Sam Palmisano has exerted a lot of energy in the last year to yank IBM toward a new computing model called "e-business on demand." It's what he and others within IBM believe will be the future of corporate computing: information resources combined with infrastructure that can be delivered much as the utility companies deliver gas, water, and electricity. When you need the extra capacity, you turn it on, and you only pay for what you use.

This concept of "utility computing" harkens back to the era when IBM was leasing mainframe capacity to all its corporate customers, prior to its anti-trust days. But what is most interesting is how IBM hopes on-demand computing will recast the role of computing within the corporate hierarchy. Why? Because on-demand computing is actually a tactic to turn the tide against the current commodity computing market trends in the industry.

How can it do this? And will on-demand computing help the iSeries gain new credibility?

The Commodity Marketplace

For more than 10 years, the computing industry has been surfing on a rising wave of commodity pricing: "Cheaper, Faster, Better" has been the mantra, and one need only look to the success of Intel and the PC to appreciate the implications.

As an analogy, consider that 40 years ago Americans scrimped and saved to purchase their first televisions. Today, the TV is simply an appliance, sitting in multiple corners of the house, constantly delivering entertainment and information, permanently on, invisible.

Computers in the corporation today are treated as appliances too. The cost of purchasing hardware and software has plummeted while the power of computing architectures has skyrocketed. The Internet has become the information "signal"--like the cable TV station--making nearly every information service readily available at an incredibly inexpensive price. This commodity technology trend has transformed almost every aspect of corporate information systems, making them so affordable that the cost of a computer's mechanisms is a comparative pittance today, while the real value has become nearly invisible.

Meanwhile, the manufacture and distribution of computing systems is driven almost entirely like any other commodity in the marketplace: Volume is high, and profit margins are low. Gone are the days when a single technological advancement in hardware or software can transform the industry. The "next big thing" is a myth of ages past, and many analysts believe that all that lies ahead is a wasteland of more commodity computer technology.

IBM vs. the Computing Appliance

For technology companies like IBM, the trend toward commodity or "appliance computing" has been a mixed blessing: On the one hand, it has increased their fortunes by broadening the marketplace; on the other, it has raised the cost of bringing successful products into the crowded market itself, while simultaneously lowering the perceived value of those products. But does it really matter to a behemoth like IBM? Absolutely!

It is this commodity pricing that drove IBM out of the disk drive business, causing the company to sell its holdings to Hitachi last year. (After all, what was the value of hawking disk drives when the price per gigabyte of storage was falling so rapidly?) It is commodity pricing that drove IBM to channel marketing of computing systems in the early 1990s. (Why staff a sales and support team with expensive personnel when brokers and Business Partners will do all the legwork for a fraction of the cost?) And it was commodity pricing that drove IBM to diversify into the services arena, creating massive computing centers and networked systems to meet the fluctuating requirements of corporate IT. (IBM reaped $10.6 billion in revenues in the second quarter of 2003 alone.)

Clearly, nearly every strategic move that corporate IBM has made in the last several years has been an attempt to escape the commodity marketplace, to transform its business model so that the value of its products might actually increase instead of decrease.

The iSeries in the Land of Plenty

In many respects, the fortunes of IBM's iSeries architecture have mirrored the struggles that IBM itself has suffered in the land of commodity computing. Its predecessor, the AS/400, tried to reinvent itself throughout the 1990s in a marketplace that portrayed its advanced architecture as an old fashioned, expensive, and benighted mini-mainframe. What the world wanted, analysts told IBM, was new, cheap, networked personal workstations.

But the true value of the AS/400--with its virtual architecture--was that it could adapt quickly. Meanwhile, the AS/400's primary competitors, HP and DEC, faltered or failed. Even during these last few years of tremendous competition and devaluation of computing hardware and software, the newly branded iSeries has remained a highly profitable--if not growing--computing brand for IBM. Yet the underlying question has remained: How can the iSeries--like IBM itself--escape the commodity marketplace? How can the "value statement" delivered by an architecture like the iSeries take control of a fragmented, balkanized industry of cheap solutions?

Redefining the Problem

In response to the market dilemma of commodity computing, last summer, on August 5, Sam Palmisano asked his strategists to design a revolutionary project plan that would unify all of IBM's hardware and software offerings, eliminating in-house competition between brands and charting an ongoing course into the 21st century. This project plan would--if successful--not only separate IBM from its competitors, but change the nature of computing itself. It would deliver the IT infrastructure not as a series of products, but "on demand," as required by the customer, as a combination of equipment, software, services, and features. Only IBM, as the world's largest technology company, Palmisano argued, could deliver such an architecture. It was to be called "e-Business on demand."

In response, IBM's strategists developed the current massive 10-year technological game plan. Step 1 requires the standardization of the customer's applications around open and flexible industry standards (WebSphere, Linux, etc.). Step 2 "virtualizes" network throughput, so that computational capacity on connected equipment is maximized. Step 3 creates computing grids that can act as the backbone of IT infrastructure, with devices delivered to customers that can be transparently connected to the grids. Step 4 creates a "utility-like" structure of services that enable the customer to buy capacity as it is used. Step 5 places IBM's business expertise into the hands of customers themselves, allowing IBM to sell its consulting services.

Almost simultaneously last spring, IBM announced its "capacity on demand" offering for the iSeries. Was this the same thing? What is the difference?

iSeries "Capacity on Demand"

As far back as 2001, IBM was marketing a license agreement to certain customers that allowed them to buy additional computing power on high-end iSeries machines at a slight premium. These customers would then pay for that capacity only as it was consumed during peak periods or if they chose feature upgrades. This successful plan was later re-branded as "capacity on demand" and bundled into a standard iSeries offering.

By comparison with IBM's larger goals of e-business on demand, the iSeries capacity on demand is a minor offering that is attempting to position itself with the larger corporate strategy. (Still, as a marketing tool for Business Partners today, it's of great value to larger companies that need that kind of growth management.)

And therein lies the rub! If IBM is serious about its e-business on demand, it's wasting precious branding capital on offerings that bear only a casual resemblance to the larger corporate strategy. CFOs and CEOs who become excited about the concept of e-business on demand may--in their rush to embrace the promise of this new on-demand technology--become jaded while waiting for the real benefits after purchasing an iSeries capacity on demand licensed machine.

The Face of Tomorrow: Will the iSeries Recognize It?

Meanwhile, should IBM be truly successful in uniting its offerings behind Palmisano's strategy, it also stands a chance that it will completely separate itself from the marketplace and its own current product line. This is exactly what happened when IBM began its ill-fated Systems Application Architecture (SAA) in the late 1980s: Customers got tired of waiting for IBM to deliver truly integrated offerings, while promised features of products like OS/2 failed to materialize. As a result, customers turned to other vendors, such as Microsoft and Novell. In other words, by trying to establish a new "standard" for computing, IBM abandoned the real-world needs of its customers in search of a holy grail so abstract that only a true techno-geek could appreciate it. And, quite frankly, it could do it again with its current plan.

Certainly, the iSeries is technically well-positioned for adaptation to an on-demand architecture. Or so its architects claim, with its virtual architecture and advanced features. But if IBM truly changes the face of computing, why would it bother to bring iSeries customers along? Indeed, a lot will change for mid-market iSeries customers in 10 years. But a revolution in computing is not necessarily one of those things that any iSeries customer really wants. And the idea of a complete virtual IT resource from IBM itself? How does that compute, when even ordering current iSeries products from IBM is a virtual nightmare?

Yes, IBM is truly gambling its future with e-business on demand, and the bottom line has yet to be reached. More importantly to iSeries professionals, the belief that the iSeries will be a viable component of that future architecture is not a foregone conclusion at all--despite the capacity on demand offerings that IBM is marketing today. And if IBM truly changes the face of computing, it may not be the face that any of us in this community will recognize.

Thomas M. Stockwell is Editor in Chief of MC Press, LC. He can be reached for comment at This email address is being protected from spambots. You need JavaScript enabled to view it..