If I had built Twitter (from a network perspective) I would have had it virtualization ready, such that if my primary carrier had to do some local maintenance, I could migrate an image and then move it back. I would have a highly distributed and available system -- without a single point of failure.

In the modern world of virtualization and the cloud - downtime isn't as necessary as it was 10 years ago. The core data itself - users' tweets - are not the items being updated. That means (in my view) that Twitter can aspire to 100 percent availability.

Databases, core software and connectivity can all more be upgraded and go down, but if the content is free and virtual it can migrate across always on virtual images wherever they might be.

Yes I'm suffering from Twitter withdrawal right now (otherwise this post would only have been 140 characters long). Let's hope that part of the reason for the upgrade is enable better scalability such that in the future Twitter can be an always on service.

It's not to much to ask for - in fact I think that if Twitter is truly to remain the killer app that it has become, it must reach the same level of reliability as the Internet itself.

(and yeaah, I know Twitter was only offline for an hour, not the end of the world, but hey I'm addicted!)

Follow me on Twitter @TechJournalist