If your company's IT department can guarantee 99.99% up time for email and other server applications they are not realistic, and honestly, are risking their jobs by making the claim if you actually hold them accountable. If Google can truly achieve a 99.9% SLA, they are better than 99.99% of IT Departments in the US.

David, what more could Google have done? They were immediately aware of the problem and posted timely updates on the status dashboard. They even posted a workaround stating that users can access their email with an IMAP or POP client such as Outlook or Thunderbird.

From an ITIL® perspective, this means Google's event management and incident management processes are very effective. Users don't want or care about mind-numbing technical details during the incident, they just want service restored as soon as possible, which is the primary objective of incident management.

A detailed explanation of the cause and a permanent resolution (i.e., service improvement) should be provided after performing a root cause analysis. Again from an ITIL® perspective, this is the outcome of the problem management process.

Your suggestion that Google needs to "add another nine" to their service level is ludicrous. Problems will always happen. Even the electrical grid fails from time to time, and although its service level might be better than 99.9%, it's still not perfect.

If an organization can't live with 0.1% down time (an average of 10 minutes per week), they should either plan for this contingency or provide the service internally. Personally, I think most organizations would be hard pressed to deliver a better service level themselves.

You and the other "prophets of doom" are blowing this incident way out of proportion. Instead of just focusing on failures and demanding near perfect service levels, we should be more concerned with how service providers respond to incidents and what steps they take to mitigate future occurrences.

Read this article for a different perspective on the Google outage.

It's a bit like airline disasters. They don't happen very often, and statistically air travel is still much safer than the car. But when they do, they affect many more people and gather a lot of column inches.

It also reminds me that way back in the cradle of IT, when I was a raw postgrad, I had to work hard to persuade the powers-that-were that if the University computer went down for two minutes when I'd just spent an hour on an unfinished problem, I lost 60 minutes' work not two. The impact of outage depends on what it interrupts and when. It's not just the uptime percentage.

Google's success depends at least partly on the assumption that their mega-scale infrastructure is more reliable than an enterprise can achieve for itself, because it's their primary business and they're better at it. It's that assumption that's being called into question.