My wife hates flying. As in, she doesn't fly. And gets really grumpy when I have to fly. As she points out, usually there aren't many survivors in a plane crash. When you go down, you go down spectacularly. She sounds remarkably like Ed Bott in his post on Google's Blogger outage.
The same week that Google made its strongest pitch ever for putting your entire business online, one of its flagship services has failed spectacularly.
Earlier this week, Google rolled out a maintenance release for its Blogger service. Something went terribly wrong, and its Blogger customers have been locked out of their accounts for more than a day.
No doubt, it's been a rough couple of weeks for the cloud. Microsoft, Google, and Amazon have all taken big hits to important services.
My wife is happy to tell me that if you're driving, you stand a chance of survival in a crash. Besides, at least when you're driving, you are in control. Or so she says. She's never been T-boned or rear-ended by another driver. And she's not a big fan of math. She's more of liberal arts sort of girl. But statistically speaking, driving is far more dangerous than flying (of course, depending upon the statistics you choose to use). Since I'm a person-miles sort of guy, though, I'm sticking to the flying is safer schtick.
So how often does an individual computer crash and burn compared to the number of times Blogger or another cloud-based service go down? How many users do you know who lose everything when their hard drive bites the big one? And how often do users and consumers end up getting hosed by the Geek Squad to deal with their computer problems.
The answer to all of those questions is "a lot."
It's true, as with my wife's flying concerns, that when BPOS goes down, a lot of people feel it in critical ways. It's also true that widespread data loss is extremely rare and the last thing Microsoft (or any of the big cloud players) needs is for people to not adopt their services because they believe the cloud is unreliable. So the entire engineering mite of Amazon or Google turns to fixing the problem instead of the local Geek Squad.
Do cloud providers need to adopt more robust practices and improve failover processes? Of course they do. As we all hop on the cloud bandwagon (some more willingly than others), outages like Blogger's will be entirely unacceptable. But do we need a hybrid strategy? I'd argue that we don't, particularly for those who adopt Chromebooks and various tablets as their primary computing devices. When all you have is the "screen," where will you do a backup? And why should you have to if cloud providers are doing their jobs?
Obviously there will be bumps in the road during this transition. But instead of adopting or developing the latest solution to only sort of use the cloud, let's improve infrastructure and bandwidth; let's develop new failover tools; let's do a better job of replicating our data.
Seriously, if I have the choice between flying and riding in a car that my wife is driving, I'll take the airplane any day of the week. I'll take my Google Apps to the Geek Squad any day, too.