Question

I'm looking to prototype a web app that will use sockets to push a gentle stream of messages to mobile web app clients. I want to pick an architecture that will work for a large number of clients if/when it moves to production (so i dont have to change later)

I'd like to start with rails because its familiar and has a strong structure from the go meaning easier to prototype. I think Faye will provide what i need in terms of a pub-sub layer but am I going to create a bottleneck by using ruby and the high number of socket connections, or will Faye isolate/protect Ruby server from that load, if you follow?

At the outset the load will not be significant so it won't matter, i just don't want to be hobbled later on when there are a lot of socket connections and i wish i used node.js ! Server side JS would be fairly new to me but I guess there are benefits in that the JS app can include the client side also

Advice appreciated.

Was it helpful?

Solution

You can take a look at https://github.com/faye/faye-redis-node.

This plugin provides a Redis-based backend for the Faye messaging server. It allows a single Faye service to be distributed across many front-end web servers by storing state and routing messages through a Redis database server

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top