Creating an Incremental Architecture for your System
summary
Argues that architecture grows by doing not grand planning, showing concrete moves: pick a minimal end-to-end skeleton on crutches (file system first, swap to DB later), prototype risky parts, and design for testability with TDD plus CI/CD so refactors stay safe. Gives a real case where AWS/Postgres replication delivers horizontal scale with minimal work, and a microservice split that failed until related data was merged back. Recommends hotspot analysis—analyze git history—to find where to start on legacy code.
AI-written summary · about
chapters (10)
- 0:00Introduction and talk overview
- 1:52What is software architecture
- 15:13Drivers for system design
- 29:15Problems with big upfront design
- 31:28Incremental design explained
- 47:31Politics and stakeholder engagement
- 50:21Start small and focus on knowns
- 1:01:15Avoiding design pitfalls
- 1:13:58Architectural tasks in workflow
- 1:24:18Q&A and closing remarks
related talks
50mTowards an Evolutionary Architecture
56mTowards an Evolutionary Architecture
48mBuilding Evolutionary Architectures: Principles & Practices
38mAgile Architecture
44mThe Art of Architecting for Scale
01h 00mEvolutionary Architectures
28mThe lean architect: Experimenting your way to better decisions
58mThe past, present and future of Architecture