The Entity Microservice Trap: You May Be Doing It Wrong
summary
Argues that splitting a monolith into entity microservices recreates tight coupling, slow cross-service joins and CAP problems and that many teams hit a wall when they treat services as CRUD wrappers. Recommends behavior-oriented services backed by tiny per-service stores (key‑value, reporting warehouse) and an event bus for asynchronous communication; example offer engine uses service bidders, a Redis cache read within 300 ms and nightly ETL cron jobs for eventual consistency. Claims real-world payoff: 50 million pounds with 50 people.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:23Fred George
- 3:23Fred Brooks Design Approach
- 4:0390's Design Approach
- 6:06Corollary
- 7:36Rapid Deployment in Perth
- 11:39Conclusion: Break DB Apart
- 12:00Asynchronous Support
- 13:07Connascence by Jim Weirich
- 20:24The Cynefin Framework
- 23:01Alternative: Event Bus
- 24:23Distributed, Heterogenous DB
related talks
48mIT Architecture : All your micro-services are wrong PERIOD!
57mEmbracing Microservices
23mWhy You Had a Bad Time With Microservices
57mSOA lessons learnt (OR Microservices done better)
38mMicroservices, a bittersweet symphony
36mDo Proper Monoliths before You Do Microservices!
54mThe Long Sad History of Microservices
47mGetting (service) design right