Google Had an Outage: Let's Not Blame Cloud Computing This Time

Every time Google has an outage the press immediately jumps on the idea that somehow cloud computing itself is to blame. This is a wrongheaded viewpoint, because it is not simply a cloud issue, it is a total systems issue, much like anything else that has to do with any computing, outages happen anywhere in the link chain. This isn’t an issue with cloud computing, this is an issue with all computing.

The reason why this is an issue is that some in the popular press have noticed that Google has had seven outages in their online services over the last year. Google is the poster child for making cloud computing work, Gmail, news, applications, advertising, and just about anything else you can think of including documents, Google has their hands in the system. The unfortunate problem is that every time Google takes a hit, people start blaming cloud computing. What needs to be noted though is that all systems are not 24X7 always on 100% anywhere or in any company. Cloud computing is like any other computing system when it comes to fail over and backups, you got to have them and they have to be automatic.

If you have suffered an e-mail outage at work (and work does come to a halt when that happens) or a shutdown of a data center at work to test disaster recovery or to upgrade power will know that these things do happen. We are lucky to get 99.99% uptime. Even Amazon Web Services has had the occasional outage, and few seem to freak out and blame cloud computing. When Seattle had the Fisher Plaza data center fire, more people were laughing because no one had a back up site that they could fail over to, it was not a data center issue, it ended up being an issue with disaster recovery.

It is wrong to blame cloud computing for this, cloud computing is not a singular application; it is a series of systems that reside off the company data center. Like any other system there should be disaster recovery, fail over, and anything else you would have if you owned your own computers in the company. Cloud computing does not make you immune to “bubba and his backhoe” cutting the fiber to the data center. What cloud computing does do, and does wonderfully is provide a space in someone else’s data center for you to do company functions.

Data centers, any data center anywhere is vulnerable to many things, from nature, to intentional accidents. Cloud computing does not reduce the physical risk to a data center; what it does is push that problem off on someone else. That does not mean that a company using cloud computing is not responsible for backups and disaster recovery, it means they can have a hot spare site ready to go within minutes of a systems failure by leveraging the many cloud computing solutions out there. An outage in Google is unlikely to influence Amazon or Rack space, and that holds true across the major providers. You rarely hear about outages at Zoho (also a cloud solution), or at other cloud based solutions like salesforce.com. That is because they are all over the place. While Google inherently has to use its own resources, an outage on Gmail is not a reason to forget to go to the cloud.