Databases unveiled; choose wisely
summary
Overview of concrete claims and tools: companies moved 300 shops into two and rebuilt passenger systems, then suffered a major outage on 3 April when missing active data stops dispatching; key-value stores like Redis and etcd power ephemeral shopping sessions and Kubernetes discovery; relational Postgres/Oracle hold master data but Postgres JSON queries initially slow; document stores (Mongo, Cosmos, DynamoDB) store train docs keyed by date+train number; graph stores (Neo4j, Gremlin) handle route planning; time-series (Prometheus, Cassandra) ingest metrics from 9,000 sensors per train. Recommends layered design, CQRS and event sourcing to switch stores later.
AI-written summary · about
chapters (15)
- 0:00Introduction to databases
- 3:34Importance of data in systems
- 6:50Seven fallacies of distributed systems
- 9:21CAP theorem and its evolution
- 12:11Use cases for key-value stores
- 14:40Relational databases in practice
- 17:50Document stores for train data
- 21:25Graph databases for travel planning
- 23:56Time series databases for monitoring
- 25:00Choosing database wisely
- 26:27CQRS for flexible data access
- 30:22Event sourcing and projections
- 32:10Avoiding database-locked models
- 34:19Consistency vs. data size trade-offs
- 36:05Conclusion and final advice
related talks
35mAn Ode to 17 Databases in 39 minutes
41mHow to choose the right database for your application
40mBeyond Relational: Data storage for modern applications
52mData storage for modern applications - Mike Lehan
29mPurpose-Built Databases for Modern Applications
33mBuilding an Elastic, Resilient Data Layer to Power New Retail Experiences
34mData Storage: NoSQL Toasters and a Cloud of Kitchen Sinks
29mPurpose-Built Databases for Modern Applications