MotherDuck
summary
MotherDuck runs DuckDB as a hybrid local/cloud system that keeps most work on the laptop while spinning a cloud-side DuckDB “duckling” per session to avoid egress and speed S3 scans. Claims: 90% of queries hit ≈100MB and typical datasets are ~100GB so single-node memory (up to ~1TB) suffices. Built: a DuckDB extension, layered caching, serverless ducklings and a web UI; prototype demo works but cache hydration, failure handling and cross-cloud egress remain unresolved.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 3:28In-memory DBs in cloud
- 7:23Local-cloud hybrid execution
- 13:40Cloud service with DuckDB
- 18:10Architecture and cloud integration
- 25:36Hybrid query demo
- 32:17Future plans & Q&A
related talks
27mMotherDuck: Taking flight with interactive analytics
51mMotherduck: Data Analytics in the Post Big-Data Era
38mFrom Big Data to DuckDB: Ryan Boyd (MotherDuck) on Hyper-Tenancy, Duck Lake, and Making Data Fun.
33mData preparation and analytics on your laptop with DuckDB
40mDuckDB Meets Data Lakes
27mThe Duck and the DataFrame: A Data Engineer’s Journey with DuckDB
28mState of the Duck
28mAnalyze data easily with duckdb - and the implications on data architectures