I hate articles like this.<p>Outages happen. Stuff breaks. Plan for it. When it happens in a way you don't have a plan for, update your plan.<p>Last I heard, reddit was operating with 1 sysadmin. If they had to employee the staff to run their own hardware AND own their own hardware, I suspect we wouldn't have reddit: it wouldn't be profitable. I'm also part of a startup, and thanks to AWS, we were able to be in two separate data centers from day 1. In the past companies, it's a struggle to get one up and running.<p>Plus, if they were running their own datacenter and had a 'network event', don't you think it would be down anyway? In fact, I'd expect more outages due to less available capacity to fail over to.<p>Running your machines on EC2 is no less risky than running them at Rackspace or any other non-cloud hosting providing. If there is an outage in core infrastructure, you're going to be down. If you think a 99.999 uptime guarantee means anything when the shit hits the fan, you've probably never worked with hosting providers before.