vs Inmon, SCD types, Data Vault 2.0, cloud warehouse comparison. Start here if your data is structured and analytical.

02-data-lake-architecture

Zone architecture (landing/cleansed/curated), file formats, governance, anti-patterns. Start here if you work with unstructured or semi-structured data at scale.

03-lakehouse-architecture

Combining lake storage with warehouse semantics: ACID on object storage, open table formats, query engines. Where most modern platforms are converging.

04-data-mesh-architecture

Domain-driven data ownership, data products, federated governance. When centralized architecture creates organizational bottlenecks.

05-streaming-architecture

Lambda vs Kappa, event-driven design, CDC, Kafka/Pub/Sub, windowing, exactly-once semantics. When batch latency isn’t acceptable.

06-open-table-formats

Delta Lake, Apache Iceberg, Apache Hudi — format internals, time travel, schema evolution, maintenance operations. The technology layer that enables lakehouse.

07-context-and-metadata-architecture

The five types of pipeline context (run, provenance, temporal, quality, business), context propagation patterns, bi-temporal modeling, schema evolution. The semantic layer that makes data self-describing.