The Midrange Manager: Do You Have Java on Your Machines?

  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times
In a survey earlier this year, I polled the readers of this newsletter, asking them who they felt had spent more on Java: IBM, Sun, Microsoft, Apple, or other. Most people felt Sun had spent the most on Java. But IBM has actually spent the most--well over $1 billion over the last few years.

So what do have they to show for it? Has it produced any revenue or that next killer application? I wish I could compare that $1 billion to what they invested in selling the iSeries during the same period; I'm certain the return on investment would be better.

But today another "Java" technology has been sold to the IBM high command: On Demand. This is the delivery of IT resources as needed, something like having the electric company turn up the power output during the peak summer usage periods. The first delivery of this technology was in the iSeries. Multiple processor cards can be installed in machines so that when customers need to upgrade, they simply call IBM and pay for the extra CPU. IBM sends them a code, and the extra power is released. IBM is also doing this for Web "services," except it looks a lot like a new name for outsourcing your IT. Here's a quote from the IBM e-business On Demand Web page:

"By giving up the burden of IT ownership, your company can vastly increase its access to computing power, expertise, and innovation. Subscribe to what you need today. Turn it up when you need more."

Pretty cool idea, isn't it? To me, this says, get rid of all your IT staff and use the big computer in the sky, and oh by the way, we at IBM offer this to you.

Like other consultants, I make my living off of companies owning their own IT assets; so do most software vendors. The only people who will be making a living if On Demand takes over the world are Microsoft (on the desktop) and the owner of the Big Iron in the Sky. That could mean the lowest bidder, which probably wouldn't be IBM. But don't worry; IBM's marketing may screw this up, too, so there's always hope.

But realistically, who is this technology for? Sure, several large-scale customers such as the Fortune 10 will have a need for this kind of thing, but who else will? In order to price it profitably, IBM would need to price it in the mainframe computing model--which is 5 to 10 times the price that a typical iSeries shop is used to. If it is less, then those top 10 companies may not be spending as much money with IBM as they are today, and I don't think reducing annual sales figures is what IBM had in mind.

The interesting thing to me is that somebody actually believes this will help IBM's bottom line. Let's look back at Capacity On Demand. If you expect to need a 25 - 400% increase in CPU capacity within a year or two, Capacity On Demand for iSeries seems like a good idea, but at what cost? I know that when I buy new equipment, I don't want two-year-old technology. Suppose I buy a four-way iSeries and only need a two-way today but expect to need the power of a four-way in two years; will I really want to activate two-year-old technology? If I could get a price that represents two-year-old technology, then perhaps I would. But I doubt the prices will reflect that kind of depreciated value. After all, IBM essentially installed double the CPU chips that you originally paid for. So their cost related to those chips is doubled. Sure, that's not a huge sum relative to the cost of the entire system, but if you were IBM and your earnings weren't as high as you'd like them to be, would you want to continue manufacturing something that has increasing costs but stable or falling retail prices?

Now, IBM is talking about adding DASD to the Capacity On Demand. Essentially, they'd be shipping a full system to everyone--max DASD, max memory, max CPUs. Need to upgrade? Just call IBM and--voila!--you've got it. Great idea if this were 1988, but today, if you need capacity, you call DELL or HP, order another server, plug it into the network, and you're done. Granted, networks are unreliable and complex and take an army to manage, but you still get a new system installed into your infrastructure, and that's what people are becoming used to: overnight ordering and user-installable components.

We have people in this market (the midrange market space of System/36/38 AS/400 and iSeries) who haven't shut their machines off for years. Those machines just keep on running, and performance is not an issue. Can you imagine new customers who install a small iSeries and decide five years later to upgrade? What do they do with all this Capacity On Demand available? What if IBM no longer supports the chip set? Will IBM give them new boxes with new hardware for a price similar to the Capacity On Demand prices? Interesting question, I think.

For new, start-up, or growing companies with lots of available cash (for example, an old dot-com that hasn't gone under yet), On Demand may provide the kind of infrastructure needed to get going faster, and that's a good thing. But in my view, On Demand and Java are very cool and interesting technologies that "techies" within IBM may have simply sold to upper management as a cool new technology. But were they simply solutions looking for problems to solve, problems that may not actually exist?