Betting on Evolutionary Architecture
summary
Argues that treating architecture as an experimentable asset lets teams place tiny bets and pivot: uses infrastructure as code, CI pipelines, canary and blue‑green deploys and phoenix environments to change topology in minutes; measures with Graphite/Hystrix, Dropwizard tracing and lightweight probes like Siege and Vegeta; experiments (bigger CPU, move compute to S3 region, XSLT tweaks, then a Varnish cache) cut page load from ~35s → 6s → 4s → 3.5s → ~0.2s.
AI-written summary · about
chapters (22)
- 0:00Start
- 1:06What Is Software Architecture
- 1:28The Things That Traditionally Have Been Hard To Change in Software Architecture
- 4:00Deployment Topologies
- 5:09Survivor's Bias
- 5:32The Dreyfus Model of Skill Acquisition
- 15:26Divergent Evolutionary Graph
- 17:42Betting on Software Architecture
- 18:27The Apollo Space Program
- 21:28Gemini Program
- 21:50Regalo Wing
- 25:29Black Scholes Equation
- 28:55Requirements for Performance
- 31:59Hypothesis Driven Architecture
- 32:10Richard Fineman
- 35:26Observability
- 36:01Tools
- 36:21Graphite Dashboards
- 36:55Monkeys of Continuous Delivery
- 40:25Cloud Native Infrastructure as Code
- 40:58Canary Release
- 41:20Phoenix Environments
related talks
42mBetting On Evolutionary Architecture
59mBetting on evolutionary architecture
48mBuilding Evolutionary Architectures: Principles & Practices
50mTowards an Evolutionary Architecture
53mEvolutionary Architecture: Principles and Patterns
40mPractices from trenches for embracing evolutionary architecture in wild
44mEvolutionary UI
01h 01mArchitecture: The Stuff That's Hard to Change