|HOME||PUBLIC LIBRARY||ANOTHER PERSPECTIVE||INFOPERSPECTIVES||CONTACT|
Henry Ford started selling Model T cars in 1908 at $825 a pop. Ten years later the price had come down to $360, and every one made was the same color. As Henry Ford said, "Any customer can have a car painted any color that he wants so long as it is black." Despite this lack of variety, the Model T accounted for half the cars in the USA. Today, Henry Ford might not make cars; he might build machines for cloud computing, where he could tell customers to take an X64 or take a hike.
You might think that any company with proprietary hardware--IBM, Oracle, and Hewlett-Packard come to mind--would hate this, but maybe that's not the case. The way it looks right now, cloud computing is not an alternative to most in-house information processing. On the contrary, it is a new kind of service that lets organizations do things that are difficult and impractical to do on systems in a corporate glass house or a small company's pack of racks.
Before I go much further, I ought to acknowledge that the big name computer vendors, particularly IBM and VMware , are trying to pitch something they call a private cloud to their customers. IBM thinks this is going to be a big business, but what IBM calls a private cloud and what Amazon, Microsoft, Google, Yahoo, and a number of startups call a cloud are really quite different. The minute you buy a bunch of servers from IBM (or any other supplier) and set them up with capacity provisioning systems that in some ways resemble the cloud offerings of, say, Amazon, you are just plain kidding yourself. You won't have computing on tap and more importantly you won't have computing on a tap you can turn down as well as up. What you have is a new kind of resource management setup, maybe a very good one, but you still have a collection of capital assets with fixed value rather than computing capacity that's not an asset but which is a variable expense.
So hooray for IBM's marketing folk and good luck to all the Eskimos who want to buy this particular refrigerator, but anyone whose company might be able to successfully exploit cloud computing had better refer to the dog leg riddle. (You don't remember it? Okay. If you call a talk a leg, how many legs does a dog have? The answer is four, because calling a tail a leg doesn't make it one.)
Now it turns out that all the brave talk about abstracting and virtualizing hardware is to some extent an empty boast on the part of the entire IT industry. You can get access to X64 machines that provide virtual environments based on Linux or Windows (or some other environments, if you really want 'em) more or less the way you buy electricity, except you can increase, decrease, and reshape your IT capacity a lot more easily than you can add or remove a thousand amps of electric service or change the diameter of your gas main. If you want to add some processing power, a few terabytes of storage and one more database engine to your Amazon cloud setup, no big deal. Same goes for getting a few hundred people hooked into a document creation and sharing system with the help of Google or Microsoft, or boosting your boutique storefront on Yahoo into a department store.
The same goes for companies who prefer the offerings of cloud outfits that are not (yet) behemoths. You can redial capacity at Rackspace Hosting in ways that probably dwarf you total in-house IT processing and storage capability. And if what you want to do is kind of specialized--you want your sales team to have access to every IT tool all your rivals have and then some--all you have to do is milk the peddler's cow that Salesforce.com has already built. (And if you don't like how they work, you can find alternatives.)
Google, when it's not building cars that robotically run over your privacy in the kind of mashup of lug nuts and wing nuts that would have made Henry Ford proud, is trying to dream up a bunch of cloud offerings that can feed the dreams of small (and not-so-small) business that think the Internet might be able to broadcast their offerings to the universe. Google wants to do this better than Microsoft, which has proclaimed similar intentions just in case OpenOffice survives the cynical, if brilliant, forces of Ellison Wonderland.
If often happens in computing that new ideas allow new companies to come to the fore. The ultimate winners might not be either the legacy operators (and Google and Microsoft are old legacy players when it comes to the cloud). Just as Google was not the first search engine, just as Microsoft was not the first outfit to make operating systems for Intel-based computers, there are lots of young companies that think the cloud might become their empire. There's Box.net, which lets companies combine best-of-breed offerings from across the industry and, if all goes well, integrate the data and functionality of disparate specialized cloud services. And there are alternatives to in-house information management systems, cloud services (in some cases built on top of Amazon's services) that let dynamic companies pool information, share (and not share) various files and file permissions, all without the kind of data administration technology that would otherwise require a server farm, a storage farm, and a very savvy team of farmers to manage it (and bring in all the new projects nice and late).
There's Dimdim, which talks as if it can do all the collaboration type stuff IBM's Lotus promises, but which distinguishes itself from Lotus because it is free. There's also Dropbox, another freebie that lets users tie together cloud services in ways that are, in some dimensions, boundless. The list goes on and on and on.
I don't know if free cloud computing services that help businesses use the economical but not ways user friendly fee-based cloud offerings of Amazon, Microsoft, Google and others are the future, but in computing free stuff often does seem to catch on, particularly when the people behind it are very, very smart and the people in front of it are not rich.
It looks like just about all the lively stuff coming up in cloud computing is getting built on X64. IBM talks about offering Power Systems capacity in the cloud or helping mainframe users build clouds in their multi-million-dollar System z boxes, but anyone can see that this stuff isn't going to get anywhere because the skills pool is thinking Ruby on Rails or serving radical XML and HTML5 to clients living inside Windows 7 or Apple's next big cat litter nixoid.
If you want some kind of economic basis for a belief in the cloud and an expectation that it is going to be a new and different world, not some junk bolted onto what's running in your data center, all you have to do is see what the top industry analyst at IDC has to say. Frank Gens has become so excited by the possibilities of the cloud computing world the he gone all generous. He drew a few pictures to explain what he thinks is going on, and they are just plain stunning.
Basically, he says by 2014 the cloud biz, the public part, which as we said is the real and really important part, will be worth something north of $55 billion a year, more than three times what it was last year. He says that in 2014, apparently his favorite year so far, net new IT spending for cloud stuff will be nearly half as big as growth in traditional IT, which is more than enough to let some new companies become real biggies. He has tied these two concepts together as follows: for the next few years traditional computing (including the stuff that's called private cloud but is just another form of buy-it-and-live-with-it-every-day computing) will grow, on average, at 5 percent a year. Cloud computing, however, will grow at 27 percent per year.
So which companies are going to be up in the clouds? Well, that's the best part of the whole story. Nobody knows. Which makes it not only interesting, but fun, too.
— Hesh Wiener October 2010