Biggest Cloud Outages in 2011

I recently saw an article on the biggest Cloud outages in 2011. Since outages (loss of connectivity) are typically one of the reservations that many people have about moving to the cloud, I was curious.

The biggie of course was the 5+ day Amazon outage in April that affected a large number of major companies.  Microsoft cloud services – Office 365 and the Business Productivity Suite – suffered 5 outages (4 in the month of May), perhaps enough to raise doubts about the reliability of those offerings.

Other than that, there were two outages caused by a major lightning storm in Ireland. Google Docs and related apps went down for about an hour in September, and Gmail lost some 150,000 mailboxes in February. Other of the “10 biggest” were in fact pretty minor – half hour to an hour, or simply multi-hour delays in delivering email.

What this all means is that cloud-based services are pretty reliable, almost certainly more reliable than the systems in the average law firm (or Comcast cable). The lightning-storm outage is noteworthy because the backup power systems failed to kick in.

This of course is no consolation if you DO get hit with an outage.  Which raises the question of what sort of reimbursement do you get from the cloud service if you get hit with an outage? Most contracts provide for reimbursement as a percentage of the monthly fee – if you are out for 1 day, you get 1/30th of your monthly fee returned. This is essentially worthless to a firm.

Software vendors long ago faced this issue and most agreements limit liability to the price of the software.  However, cloud-based services are in a slightly different category. One could reasonably make the argument that if a 10-attorney firm suffers a 1-hour outage, and half the attorneys might have been unable to do billable work during that time, then at $300 an hour the firm has lost $1,500 in billables. Obviously no vendor is going to give this sort of guarantee, but you might be able to improve your agreement over the standard chump change for outages.