Real-World NoSQL Schema Design #bbuzz
summary
Converts the MusicBrainz relational model into denormalized NoSQL documents and column families, cutting MusicBrainz -> 138 tables to 4 by embedding aliases, recordings and releases into single rows; demonstrates time-series design that compresses billions of tiny rows into per-minute columns; shows Apache Drill SQL over HBase/JSON to run joins and aggregations without ETL; explains atomic single-row/document updates, distribution by row key, and a percolator pattern for multi-store sync that keeps full‑text, recommendation, and HBase APIs consistent.
AI-written summary · about
chapters (16)
- 0:00Introduction and background
- 1:50Why NoSQL databases emerged
- 5:46Schema design in NoSQL vs relational
- 9:18Organizing data in NoSQL systems
- 14:12Time series data modeling
- 15:14Case study: MusicBrainz migration
- 21:03JSON document advantages
- 23:11E-commerce product catalog example
- 27:42MusicBrainz with JSON documents
- 29:08Choosing the right NoSQL database
- 31:17Working with JSON data types
- 32:40Schema flexibility benefits
- 33:43Analytics with SQL on NoSQL
- 37:16Querying distributed NoSQL systems
- 39:24Transactions and data integrity
- 40:26Summary and key takeaways
related talks
18mGermany's Next Top Data Model
01h 01mNoSQL: An Analysis
26mNoSQL Data Modeling for Front-end Developers
45mNoSQL: Death to Relational Databases
37mDenormalizing Rails
45mDemystifying NoSQL: The SQL developer's guide
53mFrom RDBMS to NoSQL (PRT314)
46mFrom Tables to Documents -- Changing Your Database Mindset