We put these tests together after clients with existing Rails stacks wanted our input on choosing Phoenix or Rails for their specific real-time push features. The benchmark apps are simple – no database, no sessions, but we wanted to set the PubSub patterns up in a way that would give us insights into a wide range of applications. The blog repo contains everything to setup an instance and run the benchmarks yourself:

https://github.com/chrismccord/channelsac

This won't be a popular comment, but I'm posting it anyway.

I'm annoyed by all these comparison posts that only harp on speed and benchmarks. The biggest benefit to using Rails is its ecosystem and maturity. When speed and benchmarks matter, you won't use Rails unless you're ready to throw money at hardware.

In the second sentence, "Phoenix is Not Rails" yet it goes on to compare ActionCable to Phoenix Channels.

All these posts on Phoenix vs Rails really confuse me. How does Phoenix, apart from MVC (and perhaps the folder structure?) compare to Rails?

Speed, benchmarks, and scaling aside, how does it really compare? Rails scales just fine if you throw money at it (a great thing!); Rails also has the Ruby ecosystem — I can guarantee is vastly superior than Elixir's ecosystem.

I digress.