MotherDuck: Taking flight with interactive analytics
summary
MotherDuck runs a per-user DuckDB instance in the cloud and a WebAssembly DuckDB in the browser, coordinating via a MotherDuck extension that inspects the bind/optimize phases and splits work across locations. Demo uses Mosaic visualizations on a 5 million‑row star catalog, pulling shards to the browser for sub‑10ms interactions and larger work to cloud VMs (up to 96 cores). Features include layered cloud storage with zero‑copy clone, dual execution across local and cloud engines, and an embedded pg_duckdb path to ingest Postgres data.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:05Big data to small data shift
- 3:34Why DuckDB thrives
- 6:24MotherDuck cloud features
- 8:10Dual local-cloud execution
- 11:18Demo: Mosaic visualizations
- 14:49Demo: PostgreSQL hybrid execution
- 20:42Summary
- 22:05Q&A
related talks
36mMotherDuck
51mMotherduck: Data Analytics in the Post Big-Data Era
33mData preparation and analytics on your laptop with DuckDB
38mFrom Big Data to DuckDB: Ryan Boyd (MotherDuck) on Hyper-Tenancy, Duck Lake, and Making Data Fun.
24mpg_duckdb: Ducking awesome analytics in Postgres
28mAnalyze data easily with duckdb - and the implications on data architectures
27mIn Process Analytical Data Management with DuckDB
40mDuckDB: Harnessing in-process analytics for data science and beyond