DuckLake is a new data lakehouse solution that separates metadata storage from data storage, storing metadata in SQL databases (Postgres, MySQL, DuckDB, SQLite) while keeping data in object storage. This architecture enables concurrent writes, eliminates egress fees when using services like Tigris, and allows querying from anywhere. The solution combines relational and non-relational data seamlessly, supports time-travel queries through snapshots, and can scale from laptop development to production workloads without complex infrastructure setup.
Table of contents
DuckLake is a data lakehouse for big and small data How it works The cloud’s the limit The new era of databases Sort: