The State of NoSQL in 2021
summary
NoSQL complements relational systems rather than replaces them, with denormalization and partial covering indexes used to avoid expensive cross-node joins in distributed clusters. Newer NoSQL engines add ACID support via different approaches: key-value-level two-phase commit transactions in SDKs, N1QL-style SQL transactions, MongoDB’s multi-version concurrency, and Redis’s non-rollback transactions. Expect cost-based optimizers, interest in atomic-clock-backed serializability (Google Spanner), Kubernetes operators for automated ops, SQL++ as a cross-DB query option, and a market of 200+ NoSQL engines aiming toward ~$22B.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:34Why NoSQL hasnt happened yet
- 4:15Tradeoffs
- 10:45Premature optimization
- 13:38Distributed relational
- 16:15Consistency
- 17:30Integrations
- 24:37MongoDB
- 27:12Atomic Clocks
- 28:58Distribution of NoSQL
- 30:25NoSQL popularity
- 30:54Which NoSQL database should I use
- 32:17The elephant in the room
- 33:19BSL
- 34:50Managing NoSQL
- 37:35SQL Plus
- 39:42Single Plus
- 41:37Databases converging
related talks
45mSay YES! to NoSQL: A Guide on When to Ditch Relational Databases
01h 00mFrom Relational Databases to NoSQL
01h 11mThe official ten-year retrospective of NewSQL databases
56mNewSQL: SQL Is Here to Stay
38mNoSQL Databases: Its not a free lunch
42mSay YES! to NoSQL: A Guide on When to Ditch Relational Databases
45mDemystifying NoSQL: The SQL developer's guide
01h 01mNoSQL: An Analysis