@flovilmart
- Parse, the mobile backend as a service (mBaaS)
- Free backend. Can scale with $.
- Schema-less, easy to implement changes in models
- Security with ACL
- Client SDKs the gateway to the service. Service itself opaque
- Timeline
- Aug 2011, started
- 2013, acquired by Facebook
- Mid 2015, open source SDKs
- Early 2016, Parse Server, Dashboard
- Jan 2017, end of service
- Lots of open source Parse libraries (iOS, Android, PHP, Arduino, …)
- Dashboard written in React.js, architecture is ✨
- To open source…
- From 1M users to 1 (one) (software you run)
- Product shift, different requirements. Chain reaction…
- Technology shift
- Audience shift
- Lots of complaints, e.g. too long to set up, needs to fit in smallest tier of IaaS provider
parse-server- Under the hood
- It’s an express.js app, with command line interface
- ES6, small memory footprint (~150MB RAM)
- Open source:
- in 5 months:
- 10k stars, 500 PRs, 1300 issues closed, 2500 forks
- Technology shift, platform agnostic?
- Front-end devs (mobile, web) no back-end experience, never touched a database
- in 5 months:
- Under the hood
- Audience shift:
- Existing users:
- Front-end focused, care about continuity of service, don’t want to rewrite
- New users:
- Full stack, care about API / SDK, flexibility, ease of use
- SaaS
- Business, want to make $$$ 💰
- Existing users:
- Moving forward:
- Designed to be extensible, from the ground up, through modules
- Live queries
- Native social auth, not just Twitter / Facebook, currently 8 supported
- PostgreSQL (CockroachDB)
Wed Jun 29 12:22:45 PDT 2016