Ragunathan, RoR is certainly not my choice for serving dynamic content out to the world.
However, RoR is not the root cause of failure for Twitter. Swapping it with Python-Django or Erlang-ErlyWeb or Scala-lift would probably be a 10-20% improvement at best in terms of content loading time / cpu performance. I am sure there are many, many, many mongrel instances humming away trying to keep Twitter alive :). The bottom line is: Tools and Technology do not make the difference between success and failure. Great people and sound Architectural choices usually overcome the tech. Twitter is now re-writing their architecture after getting a $15m round of funding (I believe that gives it a $70m valuation). What this means is that they finally woke up: They hired some solid guys from Google and Facebook, and they are waking up to the reality that they built something incapable of adapting to the pressures of growth. A good system running on less efficient tech would have gotten by with more servers being thrown at the problem. Glad you liked the post! Cheers, Zubin. On Mon, Jun 9, 2008 at 1:52 PM, ragu < [EMAIL PROTECTED]<[EMAIL PROTECTED]>> wrote: > That's an excellent post Zubin. Though you mentioned it is crazy to > blame it on Ruby on Rails, by suggesting alternative framework as the > solution to the problem, it actually sounds like RoR is not a good > choice when scalability matters. > > >
