We are now in the Oprah stage of cloud market development. As in, you get a cloud, you get a cloud, everybody gets a cloud. (Picture from Wikipedia.)
Game company Zynga (ZYNG) is among the many who have just announced they will build their own cloud. Zynga's will be a public cloud, in that you're invited to play games there, and it will take on third-party hosting duties for other game companies.
The enterprise computing market is moving in the same direction. While worrying about issues of security and privacy, many companies are building their own clouds and moving simple functions like web hosting into them. This is providing big money to companies like VMWare (VMW), Citrix (CTXS) , and RedHat (RHT), which sell software to these clouds, as well as Hewlett-Packard (HPQ), Dell (DELL) and EMC (EMC), which sell hardware.
So far, so profitable. But there are two problems. First of all, these aren't really clouds. Second of all, the boom is unsustainable.
If an asteroid fell from the sky tomorrow and destroyed Apple's (AAPL) new iCloud hosting center in North Carolina, Apple would be in trouble. The workload would not be transferrable. They have taken Mark Twain's advice to "put all your eggs in one basket and watch that basket."
That's not cloud. A data center may provide cloud services. But it's not a cloud. A real cloud has the resiliency and the redundancy of the Internet from whence it sprang. A real cloud can move those workloads, provide back-up, and route around trouble in the way a single fixed hosting center can't.
This truth is being ignored right now in the rush by companies to build the own clouds, by cobbling together servers, trying out cloud software, and scaling in place.
Which brings me to the second problem: The boom is unsustainable.
Forgive me, but I have seen this movie before. I have seen it many times. Most memorably, I saw it in the late 1990s when company-after-company installed fiber cables around the railroad tracks near my house. Rail tracks and pipelines were great havens for the bandwidth boom of that time.
But Wavelength Division Multiplexing (WDM), which I once called "Moore's Law of Fiber," quickly allowed a single fiber to to the work of two, then four, then many, by running traffic in different colors. Most of that fiber went unlit and unused. Google (GOOG) scooped up a bunch of it early in the last decade at bargain prices.
Something similar is going to happen here. Once everyone has a cloud, once clouds are a commodity, the cloud boom will be replaced by a cloud bust. And it's the less-economic clouds, especially the one-site clouds, that will fall first, with enterprises moving their (now mature) workloads onto cloud networks, based on redundancy and total cost of operations.
The winners at this point will be those with the largest, most efficient clouds, those whose clouds are resilient in the way the Internet is resilient, that can route-around problems and move workloads to wherever they need to be.
In a few years, it won't be everybody gets a cloud. It will be best cloud wins. And the best cloud is not a single data center. It's a network of such data centers. An Internet, if you will.