Nov. 4, 2002: In this week's commentary: As long as IBM perceives on-demand computing as something that its customers get only from IBM, its products and services will fall short of its vision.
Seven years ago this month, Lou Gerstner, then CEO and chairman of IBM, stood up to give the opening keynote at Comdex, the huge annual tradeshow of the U.S. computer industry. It had been an exciting year: Sun had introduced the platform-independent Java programming language, Netscape had completed its record-breaking $2.6 billion Nasdaq IPO, and everyone was fired up about the potential of the Internet and the WorldWide Web.
Gerstner's speech brought the full weight of IBM behind a concept that had already been floated by the CEOs of lesser companies including Sun, Oracle and Netscape. The industry, he said, was about to move beyond client-server and enter a phase of "network-centric computing," driven by powerful networking technologies. "In a truly networked world, we can share computational power, combine it and leverage it," he said. "This world will reshape our notions of computing and, in particular, our notions of the PC."
Read and React
"It's hardly surprising that the company that once dominated the market for "big iron" mainframes should still want to be the main provider of its customer's computing power. "
Last week, Gerstner's successor, Sam Palmisano, stood before three hundred of IBM's top customers and, in a speech that was simultaneously broadcast to the corporation's staff worldwide, outlined his own vision of a new phase of "on-demand computing" (see Palmisano: On-Demand is Here, It's Ready).
We've Come a Long Way, Big Blue
The contrast between the two speeches shows how far we have moved on in the intervening years. In place of Gerstner's vague 1995 notion of shared computational power, Palmisano in 2002 can point to live implementations of grid computing. In place of Gerstner calling on the industry to embrace open standards, Palmisano can cite established Web services protocols such as XML, SOAP, and WSDL, as well as IBM's championing of the Linux operating platform.
Instead of Gerstner promoting more powerful networks to transfer knowledge between incompatible computer systems, Palmisano envisions a shared, networked systems infrastructure that enables real-time process integration across disparate applications.
All of us have shared that seven-year journey. Gerstner's 1995 Comdex keynote effectively summed up the aspirations of the industry at the time. It seemed to validate many of the innovations that had been forming in people's minds in that momentous year, in particular those of many ASP pioneers.
Unfortunately, there were many false assumptions behind those 1995 ideas, in particular the notion that networking meant sucking everything up into the center. "If the communications link between the PC and the network is cheap enough, fast enough and has virtually unlimited bandwidth," Gerstner suggested, "why not migrate a lot of the functions that currently reside inside the PC to the network — the applications, the data, the storage, and even some of the processing?"
The Power of the Grid
Today we realize that networking is a two-way street. It allows us to migrate the network right inside the PC, supplementing its local resources rather than sucking everything out to leave just the hollow shell of a network appliance in its place. There is no such thing as a network center, where all its intelligence and power resides. The network is a grid that every resource contributes to and participates in.
IBM's shift of emphasis from "network-centric" in 1995 to "on-demand" in 2002 reflects that growing realization. But being in tune with current thinking still doesn't mean that IBM has all the solutions. In 1995, Gerstner cited just three products as examples of how IBM expected to fulfill his vision of network-centric computing. They were ATM networking, the planned network appliance and Lotus Notes — none of which turned out to be pivotal to the execution of the vision. Conceived at a time when client-server was still the dominant computing model, they were imbued with a flawed view of networked computing that saw just two tiers, the edge and the center.
IBM Stuck on Being the Center
Today, IBM has moved on to adopt a vision of "on-demand computing," but it is a vision that is still imbued with network-centric values. The true meaning of on-demand is a peer-to-peer model in which resources can be shared from anywhere on the network, and which has no definable center. As long as IBM perceives on-demand computing as something that its customers get from IBM, it will be imagining itself at the center of their network, and thus inevitably its products and services will fall short of its vision.
It's hardly surprising that the company that once dominated the market for "big iron" mainframes should still want to be the main provider of its customer's computing power. But enterprises that want to exploit the full potential of on-demand computing should remain wary of locking themselves into Big Blue's Big Grid.
Do you have a comment or question about this article or the ASP industry in general? Speak out in the ASP Discussion Forum.
Phil Wainewright founded ASPnews.com in 1998 and is the publisher of Loosely Coupled. He can be contacted at
IE 7
Firefox 2.0