Torquebox vs. Trinidad + extensions for my use case

I’m evaluating options for a jRuby R. deployment that needs a couple
of
load balanced app servers to do the following:

  • ~10 persistant running network services with access to the whole
    Rails
    environment (for iOS APNS, Twitter/Instagram streaming APIs, etc.)
  • backgrounding
  • scheduled jobs
  • websockets (in the future)

In an MRI deployment I’d end up with Redis for a shared session store, a
server for backgrounding and cron, and some evented server messaging
with
Rails for the websockets - yuck! Wish I could do with less.

I started looking for Trinidad extensions and experimenting with
Celluloid
for the services, and spinning up actors for websocket connections, then
I
discovered Torquebox, read through the documentation and was like “holy
crap this does everything!”

But then posts like this have me a bit worried
http://thenerdings.blogspot.com/2012/09/pulling-plug-on-torquebox-and-jruby-for.htmlas
I don’t have much background on JBoss (of course, i’m happy to go look
into what I need to know)

Are there any thoughts from the community on what’s good for an
“all-in-one” K.I.S.S. architecture?

Thanks!
Ambert

You should have a look at this book (which I wrote):
http://pragprog.com/book/jkdepj/deploying-with-jruby

it discusses a lot of the trade-offs between Trinidad and Torquebox.

I have a tendency to favor Torquebox, but @kares is doing some
incredible work right now with Trinidad and it’s extensions.

If you’re serious about websockets, Torquebox has really awesome support
for that. It even hooks into the messaging queues, so you can
essentially pipeline messages from the background jobs up to the
browser. I have a nice example of this in the book.

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs