An Ode to 17 Databases in 39 minutes
summary
Seventeen databases get rapid, concrete contrasts: PostgreSQL JSON/HStore as a practical move toward document queries, Redis as a fast in‑memory tool that is not safely distributed, Dynamo/Riak showing vector clocks and CRDTs and the tradeoff of eventual consistency, Cassandra described as production‑scale (Netflix uses it at thousands of nodes), Google’s Bigtable/Spanner claiming atomic distributed transactions with GPS/atomic clocks, MongoDb critiqued for correctness tradeoffs, and RethinkDB bleeding edge versus Elasticsearch as a robust clustered search datastore. The talk names what breaks, what scales, and what people actually run.
AI-written summary · about
chapters (4)
related talks
38mAn Ode to 17 Databases in 33 Minutes
44mAn ode to 17 databases in 29 minutes
45mDatabases for Data Science
34mData Storage: NoSQL Toasters and a Cloud of Kitchen Sinks
41mHow to choose the right database for your application
42mDatabases unveiled; choose wisely
45mNoSQL: Death to Relational Databases
01h 02mStoring Non-Scalar Data