The Four Hundred

IBM announced last week what many have long expected: that Sam Palmisano will be its next chairman. Palmisano will consolidate his control over IBM on January 1, when Louis Gerstner, who will have run IBM for almost a decade, steps down. Palmisano has been IBM's chief executive since March and its president since October 2000. He has run IBM's services, servers, and PC divisions, and is arguably the best person, other than Gerstner, to run Big Blue.

It was fairly obvious even in 1999 that Palmisano was in the running for at least one of the top jobs at IBM when Gerstner, who was set to retire in March 2001, left the company. (No one expects Gerstner, who is only 61, to play golf or fully retire just because he is retiring from IBM. He seems to have an interest in helping to transform the American educational system, and it would not be surprising to see him do something here.) Gerstner's work at IBM is done, and he has trained the next generation of managers to take over the company. We will now see what they will do with it, as well as to the IT industry, in which IBM plays such a dominant role.

Soon after Gerstner took the helm as chairman, chief executive, and president, in April 1993, he grumpily told Wall Street analysts and IBM watchers that the last thing the company needed at the time was a vision. We were just coming out of a recession (but didn't really know it yet), and IBM's mainframe business had taken it on the chin pretty hard, forcing the company to absorb an $8.1 billion loss, an embarrassment for the bluest of the blue chip companies on the New York Stock Exchange. The previous chairman, John Akers, had stirred up expectations that IBM would break itself into separate companies, to try to let these fiefdoms within IBM break free of the byzantine hierarchy of Big Blue, and perhaps fend better for themselves than they did inside IBM. But Gerstner quickly put an end to such talk, explaining what he thought IBM needed was to get its act together, not to breakup. My contention is that the IBM of 1993, in the sorry state it was in, would not have been worth much if it were broken into pieces. I also suspect that whatever Gerstner saw from inside IBM was probably a lot worse than any of us could imagine from the outside. In any event, only by keeping IBM together, pumping up its services and software message, buying back tens of billions in shares to keep net earnings growing, and peddling as much of its iron as it could in a recovering economy, could the company get itself back on an even keel. Gerstner got lucky with the dot-com boom (as did all IT suppliers) and was smart to capitalize on the e-business slogan his team coined (and turned into coin, lots of coin).

It's a tough act to follow, and Palmisano probably will not have serendipitous history on his side like Gerstner enjoyed--coming into IBM in the wake of a recession just before the biggest IT boom since the initial big bang for the widespread commercialization of electronic computers in the 1960s. Like Gerstner, Palmisano is taking the helm of IBM in the wake of a recession (at least we hope this recession has ended, or soon will). The company has a sense of its place in the IT market, and no one is worried about whether IBM is going to live or die. The odds do not favor anything like the e-business boom of the late 1990s to happen again in the 2000s--or, indeed, ever again in the IT market. And that is why, if there is one thing Palmisano needs for himself, for IBM's employees, for IBM's investors, and for IBM's customers, it is a vision. He presented the third iteration of that vision--something he has been building on since he became IBM president two years ago--at a gathering of IBM's top customers and IT analysts in New York last week.

In late 2000 and early last year, the message from Palmisano was that customers had heterogeneous environments and that was the way they were going to continue to build IT infrastructure. No one platform wins, so IBM has to support any and all platforms, including Linux. Late in 2001, Palmisano and his team started talking about the various stages of e-business adoption. IBM and consultants McKenna and Company surveyed a bunch of customers by the end of 2001. About half of the companies surveyed were in the basic e-business phase, while another 30 percent of companies were trying to integrate with partners and customers via Web-based ERP, CRM, and SCM applications. Only 1 to 2 percent were entering the "advanced e-business" stage. This is the final stage of e-business development in Palmisano's vision, and it represents the point at which a company is fully Web-enabled and can interact in real time with employees, customers, partners, and suppliers. (The other 18 percent of companies surveyed were, presumably, thinking about doing something about e-business, even if they hadn't started moving yet.)

With the vision that Palmisano presented last week, he changed the name of "advanced e-business" to "on-demand business." Apparently, the phrase "e-business" has lost its luster in the market, being so closely associated with the dot-com boom, so IBM is dropping it. And now IBM's ideas associated with its self-healing, self-administering technologies--collectively called Project eLiza--are being woven into this advanced e-business stage, and now the ultimate goal of e-business--and the star that Palmisano will steer IBM by--is on-demand computing, which implies dynamic, autonomic systems.

This is right out of science fiction, of course, and represents the dreams and desires of 30 years of systems designers and operating systems writers. The big IT vision of the 2000s is not a computer in every household, nor cheap processing capacity so more human tasks can be automated. The vision is to create machines that don't need people to operate them at all--or, at least, very little. This is an important statement to make, to be sure. But it will be quite a different thing to deliver the products, and the vagueness of IBM's promises in autonomic computing and utility computing (which is what IBM is really after) is such that it will be hard to tell if the company has reached any particular milestone. For example, IBM might have promised to double the scalability of its server line every two years or so and deliver a 30 percent price/performance improvement in a server line each year. We knew if IBM did or didn't do this. We knew from history whether or not IBM had done it before, and could more or less calculate the odds for the future, based on IBM's skills and those of its competitors. I suppose IBM's vision of autonomic, on-demand computing will be like Supreme Court Justice Potter Stewart's famous statement about pornography in 1964: "I know it when I see it." When we are no longer concerned with buying software and servers to support it, but only the processing capacity that we need to run our applications, then we will have finally arrived at Palmisano's elusive on-demand computing goal. And by then, maybe IBM will no longer be a hardware company, but merely a grid supplier, like so many electric companies that have abandoned the difficult and expensive work of building efficient, clean power plants to third parties who neglect this responsibility so they can bill customers for the use of their grids.

I don't know about you, but I dislike this vision of the future. I like the idea of regional power plants, of competition among suppliers, and of smaller, interconnected grids for both computing and electricity. I think companies like to have their own data, their own applications, and their own servers. I know I do. Because of monolithic suppliers, which often have needs distinct from ours, we often can't adapt to new technologies quickly. A big supplier moves to new technologies when it suits its short-term profit and loss statements, which usually means never. Computers, by their very nature, are more flexible than coal-fired electric plants, so, hopefully, in this future of on-demand computing, we won't get stuck in old technologies. I think that the competition in the server market has allowed companies to do the things they need to do with their machines, even if they could not accomplish everything they dreamed of.

There's nothing wrong with autonomic, on-demand, utility computing, unless you happen to be a system administrator (who will be out of a job) or a server channel partner or sales rep (who will also be given the boot). But IBM likes the old mainframe days, when it rented machines and could control the release of technologies at a pace that it--not its customers--set, and this vision of computing moves IBM back in that direction. IBM wants a predictable revenue stream, like it gets from mainframe software licenses. The problem is that no one wants to buy mainframes any more. This utility concept is the best thing that IBM--and, indeed, any IT supplier--can think of to change the subject from pushing product to supplying services. Hewlett-Packard, Microsoft, Sun Microsystems, and other dominant players in the IT market are all trying to think of a different set of marketing buzzwords and technologies to get to the same goal as IBM. And if they get there, we won't need MIS managers any more, and we won't need IT newsletters, either. I keep remembering, every time I hear about utility computing, that there used to be guys called chief electrification officers when American manufacturing switched from steam-and-pulley power to electricity and motors in the middle of the Industrial Revolution.

But we are a long way off before IBM or any other IT supplier gets us there, and, in fact, if anything, it will take a collection of suppliers and technologies to accomplish IBM's goal.

To that end, Palmisano said that IBM had dedicated $10 billion to this autonomic computing vision--presumably spent over quite a few years. "My executives buy into this," he said at the conference in New York last week. "They get it, and they are going to drive this. I'm lucky." Later in his speech, he qualified that $10 billion bet as he spoke to the several hundred chief information officers in the room. "It's a big bet. It's a bold bet--no doubt about it. But it is not a risky bet, because it is tied to your priorities, not our priorities. And if we do that, we can't lose."

We'll see. And we'll see what cultural tectonic shifts take place in the IT industry if Palmisano is right and IBM can deliver on true utility computing and autonomic computing. It was one thing to use mainframes to get rid of paper pushers and paper filers in corporate back offices in the 1960s and 1970s. It was one thing to get rid of secretaries and assistants with the advent of the PC and word processor in the 1980s. But IBM is now talking to IT organizations about getting rid of IT organizations. Not everyone, particularly those who love their IT empires and fiefdoms, is going to like this future IBM is dreaming up. But you can bet that IBM has the CEOs and CFOs of the world listening, and that it will leverage its influence with them to get what it wants, even if IT people protest that IBM wants, in the long run, to make them redundant with a computer that no longer needs people to run it.