By focusing on high-value, high-volume systems, Dell Inc. has achieved success in both strong and weak economic cycles. It has also made strong partnerships with computer chip and operating system heavyweights Intel Corp. and Microsoft Corp. Chairman and CEO Michael Dell is focusing the company increasingly on what he calls “the scalable enterprise,” which has extended Dells offerings into new areas such as printers and services. Dell explained to eWEEK Editor in Chief Eric Lundquist, Executive Editor Stan Gibson and Senior Editor Jeffrey Burt his ideas about what the company will and wont do, in an interview last week at Dell headquarters in Round Rock, Texas.
Can you update us on Dells newer initiatives with printers, services and storage?
The fastest growth has been in management and deployment services—large companies out-tasking the desktop management to us. Professional services continues to grow along with the enterprise business, as we sell more [storage area networks], more servers, more clusters, [Microsoft] Exchange deployments, Oracle [Corp. software] integrations and those kinds of things.
This will be the first full financial year of our printer business, and well probably sell about a billion dollars worth of printers. In the U.S., we have about 14 or 15 percent market share.
How did you gain share so quickly?
Were providing pretty good value. Plus, theres a network effect. If desktop customers are satisfied, they may buy a printer; if notebook customers are satisfied, they may buy a printer.
I think we can change the game in terms of cost per page and the rate at which new technology is available to customers. Ultimately, this is a business that is ripe for change. Is change going to happen in one or two quarters? Probably not. But if you look at it over several years, youll see a lot of change occurring.
So, of your newer initiatives, printers stand out as the most successful?
Today, services is a bigger business and more profitable. But if you asked what am I most optimistic about five to 10 years from now, I would say printers.
What are Dells key areas for research and development?
Were spending it disproportionately on the newer areas, like enterprise servers, clustering, systems management, storage, printing and imaging. But were still spending quite a bit on desktops and notebooks, and were actually manufacturing those products, too, which is a little different from our competitors [laughter].
Are you targeting R&D at any particular technologies, such as I/O?
A big part of systems architecture in the server world is I/O and the bus speed. New bus architectures like PCI Express are increasing bandwidth. Thats creating more scalability and more redundancy, which, in turn, creates the need for systems management. Customers want to manage systems remotely, and they want it to be an automated process. They want virtualization.
The applications are moving in a kind of scale-out fashion. If you ask whats the vision for Dell, [at an event] with SAP [AG] last week we [talked about] this vision for the scalable enterprise. Perhaps its more pragmatic and less architecture, but we think this is something that we can deliver and [that can] make a real difference in the industry.
How would you define “the scalable enterprise” in 30 seconds?
Building blocks of industry-standard server architecture components that can be almost infinitely scalable to support the largest and most robust enterprise application needs.
In 20 years, Dell has gone from being a small customer of Microsoft and Intel to being probably their largest customer. As youve grown, how much influence have you gained in your ability to set direction?
We have folks inside Microsoft, and were working on things right now in our labs that are well beyond “Longhorn” [the next version of Windows]. We spend a lot of time with Intel and Microsoft on what the next generation should be. As they roll out new products, we have to be joined at the hip. How do you create a new product if you dont know the hardware it will run on? Thats an obvious one.
Architects on both sides are sitting down and thinking of what we need in terms of features and functionality that will meet the needs in 2007, 2008, 2009. I think we have pretty good influence.
Whats missing from Longhorn
What technology pieces are missing?
A lot of what were doing in the enterprise is taking things that are already done in more robust traditional systems and scaling them down into smaller systems—taking advantage of the enormous explosion in power and cost-effectiveness of industry-standard platforms. What youll see is that more and more of what was traditionally the minicomputer market will get taken into this scalable-enterprise architecture.
Theres so much going on in the hardware in terms of additional power, not just raw semiconductor power but multiple cores on the same piece of silicon, improving semiconductor geometries; much larger caches, which improve performance.
Again, a lot of that is taking what used to be the system architecture and scaling it down, improving I/O bandwidth and moving applications down to those platforms so they can take advantage of it.
Intel is coming out with the Nocona, the Xeon processor with 64-bit extensions. That would seem to be an area where you would have talked to Intel a lot.
Were you the spur that drove them to do that?
Thats kind of inside baseball [laughter]. The company that announced and delivered products that take that approach [Advanced Micro Devices Inc.] is not necessarily the first one that created it.
So Intel may have had the technology and sat on it, and then dusted it off?
Theres not really a whole lot to add there. Ill leave it at that [laughter].
Were customers telling you they wanted a transitional chip?
The largest single Itanium server customer in the world is a Dell customer. Were seeing some [Itanium] demand there for the upper end of the mainframe replacement spectrum. In the high-volume server market, its really about software. Ive got a Nocona box under my desk, but I dont have any software. Were going to be delivering a machine with 64-bit extensions way before theres a lot of software.
Putting Dell in the
On-demand computing picture”>
You hear IBM talk about on-demand computing or utility-based computing. How do you define those, and where do you put Dell in that nomenclature?
The marketing has gotten way ahead of the reality. When I go out and talk to customers, thats what they tell us. So our scalable enterprise is a more logical and understandable approach to whats really going on. The notion that the company you mentioned can help you save money doesnt make any sense at all. Look at their P&L [profit and loss]. Theres nothing about their P&L that suggests youre going to save money.
Scalable-enterprise computing could be inside your own business; it could be inside an outsourced provider or an application service provider. Were not trying to dictate to the customer whether they should do one or the other. Let me be clear here. Were not a consulting company. If you come to us and ask, “Where should my applications be?” were not making ourselves out to be experts on where your applications should be.
Dell introduced a Layer 3 networking switch recently. The game plan was always to keep moving up the IT stack. How far and how fast will you go?
Were pretty happy with the networking business. The units are growing at a rapid rate. Going up the stack does not have nearly the priority of some of the other things we talked about, like storage. Or services. Or printing.
One of the things we have to figure out at Dell is what were not going to do. Were not going to do optical switches and high-end consulting.
What about Tablet PCs? They seem to have attained a niche. Are they Dell caliber at this point?
Out of 170 million computers sold every year, the tablet market is—no one is really sure—300,000 units per year? Thats a couple of hours of sales for the industry. You can calculate it out yourself.
Are you seeing Windows XP upgrades driving a wave of demand?
Yeah, to some extent. You also have a lot of people out there with installed bases that are just pretty old, and they recognize the need to upgrade. When a new operating system comes out, you dont have a lot of big corporations jumping on it right away, which is why the software thats critical for 64-bit will take a long time.