It’s 2013, and yet two big questions still dominate the discussion any time a sufficiently large number of cloud computing types gather in the same room: How many players can the market support, and are cloud resources a commodity?
The topic arose at the clouderati-filled Cloud 2020 meetup in Las Vegas last week (where someone suggested we’ll have a cloud duopoly of Amazon Web Services and Google) and it’s back in the public eye again this week with the general availability of Google Compute Engine. I think we might get an idea how the cloud computing market will play out by looking at the fast-food industry.
The analogy goes like this: Fast food restaurants offer their consumers essentially the same things as public clouds offer their customers — convenience, speed, standardization, flexibility and everything else that comes with not having to prepare a meal from scratch or deploy applications on physical gear. And if all anyone wanted was fast, cheap hamburgers, fries and maybe some sort of chicken sandwich, the more than 33,000 McDonald’s across the world would probably do the trick.
However, when I come to any major intersection in a big city (and even in some small towns), I usually see no less than two national fast food chains taking up corner real estate. If I drive a little down the road, I’ll likely see a few more, and possibly some regional chains thrown in, as well.
Not all hamburgers are created equal, it seems.
Why should cloud computing be any different? If all anyone wanted was a virtual server, they’d probably go with the omnipresent Amazon Web Services. But when features, price, security, network connectivity and related services come into play, it becomes easy to see why there’s such an appetite for more options.
Amazon is to McDonald’s as Google is to …
Amazon Web Services = McDonald’s and Yum Brands rolled into one: AWS is to the cloud what McDonald’s is to fast food. It was the first, it’s the biggest and it’s the best known. All things being equal, there would be no reason for anyone to go anywhere else for cloud computing because AWS delivers reasonable services at a fair price (sometimes downright cheap), is omnipresent and can pretty much handle whatever scale you throw at it.
Only, if we consider the virtual server the hamburger of public cloud, the object store the French fries and the cloud database a chicken sandwich, AWS starts to look like a lot more than just a McDonald’s. You might look at it more like Yum Brands, the parent company of Taco Bell, KFC and Pizza Hut. The Amazon platform is about far more than just machine images and some standard storage and database features. It has myriad services covering everything from configuration to big data, and they’re all designed to integrate tightly with one another — like one of those KFC/Taco Bell combination restaurants that dot the urban landscape.
Rackspace = Wendy’s: Wendy’s is the No. 2 fast-food franchise in the United States, a title I think Rackspace probably holds in the cloud space (although assessing cloud market share is a little more difficult than assessing fast-food market share). And much like Wendy’s places a premium on the quality of its products, Rackspace places a premium on the quality of its service. CEO Lanham Napier has gone so far as to say it’s “playing a different game” than Amazon.
What he means is that Rackspace doesn’t need to compete with AWS by constantly driving down prices because Rackspace customers value service and will pay for it. Maybe, but the company might take a hint from what’s happening with Wendy’s as it struggles to maintain its No. 2 status against a feisty Burger King that’s largely following the McDonald’s playbook. If market share is important, higher prices aren’t often the best recipe for maintaining it.
Google = Burger King: That cloud version of Burger King nipping at Rackspace’s heels is Google. It already has all the standard fare in servers, storage and databases, but it’s also hipper than the rest (or at least it tries to be), it takes some chances on product design (sometimes to the love-it-or-hate-it extreme) and, like Burger King with the Whopper, what it does well, it does really well. In Google’s case, that’s perform at scale.
If Google keeps adding services and cutting the costs of everything, there’s no reason it can’t become the world’s No. 2 cloud provider — some have already bestowed that honor upon it — and maybe challenge AWS a decade down the road.
Microsoft = Arby’s: Despite Microsoft’s best efforts to market it otherwise, Windows Azure is still largely viewed as a cloud platform for running .NET applications and generally doing all things Windows. Not that that’s a bad thing — a lot of people really like Windows and, by many accounts, Windows Azure is a fine platform. It’s like going to Arby’s: the menu offers a lot of things, but you go for the roast beef.
Joyent, Virtustream, CloudSigma et al = In-N-Out Burger, Culvers, Five Guys et al: These cloud providers, like their analogous restaurant chains, are damn good at what they do and their patrons are loyal. They’re typically designed for maximum performance, maybe security, too, and will play around with new infrastructural or programming components in order to maintain their edge. They might even be the best at certain things and have some major customers (I’ve seen Maseratis leaving the In-N-Out drive-thru), but cost, geography or the desire to get a chicken sandwich, too, limit the number of users they can attract.
VMware = Del Taco: According to my colleage Barb Darrow, VMware’s new VMware vCloud Hybrid Service will “be run from partner data centers and sold by VMware’s channel but managed by VMware.” Del Taco sounds like a Mexican place but also has hamburgers, fries, shakes and even iced coffee. And I don’t know anyone who eats there.
OpenStack = Frozen French fries, or cheeseburger-flavored Doritos: It really depends on who you ask (some would even say it’s like kale). If you’re grilling burgers and cooking fries, you’re essentially trying to recreate the fast-food experience at home. On the bright side, when you’re making the hamburger patties and cooking the fries, you can control how much salt you add and ensure everyone who handles them washes their hands. It might turn out great, but it’s never really the same.
Perhaps I’m being overly pessimistic, but I’m beginning to suspect that OpenStack-based public clouds (of the non-Rackspace( rax) variety) will end up being a lot like cheeseburger-flavored Doritos. In name, they’re like cheeseburgers, but after a few bites you’re left saying, “Hey, Doritos doesn’t make cheeseburgers …”
Everyone else = everyone else: Even after all this, we’re still left a bunch of different cloud providers and a bunch of different fast food chains. You might compare the telcos to Jack in the Box, Carl’s Jr. and Hardees in that they’re big and make money, but they’re pretty much non-factors in the grand scheme of things. Then there are your various web hosts and others, which might compare with some local chain restaurants. And different countries will certainly have their own cloud providers just like they have their own takes on fast food.
In the end, though, it’s just hard to see how cloud computing becomes a two-horse race any more than the fast-food industry is a two-horse race. Sure, there are three clear leaders (with No. 1 having a big lead), but there’s plenty of business to go around because aside from some core similarities, no two providers are the same. And as long as more applications are developed and need a cloud to call home, there will be developers and CIOs with very different ideas of what makes a cloud platform great.