OLAP / Warehouses
Column-oriented storage for heavy analytics. When you need to `SUM()` a billion rows.
Row vs. Column Storage
OLTP databases (Postgres, MySQL) store data row-by-row on disk. Great for fetching a single user record. OLAP databases (ClickHouse, Snowflake) store data column-by-column. If you run `SELECT AVG(age) FROM users`, a columnar database only reads the 'age' file from disk, skipping all other data. This makes analytics orders of magnitude faster.
ClickHouse
Ludicrous Speed
Open-source column-oriented DBMS. Capable of processing hundreds of millions of rows per second per server. Used by Cloudflare and Uber for real-time analytics.
Snowflake / BigQuery
Managed Cloud Warehouses
Separates compute from storage. You dump terabytes of data into cheap object storage (S3) and spin up massive compute clusters on-demand to query it. Expensive, but zero operational overhead.