Challenges in Implementing MicroServices
summary
Argues microservices work best for fuzzy, fast experiments and insists on an event-driven stack: Kafka (250k msgs/sec) as a Rapids, small Rivers with ZeroMQ, services publishing conclusions (not raw data), and one persistent store per service (Redis for fast reads). Recommends Docker containers and cloud provisioning for seconds-scale deployment. Cites a rental-car system: ~300 microservices, ~$6M build producing ~$40M extra revenue, and the 100-lines / single-maintainer rule for true microservices.
AI-written summary · about
chapters (26)
- 0:00Intro
- 2:38The Cynefin Framework
- 6:17Hardware Lead Times
- 8:24Services Impact
- 9:47Summary Principles of MicroServices • Very, very small
- 12:06Common Theme of All Challenges
- 12:38Challenge 1
- 13:16Chad Fowler vs. Fred George
- 14:22Rapids, Rivers, and Ponds
- 16:00High-Performance Event Bus
- 20:13Asynchronous Services Example
- 24:51Service Taxonomy
- 26:40Challenge 2
- 27:48Challenge 3
- 28:03Fred Brooks' Design Approach
- 28:2290's Design Approach
- 28:55Design Approach Corollary
- 30:05Challenge 4
- 31:103 Companies - 3 Variants
- 32:36A New Hope...
- 33:02Challenge 5
- 33:44Open Source Frameworks
- 35:16Challenge 6
- 36:12Conferences / Meetups Starting...
- 36:49Challenge 7
- 38:00Challenge 8-n
related talks
50mChallenges in Implementing Microservices
49mChallenges in Implementing Microservices
47mIt is not just Microservices
16mMicroservice Challenges
38mMicroservices, a bittersweet symphony
54mThe Long Sad History of Microservices
20m3 reasons why micro-services aren't the miracle cure
43mGo micro with Microservices