Don't Drop ACID - Transactions in Distributed NoSQL
summary
Explains why early NoSQL dropped full ACID, then demonstrates a working distributed multi-document ACID transaction on a document database and the cost tradeoffs involved; compares that to simpler patterns that rely on single-document atomicity and specific data modeling and querying techniques so consistency needs can be met without full transactions; enumerates practical limits (latency and cross-shard coordination), shows the demo completing successfully, and clarifies when each approach is actually safe to choose.
AI-written summary · about
related talks
58mDon’t Drop ACID - Transactions in Distributed NoSQL
42mMyths of NoSQL
01h 00mDistributed transactions implementation trade-offs
47mBeing transactional in 2019 - what does this actually mean?
29mModern Database Architecture: From SQL to NoSQL & Distributed SQL
52mTransactions: Myths, Surprises and Opportunities
13mThe ACID BASE Conundrum
49mThe cost of MongoDB ACID transactions in theory and practice