The cost of cloud downtime [Video]

Twitter makes you think. Last week, hundreds of tweets were flying when a cloud host had connectivity issues with one of its datacenters. I know intermittent IT infrastructure issues are nothing new, but I started wondering: What happens when something fails and you don’t own the infrastructure?

Thanks to Twitter, I got to watch the journey from frustration to satisfaction firsthand. Around 3:50 p.m., the host’s engineers started working on the issue, and to their credit, they posted regular status updates to an external blog for their customers right up until 4:25, when service was restored. So what started as a minor service outage lasted approximately 35 minutes. That’s 35 minutes of downtime for a customer. If I were a customer, I’d think 35 seconds of downtime was an issue.

And it wasn’t just one customer. Even if there were only 100 customers on this infrastructure, that’s roughly 58 hours …