Once upon a time, whenever anyone asked, "But are there any big applications built on Rails?" The answer was usually, 43Things, anything from 37Signals, or Odeo. But over the past year, there's no doubt that if there is a poster child for Rails, it is now Twitter. With such notorious bouts of downtime, a worse poster child Rails could not possibly hope for. But is Twitter even the largest application out there running on Rails? Does it even matter?
"Twitter is almost certainly the largest site running on Rails, so fans of the framework and its developers have been quick to deflect the criticism and point it back at the engineers at Twitter [to explain downtime]," wrote Nik Cubrilovic in a recent post on TechCrunch calling out Rails as a poor choice for large scale app development. The debates over what causes Twitter's frequent outages (we think it's a database issue) and whether Rails is good for large apps aside, Twitter might not actually be the biggest Rails-based app out there anymore.
Some back of the napkin math by noted rails developer Evan Weaver (who recently went to work for Twitter), finds that while Twitter might be huge in terms of monthly pageviews, the Facebook app Friends for Sale, may still be bigger. And Yellopages and Scribd are similarly massive.
Ignoring the oddities in Weaver's computation (like, for example, that even though he works at Twitter he only guesses how much traffic the API is fielding), which he admits result in "wildly inaccurate values," he makes one very good final point: It doesn't matter!
"It is important to keep in mind how useless this information is. It doesn't even make sense to say 'Rails site' or 'PHP site,'" says Weaver. "Livejournal uses Perl, Memcached, and MySQL, among other things. Does that make it a Perl site, a MySQL site, or a C site? I don't know what Scribd uses, but it's pretty likely that their document pre-renderer is Java or C, not Ruby. Friends for Sale uses Nginx, Rails, Memcached, MySQL, and Linux. Ruby is really just a little piece of the pie."
Comments
Subscribe to comments for this post OR Subscribe to comments for all Read/WriteWeb posts
would like to point out that there is a difference between 'rails' and 'ruby' - and that a framework like rails can act as a chokepoint for an app in a way that choice of language to code in does not.
Posted by: mathew | May 27, 2008 10:50 AM
Pretty sure that's called splitting hairs?
Posted by: Jon Gilkison | May 27, 2008 10:53 AM
Of course it matters. MySQL, Perl, etc have been used in many, many large (however you define that term) sites so we know that they CAN scale and whether they do or not is an implementation issue. WE don't have the same broad set of examples for Ruby or for RoR, yet RoR is gaining a lot of popularity among developers. So it's a natural question to ask - if the app gets big, will RoR fail to scale?
Whether the answer is "No, look at Twitter" or whether Twitter's issues should be laid at the feet of something else we don't yet know. Hopefully the Twitter will do a public post-mortem so we can all learn from what they're going through.
Posted by: rick | May 27, 2008 11:32 AM
@matthew
absolutely correct. Its not the language itself that's wrong, its how the language is used. Any language used can go wrong with bad coding.
In an application there's multiple areas for bottlenecks. The thing is you don't want to introduce something that's known to create more bottlenecks, such as using a framework that's known to cause issues.
Posted by: cease | May 27, 2008 11:33 AM
I hope Rails rise above the whole Twitter mess because it sure got an undeserved black eye because of it, enough said.
Posted by: Bob Ngu | May 27, 2008 10:01 PM
Of course and i thought rails was just at the top come on guys...it is just redicolous.
Posted by: Sesli Sohbet | May 27, 2008 11:50 PM
A lot of naive people here, including the author.
You say that Ruby is just a piece of the pie? Did you think that some people though databases were Ruby too, and web servers where Ruby too?
I don't think anybody's denying that except maybe someone who just started programming yesterday.
For one thing, since you're reporting, why don't you make a phone call and find out what their chokepoint is? Instead of just speculating your ass off, and me wishing I could get my two minutes back reading it, and 1 minute back writing this.
Posted by: Scaler | May 28, 2008 3:30 AM
Look at http://rails100.pbwiki.com/ for biggest rails sites.
Posted by: edbond | May 28, 2008 6:03 AM
How do you measure "biggest"? Is it traffic? Requests/sec? Unique visitors per month? Number of members? Amount of data? Size and complexity of the model?
Posted by: jeremy.weiskotten.com
|
May 29, 2008 7:21 AM
See here as well:
http://pantherfotos.com
Posted by: druido | May 29, 2008 7:34 AM