Friday, March 6, 2009

Nicholas Carr: IT and economy are cloudy, but the Big Switch is on

For those at IDC's Directions conference this week in San Jose (my highlights are posted here) who hadn't yet read his book, The Big Switch, Nicholas Carr used his keynote to walk through his "IT-is-going-to-be-like-the-electrical-utility" metaphor, and his reasoning behind it. Many, I'm sure, had heard it before (some even complained about it a bit on Twitter). But that doesn't make it any less likely to come true.

IT, he argued, is the next resource to go through a very similar transformation to what happened with electricity, moving from private generation of what was needed to a public grid. Grove's law about bandwidth's growth has "been repealed," said Carr. It is "the beginning of the next great sea change in information technology" -- the move to utility or cloud computing. (For more detail on this, Andy Patrizio of also posted a good run-down.) So far, so good, but here's a question: has anything shifted since Carr's book came out? How far have we come?

Big changes in attitudes on cloud/utility computing since last year

In the 14 months since The Big Switch was published, people's attitudes have changed dramatically. What was greeted (even in Carr's estimation during his keynote) with skepticism early last year is pretty much a done deal this year. We're moving to "simply a better model for computing without hardly even noticing it," he said. "Cloud computing has become the center of investment and innovation."

Now, you can argue about the speed and impact of the changes, but it's hard to argue that the change isn't happening. Gordon Haff of Illuminata published a good thought-piece on whether or not what's happening is really as profound as Carr suggests, and whether data centers are even seeing those changes yet. Healthy skepticism, for sure, especially when talking about the data centers of big organizations.

However, Carr used his time onstage this week at the IDC conference to extend the vision his book laid out and talk a bit about the intermediary steps we're going through in the "big switch." In his view, cloud computing can be many things for organizations. It can be a new do-it-yourself model for IT. A supplement to whatever IT exists that's "shovel-ready." A replacement. A democratizer. Even a complete revolution in which IT and business finally get on the same page, shocking as that may be.

No matter which view (or, more likely, combination of views) ends up being true, Carr was clear to IT folks and businesses alike, if a little understated: "It behooves you to look into this. Figuring out how to harness the power" of a public cloud "may be the great enterprise of this century."

I think on this point Carr, IDC and their analyst brethren, and vendors like my company, Cassatt, and others are all pretty unanimous in what they're saying. Despite some false starts at the beginning of the new century, the big change for IT -- being able to actually use utility-style computing -- is really here. It's called cloud computing. Of course, it's still early days, but let the fun begin. Actually, if you hadn't noticed, it already has.

Now, comes the messy part: making this new model work

Carr was pretty clear that this is going to take a few steps. No arguments here. And while he did joke about the term "private clouds" being an oxymoron if there ever was one (my Appirio Twitter friends loved that one; see also the many previous posts here and elsewhere arguing that topic pro and con), he also talked about private clouds as one of the many steps along this path. In looking at applying the cloud model to the data center infrastructures organizations already have, we are asking "what can we learn about how the cloud operators do IT -- and revamp our own data centers on the cloud model." To Carr, internal cloud computing sounded like a "big opportunity while a public cloud is being built out."

This is also where things get messy for today's big IT suppliers, thanks to a little thing called the innovator's dilemma (from Clayton Christensen). "Big traditional vendors are in quite a fix today," said Carr. "You have big vendors making investments in cloud computing, but it's almost on faith because they haven't figured out how to make money on it yet. The problem is not in the technology, it's in the business model."

Add to this a tidbit from a survey IDC's Frank Gens talked up earlier in the day: the least important buying criteria for cloud services were that the company be a large, established company, or that the company had done business with an organization before. Well, now...that certainly opens up the playing field quite a bit.

Carr pointed to a number of hurdles that still exist, like having to hit a level of reliability beyond what's been required today just to prove all this cloud stuff is safe. However, based on what I hear from all sides (especially customers), I think the engine of change has really started up on cloud computing. Everyone is talking about it, testing the hypotheses from all angles. But that's the easy part. More importantly, though, people are trying it. Some are moving fast; some are moving cautiously. Some are trying external clouds; some are applying the cloud concept to the resources inside their own data centers. And, the "cloudy" (er, grim) economy, as Carr called it (and as Frank Gens noted earlier in the day), is probably helping in its own way, too.

As unlikely as it may have seemed when his book came out last year, Carr's "big switch" is on.


Anonymous said...

I'm sorry, but there is scant evidence to support the view that Carr's 'big switch' is on. While I concur that it is undoubtedly a direction organisations will move in, it is highly unlikely that Carr's 'vision' will be achieved in any short time scale for reasons both architectural and organisational. IT does not function like electricity - not unless we see the latter from the perspective of microgeneration. Which, of course, we don't.

Cheers, Jon

Jay Fry said...

No argument about the timeframe, Jon. The shift to cloud-style architectures and approaches for data centers is going to certainly be a gradual one, not something immediate.

The evidence that we at Cassatt are seeing of a "big switch" toward utility computing in the form of both internal and external clouds is directly from our customers and prospects. I know there is tons of hype around this topic at the moment, and it's really hard to separate what's real from what's vendor-generated marketing fluff. But, end users actually are taking initial steps into this space based on the kinds of questions they are asking us and the projects we are beginning to help them with.

Sounds like it would be useful to spend one of my next blog posts on the types of things they are actually asking our field people about to help provide some examples here.

Thanks for your thoughts, Jon.