Data management in a Microservices world
summary
Argues that data is the hard part of microservices and gives concrete trade-offs: cites 350 databases chasing niches, calls Monzo’s 1500 microservices an anti-pattern, and shows Netflix/Twitter graphs to prove service sprawl. Compares single-model vs multimodal storage, names Redis, MongoDB, Neo4j, Hadoop and Kafka, and says Hadoop becomes “garbage pile” while Kafka does not replace databases. Recommends thinking about processing, consistency and latency, and proposes Oracle REST Data Services or a data-as-a-service approach instead of one DB-per-service.
AI-written summary · about
chapters (15)
- 0:00Intro
- 1:30Microservices
- 6:27Microservices Architecture
- 10:58Polyglot persistence
- 16:10Databases
- 19:28Persistence
- 26:01Relational model
- 30:16Multimodal databases
- 31:48Hidden panes
- 34:33Event sourcing
- 35:53CQRS
- 36:35Kafka
- 37:46Cloud
- 40:49Data as a service
- 43:34Questions
related talks
30mDatabases in the Microservices World
58mDatabases in the Microservices World
49mChallenges in Implementing Microservices
48mIntegrating microservices by accepting the data dichotomy
46mData in a Microservices world from conundrum to options
47mHow to Choose a Right Database for Your Microservices?
43mPanel: Microservices, Where Did It All Go Wrong
48mChallenges in Implementing MicroServices