Straight talk on adaptive computing

Conjuring up a vision of computing's tomorrow is one thing. Convincing buyers that you're the one who has figured out the future is quite another.

Up for grabs is the legion of IT customers still smarting from the buying binge of the late 1990s. Unlike the go-go days, the contemporary reality is light years removed from the era when money earmarked for new projects seemingly was no object.

Within the past year, however, several computer companies have tried to shift the debate. Yes, they allow, acquiring technology for technology's sake is a taste that went out with the Backstreet Boys. The new mantra is to help companies better align their technology and business needs. Fix that gnawing problem and IT return on investment turns into a veritable cake walk.

So it is that the battle between competing visions has been joined. IBM talks about On-Demand Computing, Sun offers its N1 initiative and Hewlett-Packard presents Adaptive Computing.

The risk is that these broad strategic catchalls are open to wildly different interpretations and no small amount of vagueness. Indeed, earlier this year executives acknowledged that IBM's own sales force had little idea what On-Demand Computing was supposed to be about.

CNET News.com recently met with Nora Denzel, senior vice president of HP's Adaptive Enterprise, to find out what she sees on the IT horizon from the computing giant's perspective.

Q: Stripped of all the jargon and market-speak, can you succinctly define what Adaptive Enterprise is supposed to be about?
A: I define AE as a business strategy for customers who want to respond in real time to changes affecting their business.

That could be boilerplate applying to any company. What's the special sauce?
The secret sauce that HP brings is the ability to link business processes--which obviously are a manifestation of a company's strategy--to IT gear. The big breakthrough is when those two things are synchronized, so changes in the business environment can dynamically trigger the IT changes necessary to support that business change.

Can't you get that by going to any reputable company out there? Sun, IBM--that's what they're about. Am I missing something here?
HP's approach is fundamentally different in that you don't have to put in a lot of new technology to get to an adaptive state. If you've chosen .Net or J2EE--we don't offer either of those and don't have a vested interest in you swapping one out to take the other one. If you've chosen Oracle or DB2, again, we don't have a vested interest. And that's what separates us from an IBM, which says, "There's one path--and that path is through my services, my software stack and my hardware stack." If you're considering changing, we can give you advice, but not advice based on a religion or (proprietary) technology.

So what is it?
It's a business strategy for customers who want to respond in real time to their businesses. The secret is when you link the business processes together to your IT gear, then you can automatically roll those changes through and respond.

So how is that any different from what system integrators are doing? When an EDS goes into a company, they presumably are meeting the business needs of the client.
Adaptive Enterprise isn't just a program to get you to outsource. I think the discussion that an EDS would have is "We'll host that and you pay for it--and it's our people and our badges that will do it."

Well, not necessarily. An EDS is as much a system integrator as an outsourcer. The system integration part is what I think I'm hearing from you. That HP will make things work together--the business processes, the technology.
The other part is especially when you're at a new point in how computing will be done, meaning that you link to the business processes. There are huge ramifications. The skill we bring is a very good understanding of hardware and topologies and networking and the manageability aspects of it--along with a very good understanding of the services aspect.

There still seems to be confusion surrounding the topic. At the Gartner conference last month, some IT attendees said they still say they didn't understand what Carly Fiorina wanted to convey with HP's Adaptive Enterprise. Do you feel the message is unclear or needs rethinking?

I disagree that it was unclear. Adaptive Enterprise defines an entity where a company will be able to dynamically readjust to changes that affect its business.

I disagree that it was unclear. Adaptive Enterprise defines an entity where a company will be able to dynamically readjust to changes that affect its business. Carly also said it was an approach. Since you can't buy an Adaptive Enterprise, we have to sit together and build toward it. What I was surprised at was the confusion with people asking, "Well, is it an approach or an end state?" This isn't a product. It's not one piece of hardware where it runs at either this speed or that speed. It's not like that.

Whether it's called utility or grid or any of the words that are used, I think the next 10 years will be typified by the linkage of business processes to your IT resources and dynamically reallocating--I like to say it's the automatic supply and demand matching of your computing resources. So they become much more resilient and adaptable.

But aren't Sun and IBM doing that to make that happen? And IBM and Sun are investing a lot of money to make sure their software works with their respective utility computing programs. What are you doing besides saying we'll sit down with you and work on it?
It's interesting. We've spent $2.5 billion in Adaptive Enterprise. That's a pretty healthy R&D investment, so I disagree that we're not investing in technology.

Is outsourcing part of AE?
It can be, but it is not the only play we're running. First we sit down and ask about the business strategy to set some boundaries for us. Some customers tell us up-front they own the gear, the people are on their payroll and they want to own (their IT operations). Others don't. So it just depends. You can't pigeonhole. Our strategy is to let you become a company that responds in real time to these changes; the secret is linking together the business processes to the resources.

In the broadest definition it just sounds as if this is another form of services business where you want to make money. You're not selling new hardware or software--it's services.
We're entering a new decade of computing--whatever it is called. All it means is that you will use IT resources in the same way that you use a utility like a light. You take what you use. You don't have a generator at your house; there's a power grid somewhere else and you tap into it.

But that's still not most people's experience with computing.
Whether you think that will happen or not is something people will debate for years to come. But in this next phase is this notion that you don't need to overprovision anymore. You pay for what you use. You get to dynamically redeploy assets to something that's more important than what they're on right now. There's an intersection of some of the capabilities of utility computing that are in the Adaptive Enterprise. But I don't think those are synonymous terms, because utility is an overused, ill-defined word.

What's required?
You have to have the inherent capability to respond quickly to any change that affects your business.

But again, how does that differ from what's been around?
Let me give you a customer example. Target has a business policy where someone could buy a product on their credit card and return it for cash. This was a customer convenience. The problem was when someone stole the credit card, Target took five days to respond--and the thieves knew it. So the business decision to return products for cash left Target in a position of potential loss. Now it's less than a minute.

So what's the so-called paradigm shift?
The paradigm shift is that whereas the IT center was once considered to be run by technologists and to be a cost center and supportive of the business, the shift is that the IT department's goals are being inextricably linked to the business strategy and processes.

But hasn't that always been the goal?
In the early days of computing, it was quite difficult to do that. For example, you had computers that lacked industry standardization and so communicating between them required special translators for the systems. It required lots of manual intervention and interfaces that didn't exist, and you had to create the software itself, which was always subject to change whenever you got new releases of the vendor software. The computing infrastructure was somewhat brittle. As a result, it was overprovisioned and you had many more people than needed to manage it.

We are going though a change and it is a shift where IT is run now like a business for the business. It's not a support, cost-center type technology job where the decisions are made on the basis of "this box is so much faster than the old box" or "this is the killer application." It's now about what's the return of investment.

But pardon me, that's been talked about for more than a decade and a half--literally. The other thing is that you talk about overprovisioning--but with the cost of hardware coming down, is that still so much of a big deal? That is, you buy a $10,000 server whereas 10 years ago that would have cost $250,000 apiece?
Like some other things in the world, return on investment was talked about during the bubble, but very few people were practicing it. What's new is that people are giving serious thought to return on investment and saying that IT needs to be run as a business with business metrics. There's much more talk about the financials. That is a definite, profound shift in terms of financial accountability for IT projects. So I disagree with your statement that this isn't new.

Is there culpability on the vendors' part because most of the big IT providers were like the proverbial fox guarding the hen house? They had product arms and services arms, and they were the ones pushing the big multimillion-dollar deals. It wasn't because the CIOs or CFOs weren't thinking about the ROI part of the equation.
The answer isn't as simple as you outlined. I think there was a feeling among many in the industry that they would get left behind if they didn't gear up for the Internet and then Y2K. But we didn't have Web services standards that linked things together. They had been talked about and bounced around the standards committees. But there's since been a huge push for a common language that makes up a business process. And once that happens, you require less IT gear.

In talking with companies that offer these big-sounding initiatives, it sounds as if there's an assumption that IT is in danger of being overwhelmed and that they don't know how to build flexibility into their systems.
It isn't the IT department. The business is in danger of extinction if they can't respond in real time to those changes. And the only way they can do that is to link the business to the computer resource.

So what does that suggest for the role of CIO in deciding upon IT projects?
The lines between business and IT are blurring. One CIO told me they don't have IT projects anymore. It's a business project with IT ramifications in it as well as others. But they don't entertain and kick off IT projects without the CIO being integral. The role of the CIO is changing and that's an exciting part of what's happening in this decade.