If a Butterfly Flaps Its Wings, Will The Cloud Go Down? Why A Bulletproof Cloud Is Hard To Build

Microsoft BPOS customers were without email last week due to a cloud failure. Two months ago, Google had it’s own cloud outage and problems losing Gmail data. Amazon had a major failure last month. It’s enough to make you want to keep that legacy on-premise infrastructure.

I remember the first time I was exposed to the “butterfly effect”. It was the mid-90s and I was reading James Gleick’s book Chaos. The question posed was, if a butterfly flaps its wings in Brazil, does it create a tornado in Texas? If that doesn’t make sense to you, let me explain.

Gleick’s book is about chaos theory, and one of the tenets of chaos theory is that small changes in initial conditions can have large system level impacts. So what does this have to do with cloud?

Amazon, Google, Microsoft, and other large public cloud players are building infrastructure at a level of scale never before achieved. This means they have to innovate. They are encountering problems that those who blazed the trails before them never encountered, and they have to find solutions. I can’t prove this next statement, but it seems like a reasonable assumption to say that innovation at scale on large complex systems and perfect reliability are mutually exclusive. What I mean is, while innovation will always lead to less downtime, major outages will never cease because we are continually pushing the limits of scale and complexity, and it’s those unknown unknowns that get you. A butterfly flaps its wings and Amazon Web Services goes down for a half day.

Being in a business that backs up cloud data has been a unique experience because every day, a few more people come to understand the principle that innovation precludes perfection. Two years ago, everyone thought we were crazy. Why would you backup Gmail? Who needs to backup Facebook? Do I need a backup of Salesforce?

In 2010, the earlyvangelists showed up. They realized that we didn’t solve the hard drive data loss problem by building 100% bulletproof hard drives. Instead, we copied the data to other hard drives, to tape backups, and in many cases, moved those other copies to offsite locations. They realized that Backupify was that safe offsite location for business cloud data.

This year has been the year of the early adopters. As more people want to move to the cloud but are worried about data control, data backup, business continuity, and more, our signup rate, close rate, and pretty much any rate associated with backing up more data has increased.

Cloud vendors always fight us in the beginning. Gmail backup was not a popular topic at Google at first. I had several phone calls with Google’s PR team because they didn’t like our use case, and didn’t want us focusing on data loss in the cloud. But over time they came around and saw that we helped them overcome a sales objection that was real for many users. They realized that backup was a good thing.

So back to my original question about butterflies and clouds… how do you build a bulletproof cloud? You don’t. Instead you build a bulletproof data management process. I’m not sure any large-scale cloud growing at a rate that requires regular new innovation can ever be 100% reliable. The way you get around that is Backupify. Have a second copy of your data in a totally separate place. If you lose either copy, it doesn’t matter, because you can replicate the data from the other.

Every time a cloud goes down, Techmeme is filled with posts about why cloud is or isn’t dead, whether we will see more or less of these failures in the future, and a whole host of other opinions that don’t really matter much in the big scheme of things. The problem was actually solved a long time ago by smart I.T. people, who are doing the same things with their cloud data that they did on-premise. There will always be butterflies, thus there will always be backups. If you need one for your mission-critical business data, check us out.