📖 Read the The System Design Checklist.
This is a checklist of things to think about when designing or re-designing a software system or parts of it.
For example, a specialised, shorter version of this checklist can be used to design:
- an Minimum Viable Product (MVP) for a pre-PMF company
- a microservice for a post-PMF company
- a migration process from a monolith to a microservice architecture
- a logging and monitoring system for external auditing
- a frontend component
- a new API endpoint etc.
This is not a tutorial on software system design. However, as a refresher, I added a few notes about why certain questions are important and some opinionated suggestions. Feel free to skip over them.
I've successfully used checklists my entire life, as a software engineer, during my time as a CTO of an early-stage fintech app, where I designed software for high-street banks and when I prepared for the AWS Solution Architect exam. I created a mental checklist to:
- extract the system requirements from the question
- determine the desired system properties
- break down the system layers
- consider trade-offs at each layer
This is how I was able answer 60+ scenario-based questions in 120 minutes and pass the exam 🙌.
We are bombarded by new information every day. There are many books and articles on software system design. Technologies, frameworks, languages, practices change at huge speed.
However, most of them lack a practical format, to apply the knowledge immediately. Also, our brains can't cope well with so many long-form texts on a daily basis.
I noticed a move towards checklists and flowcharts in the developer community. Therefore, this repository is the first step towards a practical tool. I want to help not only first-time CTOs, but any software team before a routine design task.
I hope the checklist accomplishes three goals:
- Helps them discover unknown unknowns by sparking curiosity and reflection ahead of the design task.
- Inspires them to use it in Pre-Mortems, to better assess high-impact, high-risk systems, and in Post-Mortems, to update it with learnings discovered the hard way.
- Used and refined frequently, it builds good habits leading to high quality work.
“Checklists […] remind us of the minimum necessary steps and make them explicit. They not only offer the possibility of verification but also instill a kind of discipline of higher performance.” — Atul Gawande, The Checklist Manifesto
If you like the checklist, let's make it more useful together! 💪
Mistakes are human nature. Because technology advances so quickly, what was good-enough yesterday may not be tomorrow. Nobody can know everything at all times.
You're very welcome to contribute to fix mistakes, and refine the list of questions and options by submitting a Pull Request.
I'm running some tests with real teams from a few companies to benchmark if the usage of a checklist results in:
- shorter
commit → review → deploycycles- improved code quality
- increased number of issued discovered ahead
- fewer severe incidents
- better knowledge loops in and between teams
If you want to try it, email me at [email protected]. Feel free to leave a comment with suggestions to make it more specific or shorter for certain cases, like the ones I mentioned in the first paragraph.
- Install Docker on your machine
- Run Jekyll:
docker-compose up - The browser should livereload at
http://localhost:4000