Architecture
Decisions and trade-offs.
Direct comparisons between approaches and frameworks for the structural choices that recur across warehouse projects: ETL versus ELT, dimensional versus vault, lakehouse versus warehouse, and the platform selection that follows from them.
5 entries
Comparison
Data warehouse vs data lake vs data mart vs lakehouse
Data warehouse vs data lake vs data mart vs lakehouse: four distinct architectural commitments, what each one actually is, how they compare on storage, governance, query engine, and workload, and when each is the right choice in a 2026 stack.
Comparison
ETL vs ELT
ETL vs ELT: what the order of operations actually changes, why cloud columnar warehouses shifted the default from ETL to ELT, the trade-offs that determine which pattern fits a given workload, and a note on where reverse ETL fits.
Comparison
OLTP vs OLAP
OLTP vs OLAP: what the two database categories are actually optimized for, where the workload boundary used to be sharp, how columnar warehouses and HTAP systems have blurred it, and the trade-offs that determine which side a given workload belongs on.
Decision
Referential integrity in a data warehouse
Referential integrity in a data warehouse is a decision, not a default. A framework for choosing between database-enforced foreign keys, informational constraints, ELT-layer assertions, and unenforced declarations on Snowflake, BigQuery, Redshift, Databricks, and lakehouse table formats.
Comparison
Star schema vs snowflake schema
Star schema vs snowflake schema: when to denormalize the whole dimensional model, when to keep hierarchies normalized, and what changes on modern columnar warehouses where the textbook trade-offs no longer hold.