In September I started an OSS project called Exosuit. From the beginning I envisioned that there would be an economic component to Exosuit, that some functionality would be free and some functionality would be paid.
Here’s what I currently have in mind.
The free tier will offer the core promise of Exosuit: the ability to deploy to AWS as easily as deploying to Heroku.
The premise I have in mind is that you, the developer, have an idea for an app in mind, and you want the path that leads from “I have an idea in my head” to “I have an application running in production” to be as quick and frictionless a path as it can possibly be.
This ease and simplicity will have a trade-off, though: the infrastructure you get when you run
git push exosuit master (or whatever) will, out of necessity, be less robust than a full-blown production application with load balancers, scalable instances, etc.
Instead of a complex, sophisticated AWS infrastructure you’ll get a single EC2 instance, a single RDS instance, and the necessary security groups to wrap each. To me this still seems plenty good enough for a brand new app that’s just at the idea stage.
The paid tier will provide tools to help you “graduate” from the simple free tier infrastructure for an infrastructure more appropriate for a successful business application with real users.
Some ideas include:
- Easy way to put instances behind a load balancer
- Easy ability to scale instances (cattle vs. pets)
- Convenient way to build and manage production and staging environments
- Tools for common operations, e.g. copying production data to staging, recovering from RDS backup
- Domain management, including SSL
All this is totally tentative and might change. For the foreseeable future I’ll be focusing on the free tier because I see a compelling free tier as the best possible marketing for a paid tier.