Marcelo Calbucci

    It is very common for Web 2.0 companies to suffer major performance or scalability issues (usually both). This is not different from outages on the Web 1.0 age. Even major companies like Amazon and eBay suffered a lot of growing pains with reliability of their service.

    Recently, a key person at Twitter (a growing Web 2.0 startup) complained about the pain of making Ruby on Rails scale.

    About 18 months ago, another entrepreneur from Seattle asked me what he should use to build his new service. He knows quite a lot of technology, but he is not a developer, so, for him, it was a matter of what skills he would be looking for on a developer.

    I'm no fanatic. I believe that if you are an expert on PHP, you should build your company on PHP. If you are expert on ASP.NET, you should use ASP.NET. If 50% of your developers know ASP.NET, and 100% know Ruby on Rails, you should use Ruby.

    On his case, he was in doubt between Ruby on Rails or ASP.NET. I told him both would be fine, although I don't really know anything about Ruby, except that Ruby on Rails is an unproven technology. It doesn't mean it is not going to work, or it won't scale, or it won't be reliable, etc. It is just there are not enough sites built on it. Nobody questions a website built with Apache and C++ ISAPI. It has been done a million times and it can scale beyond anything imaginable.

    He has chosen ASP.NET. His service is running and growing fast and he had not had any problem with scalability of his application or architecture yet. He did had some storage issues though, but that was independent of framework or platform.

    The lesson: Build stuff using the tools you know. Better yet if what you know is guaranteed to work.