Is Twitter blaming the user for its downtime woes?
At first, it seems as if the popular web site Twitter is blaming the user for its downtime woes, especially heavy users like Robert Scoble, who have in excess of 20,000 friends and followers. In fact, the downtime scalability issues come from several directions at once.
The first issue is the platform used to build Twitter. Ruby on Rails is a popular programming framework that is being used to create a wide variety of simple applications like Twitter, Basecamp, 43Things, ODEO, Tadalist, Penny Arcade, Wayfairing, Jobster and more.
Ruby on Rails is so popular because it is a fairly clear framework (Rails) using a concise, easy to use language (Ruby). This means you don’t have to worry about multiple dialects, languages and syntax issues while programming – you can just concentrate on getting your application up. It is well suited to simple, low key applications with a tight focus.
It is that simplicity that is also the downfall of Ruby on Rails when an application created with it starts to grow exponentially with popularity. Between the use of Ruby on Rails and a lack of investment in server side infrastructure, site like Twitter are doomed to have problems. Add in the release of an API to developers and the use of site like FriendFeed and its constant calls to the API with no setting on FriendFeed to dial it back a notch to under 10 API calls at a time, and you have trouble.
In the end it isn’t the users fault that Twitter is down, or not all of them anyway. Luckily for the Twitter addicts like myself, the powers that be have plans to improve the scalability of the site. Meanwhile, they have taken steps to make the site go into Grey Mode in times of stress, and have plans in place to handle spammers, lengthy follower lists and other issues. They were interviewed by Robert Scoble earlier today and addressed many user concerns:
Related Posts:
