YouTube Facebook LinkedIn Google+ Twitter Xingrss  


Bio·IT World Executive Editor—IT John Dodge spoke with Dell about his company's first major foray into the life sciences at the University of Buffalo and his plan to grow Dell's professional services.
Oct. 9, 2002 | Coming off a gangbuster third quarter that yielded a $50-million profit versus a $101-million loss a year ago, Dell Computer Corp. is finally starting to talk publicly about the life sciences. CEO Michael Dell says the company has been selling to drug discovery companies and research centers all along. The news is that latecomer Dell is starting to focus on several vertical markets, including the life sciences. Until now, Dell classified companies by size and geography, but not by industry. Bio·IT World 's John Dodge caught up with Dell at the State University of New York's University of Buffalo (UB), where along with college officials, he announced the installation of a huge Linux cluster consisting of 2,000 Dell dual-processor servers to support the work of Jeffrey Skolnick, a renowned computational biologist. Skolnick, whose appointment as executive director of the Buffalo Center for Bioinformatics Excellence was effective Sept. 30, was recruited by New York Gov. George Pataki, who spent $1.9 million to lure him and two fellow researchers from the Donald Danforth Plant Science Center in St. Louis, Dell talks about the UB project and what to expect next from this company that has generally been perceived as consumer-based, rather than business-oriented.

Q: What is Dell's life science strategy? 
A: Bioinformatics is embracing industry standard computing so our products have broad applicability in life sciences, whether it's in commercial or university research. Dr. Skolnick [is] the expert in life sciences, and we're the expert at providing industry-standard technology. He does not look to us to do his research. He looks to us for cutting-edge technology around clustering and compute engines and to have the infrastructure to deliver that.

Q: So what you're saying is the "volume economics" that has always driven your business won't fundamentally change? 
A: The fundamental change is that now [research] is done on industry standard platforms, whereas before, it was done on proprietary platforms. That means the cost of doing this research comes down dramatically. It means clusters can be installed in a lot more places much more quickly. Instead of a year, it's sixty days. Instead of huge amounts of money, it's a lot less.

Q: Would you say the life sciences are coming to you instead of the other way around?
A: I don't think it's quite coming to us. We certainly had to learn about each of these industries whether geophysical simulations or media content creation (two other verticals Dell is targeting).

Q: The IBM Life Sciences business unit has 250 people, many of them Ph.D.s. It puts drug discovery algorithms in the public domain. In other words, it spends a lot of time and money trying to understand drug discovery. Could you see Dell having similar vertically oriented experts? 
A: We've had a team of Ph.D.s focused on clustering and high-performance computing for years.

Q: But you mean computer scientists and electronic engineers, not biologists or chemists, don't you?
A: They're more focused on the horizontal application of our tools. The customer has a lot of that knowledge. Dr. Skolnick did not come to us saying, 'I need help with my algorithms.' He said, 'I need computing power.'

Q: Would Dell go into big pharmaceutical companies like Novartis or Merck with biology Ph.D.s like IBM to help bridge that gap between the computational and the science sides of drug discovery?
A: Our Dell Plural group (Dell recently bought Plural Professional Services, a New York e-consulting firm using Microsoft Windows) has done quite a lot of work in knowledge management in the pharmaceutical field, where you have 20 different facilities and hundreds or thousands of researchers working on the same thing at any given time. That's an example where [we] get more vertical. Our professional services business did 2,000 engagements last year. And those are getting more and more industry-specific. But I still contend that the overwhelming force here is industry standards, and that the costs of proprietary technology are prohibitive.

Q: Plural was the first of several targeted acquisitions Dell has said it will make. Could you see yourself doing something in the life sciences? And will these acquisitions be vertically oriented or, as tradition dictates with Dell, more horizontal in nature? 
A: We're looking at growing professional services considerably. We wouldn't rule out anything in particular.

Q: Would you do something on the scale of IBM's acquisition of PricewaterhouseCoopers' PwC Consulting? 
A: We've been focused on much smaller organizations that have a fit with the products we are selling today and the kind of services our customers are looking for. We're also organically growing that. Our services grew 200 percent even if you take Plural out (not including support services, which makes up the majority of Dell's 7,000-strong services group).

Q: How many people in professional services? 
A: It's hundreds.

Q: What will it be in a year? 
A: It'll be more. It's growing very rapidly. We want to add to it. Go back and look at clustering, and we installed this thing in around 60 days.

Q: Dell has only three rankings in the ranking of supercomputers. Hewlett-Packard leads with 168, and IBM is close behind at 164 (as of June). Where will Dell be in a year? 
A: The big thing there is getting them registered. [HP and IBM] have gone to great lengths to get them registered. To get on the list you have go through an application process, and it requires a fairly involved effort on the part of the customer. We've done a number of installations where the customer has no interest in that at all. I don't really have a great answer other than that. Dr. Skolnick told me this is the third most powerful computer system in the world. My mom's proud!

Q: IBM mainframes have shared memory, tighter integration than the cluster, more remote capabilities, self-diagnostics and extremely fast compilers. When will the clustered world come up to that level? 
A: There are a lot of people working on tools for that, such as Platform Computing and others. As the architecture evolves, microprocessors are more aware of [high-performance computing] as a principal application. So they're thinking about how to get massively scalable architectures. Look out three to five years from now.

Q: If clustering with industry standard computers is quicker and simpler, why do you have to grow your professional services business? There once was a school of thought that fewer support people and consultants would be necessary. 
A: What we are doing is much different than IBM and EDS [Electronic Data Systems]. There is certainly a group of applications or customers where there is a great desire to have those services that go along with the products. When you have something as complicated as [the UB cluster], it's a lot of work. We have 8-node and 64-node off-the-shelf stuff that's pre-done in the factory. (The UB cluster server racks were assembled in Long Island and shipped to Buffalo.)

Q: From the time you started assembling them on Long Island to the time they were here up and running, was it still less than 60 days? 
A: Yes! (UB President William R. Greiner said not even a single order had been placed 90 days prior to the Sept. 3 press conference.)

Q: What's the promotional value of doing something like this, standing up with the mayor of Buffalo, the State University of New York chancellor, and publicly talking up one of the world's largest clusters? 
A: You tell me ... (smiling).

Q: Does this change the perception that Dell is more than notebooks and desktop PCs? 
A: The thing I forgot to say [at the press conference] was that we sell more servers than anybody else in the U.S., according to IDC. We've been at it for seven years since the PowerEdge line was introduced. We first went into the server business in 1989, but we weren't as successful then as when we came out with PowerEdge. Sometimes when people think of Dell, they think of Steve [in the "Dude, you're getting a Dell" ads]. It turns out about 85 percent of our business are businesses, institutions, and government. The vast majority are not consumers. People don't always see that, so [the UB cluster] is a nice contrast to that.

The UB cluster also emphasizes an area that's important and growing. And we think there's a continuation on the theme of democratization of IT that started with the PC. It's the idea that someone can put an 8-node cluster in their office and do research all by themselves. Think ahead three and five years and you've got individual computers that are super powerful.

Q: Do you have other deals on this scale cooking? 
A: We have a couple and some that we've already done that are pretty significant that would qualify.

Q: How much more important have universities and research centers become based on new techniques for early stage drug discovery? 
A: What's happening is the universities play a critical role in the early role of [drug discovery]. The cost of doing this has gone way, way down. What's exciting is you are going to see lots of [clusters] pop up.

Q: You just came off a great quarter with 25 to 30 percent earnings growth forecast for the next quarter. How much of that came from taking market share and how much came from industry-wide growth? 
A: The market didn't really grow, so we grew market share.

Q: You have deals with companies such as EMC. Does its historically low stock price present a buying opportunity? Would you consider a deal that big? 
A: We don't comment on speculation like that. As far as EMC goes, we have a great five-year partnership with it. There is a 16-terabyte (10 useable) SAN in this cluster from EMC. We're shipping about 250TB of storage a day now and the Dell/EMC component of that is the fastest-growing. We have a very successful partnership with EMC, and we're expanding it with new products like the 600 Series [Dell/EMC CX600 enterprise storage platform] that help us drive cost down.

Q: Who is your biggest competitor in high-performance computing? 
A: The ones you mention — IBM and HP.

Q: Did HP and Compaq lose momentum during the pre-merger fight and the post-merger integration? 
A: I'm not sure when all that ends, but I'm enjoying it so far.

For reprints and/or copyright permission, please contact  Jay Mulhern, (781) 972-1359,