How Not to Die Trying to Evolve Your Architecture
summary
Argues that an evolutionary architecture is practical when teams optimize three things: fast small deployments (automatic pipelines, continuous delivery, blue/green and canary releases) to reduce lead time; clear module boundaries by defining an architectural quantum (independent deployable with high cohesion) to avoid distributed monoliths; and measurable fitness functions run as pipeline triggers or continuous monitors—examples given: Snyk and Dependabot for security, ARUnit/GoodFences for layering rules, and tracking deployment frequency and lead time as objective metrics. Governance uses lightweight design-doc facilitation.
AI-written summary · about
chapters (24)
- 0:00Intro
- 2:42Evolutionary Architecture
- 4:06Incremental Change
- 5:00Develop
- 7:16Operation
- 9:28Architectural Coupling
- 11:47Architectural Quantum
- 15:47Fitness Functions
- 16:35Types of Fitness Functions
- 19:40Fitness Functions Examples
- 23:21How to Use ARUnit
- 24:30Evolution Architecture
- 25:20What are eles
- 26:19tradeoff
- 27:55best theoretical fit
- 28:39the enemy of good
- 30:16how can we work
- 32:50allowing change
- 33:58experiments
- 34:49centralized decision making
- 36:05distributed decision making
- 37:00lightweight governance
- 41:01feedback
- 41:33recommended readings
related talks
01h 00mEvolutionary Architectures
42mBuilding Evolutionary Architectures: Principles and Practices
49mThe Principles of Evolutionary Architecture
35mPrinciples of Evolutionary Architecture
39mPrinciples of Evolutionary Architecture
39mHow will Evolutionary Architecture evolve?
01h 03mEvolutionary Software Architectures
33mBuilding Evolutionary Architectures