Server Ecosystems: Take a Ride on a Slide

by Timothy Prickett Morgan

I was asked recently by Ben Reitzes, the information technology analyst at brokerage house UBS, to host a teleconference call with institutional investors in the high tech market about the server business. I was terrified at the prospect of public speaking, but I am also adventurous enough to want to conquer a fear and learn a new (and perhaps marketable but definitely useful) skill. So when Ben asked me to do it, I said yes. That presentation is where many of my ideas about server platform ecosystems started to crystallize.

Because I am a writer and analyst who probably writes too much for his own good and maybe for the good of tired readers, and because I work from a home office with kids running around all over the place so my wife can go out and work long hours to bring in big game as a litigator, everything in my life is done at the last minute, including preparing for that speech that I was afraid of. I threw together some quick numbers, tossed in a few jokes, and because I didn't have time to really prepare because it was the Memorial Day holiday weekend (that was my official excuse, anyway), I got on the call and wung it on Tuesday morning after the holiday. (Wung is the unofficial past tense of wing in my book, er, newsletter.) So my first presentation as an authority in the server racket was extemporaneous.

You can get the presentation I did, Ben's research note on it, and the replay information if you want to actually hear what I sound like when I am scared by clicking here. I apparently did OK, with over 100 people on the call. I talked for 25 minutes in my presentation--I have only the barest recollection of what I actually said--and then answered questions for another 40 minutes or so. Apparently he usually gets around 50 institutional investors on the call and maybe the call lasts 40 minutes. "TPM, your stock value just went way up," Ben told me. It was a rush to survive the experience.

I want to share some quick observations from this experience, and then I will wrap up this series on the OS/400 ecosystem and how it relates to other ecosystems.

First, I decided to take a tongue-in-cheek approach for the UBS presentation, and then nearly swallowed my tongue on the first sentence. The presentation was entitled "Which Server Platforms Will Win?" and to be clever, the second page of my presentation said this:

I thought that was funny. But I hadn't thought ahead enough to realize I would be talking to an audience with dead microphones. Yikes! When you tell a joke and no one laughs, you feel like a comedian that is bombing. And that is a pretty tough moment for the first seconds of a first speech. My adrenal glands all kicked in, and I still have no idea what I was babbling for many minutes.

A person who was afraid of looking like a fool would not even bring any of this up. But it brings me to my second point: It is really difficult to think on your feet like that. Having run our little IT Jungle publishing business for nearly four years now, I have learned a lot of lessons about sales, budgeting, making a payroll, and all of the other things that every company faces--including managing information technology. My compassion for business managers and hard-working employees has only grown with time because of this experience. I know I sit here in my white tower--well, really, the home office in my cooperative apartment in Upstate Manhattan--sometimes praising and often criticizing what others are doing in the IT business. That may be my job, but I have respect for the people who can walk the walk and talk the talk. It is not easy to get out there and promote any idea to a group of strangers. Making that presentation made that resoundingly clear to me in a different way. Text is easy for me, after so many years of doing it. But speech--at least for now--is difficult. I respect people who have this skill, and that includes some of the people in the IT industry that I otherwise loathe.

Now, let's get to work. While I was joking in the slide above that if you look at the server shipment data, it is easy to conclude that the server business is an open and shut case, this is not true. Look at this pie chart as the first picture in the presentation and you might be wondering why I bothered:

As you can see from this pie, in 2004, a stunning 92 percent of the servers shipped were based on the X86-X64 architecture, according to Gartner. Servers based on the IBM Power RISC architecture--the pSeries and the iSeries--accounted for a mere 2 percent of the 6.7 million machines sold, and the iSeries was probably only about 20 percent of the 133,000 RISC-based servers IBM sold. When you do the math, 20 percent of 2 percent is not a big number. It kind of takes your breath away, in fact. And that is why I spent three weeks working backwards to this point. Because as the next pie chart shows, shipments ain't everything. Money is a pretty big deal. Server vendors, it turns out, like it as much as we do.

When you adjust for the money, suddenly the blue section (which is dominated by mainframes and other funky machines) and the various RISC architectures are not doing quite so badly in the $49.5 billion server market according to Gartner's data. In fact, IBM and Sun Microsystems are seeing shipment growth in their RISC server lines, and the only reason they are seeing revenue declines is the cut-throat competition among the Unix vendors as they simultaneously compete against X86-X64 iron running Windows and Linux. Take a look at the table below:

As you can see from the table above, it has not been an easy three years for the server business between 2002 and 2004. In late 2001, the server business was on the rocks, and then 9/11 hit, the worldwide economies got a jolt, and it just plain got bad. Dire economic situations always cause dramatic shifts in computing architectures, and to illustrate my point, I drew this chart below:

That slide pretty much encapsulates my entire career in the IT industry, and I will remind you that I was for many years the contributing editor to a mainframe newsletter in addition to my work on The Four Hundred for all of these years; that mainframe letter has now been out of print for many years, and I lost my job in 1996 as I rode that slide down.

Click to continue reading the next page of this story >>>