Rails to Phoenix
summary
Bleacher Report rewrites a Rails site into Phoenix and runs at 1/15th of the servers, averaging 10–30 ms with only one spike near 800 ms; independent benchmarks show Phoenix serving 2.5× more records and 4.5× larger payloads while running 1.5–2.5× faster than tuned Rails. Explanation cites Elixir on the BEAM, Plug and Mix tooling, per-process GC, compiled routing, and OTP features; Phoenix scales to two million concurrent WebSocket connections and adds Phoenix Presence to reconcile distributed client state.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:43Is Performance Enough?
- 5:19Elixir consumes all CPU cores
- 5:31Elixir is a compiled language
- 6:06Phoenix makes smart use of the BEAM
- 6:35The database is not slow
- 9:11Can Phoenix's performance benefits free engineers to work on bigger ideas?
- 10:33Optimize for developer happiness
- 13:02Variables are immutable
- 18:29Elixir & Phoenix optimize for long- term developer happiness
- 20:52Performance is the least interesting thing about Phoenix
- 24:57Pre-existing domain knowledge speeds up adoption
- 27:45Implementing features costs less money and requires less engineering support
related talks
32mRails to Phoenix: How Elixir can level-you-up in Rails
41mPhoenix Framework
01h 00mPhoenix a Web Framework for the New Web
40mTaking Off with Phoenix
44mRise of The Phoenix - Building an Elixir Web Framework
45mPhoenix a web framework for the new web
32mWhy I choose Phoenix
41mElixir & Phoenix – fast, concurrent and explicit
