James Burgess

As you may have noticed, Twitter has taken over the world. Well, not quite. The primary reason for that is the huge problems they’ve had with scalability – but I suppose that’s what you get when you choose Ruby on Rails.  However, the guys over at Twitter are considering changing architectures to something more scalable.

The question now, is, will this harm Twitter in a big way? The service has experienced some significant downtime (about 2%), equating to about 5 days of downtime during this year alone. Many of these problems have been attributed to the underlying Rails architecture, which seems to have caused Twitter nothing but hassle since they started trying to scale it. We just have to wait and see what they do about it…

I think there are two main issues for Twitter to overcome:

1. What architecture to move to? The possibilities are endless (well, not really, but the phrase works well). Prime candidate is probably moving to PHP or, more controversially, Java – both of which will have to be developed pretty much from scratch.
2. How to make the switch without losing functionality or uptime? It’s going to be a very brave Twitter that chooses convenience over uptime – and I think that many users will find more downtime to be a little too much to keep them hooked. The three-day outage of earlier this year was a real blow to the Twitter demographic. I, for one, saw the scale of the downtime and considered abandoning the whole idea of micro-blogging.

The high and low of this situation is that Twitter needs to fix itself, and very quickly. If it keeps suffering downtime, it’s going to lose users quickly – but if it switches from Rails to another back-end, it can expect significant teething troubles, which, again, may result in many lost users.

I’m not sure how Twitter can, or will, recover from this, but I can’t see it as anything but a lose-lose situation for them. Of course, there’s a slender chance of a win, but its more of a ‘save ourselves from completely and utterly falling off the face of the Earth’ than a ‘win’.

It would be a shame to lose Twitter as an idea, but someone needs to lose this implementation of it, and quickly.