(IDG) -- IBM, the one-time manufacturer of time clocks and electronic typewriters, is banking its future fortunes on a relatively new software start-up called (drum roll) -- IBM.

How can a company be an 87-year old, $79 billion manufacturing behemoth, and a software start-up at the same time? By carving out a tangential business, turning it into a $13 billion operation, and using it to drive the entire company's vision for computing in the Internet age.

During the past three years, IBM has reinvented its approach to developing and marketing its software from top to bottom. Not just that: Like a brash software start-up with little to lose, the company has boldly lead the industry in attempting to establish open and democratic standards, such as Java, despite its gold mine of existing proprietary products.

This is a radical departure from the past 40 years in which IBM's software was developed on isolated islands and slavishly tied to its respective hardware platforms. Software was merely something sold to users along with its real product -- hardware.

Since early 1995, IBM has gathered up these lost software tribes and given them an organizational home. It is part of the company's attempt to create a cohesive but free-thinking software company with legs fast enough to keep up with competition in the digital economy of the 21st century.

Granted, IBM had to do something. Its hand was forced by Microsoft, which is reaching for IBM's core customers, and by a threatening wave of Internet-based companies rolling its way.

And whether IBM's software legs, supporting the weight of the industry's largest installed base, can run as fast as its heart desires remains to be seen.

MORE COMPUTING INTELLIGENCE

IDG.net home page

InfoWorld home page

InfoWorld forums home page

InfoWorld Internet commerce section

Get Media Grok and The Industry Standard Intelligencer delivered for free

Reviews & in-depth info at IDG.net

IDG.net's personal news page

Subscribe to IDG.net's free daily newsletter for IT leaders

Questions about computers? Let IDG.net's editors help you

Search IDG.net in 12 languages

News Radio

Fusion audio primers

Computerworld Minute

Ultimately, the success of the high-tech world's oldest start-up depends, not just on the timely delivery of a compelling set of software products, but on its unproven capability to market a unique vision. In this case, the vision is e-business.

"Our No. 1 integration initiative is e-business. It is how we can take the lead in helping all institutions of society do what they do better through the universal connectivity of the Internet," says Irving Wladawsky-Berger, general manager of IBM's Internet division, in Somers, N.Y.

In more finite terms, IBM's e-business strategy is to get customers to use its middleware software -- including DB2, CICS, and MQSeries -- with its Web-oriented midrange servers, such as WebSphere, to process on the Internet much more than the 20 billion transactions per day that IBM hosts already handle around the world.

Industry observers give IBM, specifically its steely President and Chairman Louis V. Gerstner, due credit for articulating the e-business idea. That he promotes such a concept at all is a big step; five years ago Gerstner said vision was secondary to fixing nuts-and-bolts problems. Today, this vision has helped distinguish IBM's vision from that of brash Microsoft Chairman Bill Gates.

But many of those same observers who credit Gerstner have considerable doubt as to whether IBM can materialize a vision as solid as the company's logo itself.

Theoretically, this should not be impossible, especially for a "start-up" that is second in revenues only to Microsoft's $14 billion, and that spends approximately $3 billion per year on software research and development. But many state that Big Blue has a long and dusty march ahead before its e-commerce vision shines with the strength of software start-ups that are younger in reality, not just in spirit.

Vision marred by cataracts

Among IBM's challenges is tying its e-business software vision to products that appeal to more than just the existing customer base, critics contend.

"When IBM came out with its e-business thrust I was optimistic," says Jeff Tarter, editor of The Soft Letter, a newsletter based in Watertown, Mass. "I thought they had come onto something brilliant. But so far that vision has little meaning to anyone except that IBM will sell its existing customers some stuff that works with e-commerce."

The industry's largest IT customers, too, are at best cautiously hopeful that IBM can both walk the walk and talk the talk.

"I think Gerstner deserves credit for crystallizing a direction for a company of IBM's size to go in, but I can't build a career on a road map and a slogan," says one high-level IT executive at an East Coast Fortune 10 company. "They have to deliver all the pieces that add up to some clear value-add for online enterprises. They have those pieces on paper, but it is their game to lose."

IBM executives respond to such skeptics with the confidence of hang-loose, Silicon Valley-based 20-somethings, particularly when speaking of their nemesis Microsoft.

"What we sell is less glitzy than a Microsoft," says Mark Elliot, general manager of sales and marketing at IBM's Software division, in Somers. "But we make the world's best pipes and internal plumbing, and at the end of the day it is the plumbing you can't see that makes all the difference to running an online enterprise. What Microsoft sells by comparison is fixtures."

Although bad-mouthing Microsoft's glitz, IBM hopes to emulate it. It is marketing -- that art of getting someone to buy something they don't think they necessarily need -- that IBM software has been missing all along.

Now, "the challenge is taking the IBM logo with its 80-year-old brand image baked into it and attach it to a thriving software image," Elliot says.

Elliot should know. It was Elliot and Steve Mills, a veteran at IBM and now general manager of IBM's Software Solutions group, in Somers, who first proposed to Gerstner about four years ago that IBM build a truly independent software company. But in the months just prior to that proposal, the two issued an industry survey to find out what sort of software image IBM had in the market. The reflection was like that of a ghost looking in the mirror.

"The survey came back and told us we didn't have an image," Elliot says. "In the 1993 time frame we were generating about $8 billion in software and still no one thought of IBM as a software company."

It's not easy being middleware

Exacerbating IBM's image problem is the fact that its e-business vision revolves around middleware -- those gray and unglorious undergarment products.

Pitching its middleware products at hip new companies such as Amazon.com and E*Trade that make up today's most successful e-commerce sites is not IBM's strong suit.

But IBM expects this to change as more e-commerce sites experience high transaction volumes that require scalability, availability, and security -- IBM's traditional areas of strength. This high volume will be encouraged by performance improvements to the Internet of anywhere from 100 to 1,000 times during the next three to five years. A lot of that improvement will come with Internet2, a joint project involving the academic community, the U.S. government, and selected industry companies, most notably IBM.

And IBM already is making some inroads. The founders of SpeedServe.com, which sells books, videos, and games on the Internet, picked IBM's mainframe platform because it wanted a platform that could scale quickly and give the company continuous availability and security.

Meanwhile, the task of trying to market this unglorious middleware software against the formidable marketing machine Microsoft and its Windows NT, doesn't appear to faze IBM executives.

"The head-scratcher with Microsoft's [middleware] strategy is that it is all conceptualized around an NT-centric thought," Mills says." I think the more commonly accepted notion of middleware out there is something that knits together a wide range of different platforms."

The Microsoft dilemma

Another challenge is cooperating with its main competitor -- Microsoft. The market dictates that IBM software work well on its competitors' platforms, but the Internet-based integration technologies that IBM has chosen are at odds with those that Microsoft has picked.

Simultaneous cooperation and competition with Microsoft has been at the heart of IBM's software strategy since its new beginning. After Gerstner granted approval for a new, independent software business, the organization's first decision -- described by insiders as "gut wrenching" -- was to put its key software products, most notably CICS, MQSeries and DB2, on its competitors' platforms, specifically Microsoft's NT.

"My strategy embraces Microsoft and includes them in everything I do," says John M. Thompson, senior vice president and group executive at IBM's Software group, in Somers.

This decision to push software for mixed environments in favor of its proprietary ones signaled a new maturity within IBM. It also put an end to the cold war during the early 1990s that began after Microsoft abandoned joint development efforts for graphical operating systems. That war badly scarred IBM, tore at the loyalties of its corporate users, and slowed down customers' buying decisions. Now customers are happy to see the evidence of IBM's more neutral attitude.

"They are pretty willing now to help us integrate some interesting non-IBM products," says John Henderson, a buyer at a large Indianapolis-based medical products distributor.

In fact, IBM today has more Windows-based products in its lineup than any other competitor, including Microsoft. And although many of IBM's efforts in corporate markets, such as e-commerce, will center around the mainframe version of DB2, it is the NT version of DB2 that is clearly driving its database business in small and medium-size companies.

"We have over 250 people working to either recruit or enable applications for the DB2 platform. Much of that work -- maybe 80 percent -- is on the NT platform," says Janet Perna, IBM's general manager of data management, in Somers.

But IBM hasn't completely bowed to Microsoft's popularity. Underlining its commitment to mix and match products and support open environments, IBM has adopted Java with a zeal that would make a missionary blush. The company has made the Sun Microsystems-developed technology integral to its technical strategies from multimillion-dollar mainframes to $100 handheld devices.

For instance, IBM is using Java as the technical heart of its San Francisco Project, a set of frameworks for building business applications, such as integrated financial systems, by combining pre-built components such as accounts payable and receivable modules.

Also, IBM along with Netscape, Oracle, and Sun, has embraced Enterprise JavaBeans as a programming model and it has begun using the Extensible Markup Language (XML) to tag data for easy exchange between its middleware and database products and products from competitors that support XML.

This leads IBM to a software conundrum. Microsoft hasn't adopted the same "standards" as the rest of the industry, most notably the JavaBean component model. How IBM manages to reconcile its support for NT with its support for JavaBeans remains to be seen.

Revenues look good

But unsolved problems are in the nature of the industry, and they don't appear to be hampering IBM software. To prove its software operation is now working, Thompson says that since 1994 IBM has been able to remix its host to distributed software revenues from a 90 percent to 10 percent split to 65 percent to 35 percent. The overall growth of IBM's software sales has gone from 3 percent in 1994 to 10 percent as of this year's second quarter, he adds, with distributed software sales growing roughly 20 percent per year for the past four years.

Critics contend that much of the company's gains in distributed software are due more to the trend toward distributed platforms than to IBM's reinvented software group. Thompson responds that IBM has outpaced most competitors in the distributed arena.

At the same time, a counterbalancing trend, in which companies are re-centralizing servers and using Web-based thin clients to distribute data, appears to be helping IBM as well. Thompson notes that while it grew its distributed business, IBM also grew its host software business. During the past four years those sales have gone from about negative 4 percent to about 4 percent growth this year.

"We have counted on customers realizing they need big processors to anchor their Web serving e-commerce applications, and that anticipation appears to be paying off," Thompson says. The revived corporate interest in mainframes has resulted in impressive gains for IBM's host-based software. Thompson says its Tivoli management software, DB2 Universal Database, MQSeries messaging software, TX transaction series have all grown at more than 50 percent of margin during the past year.

IBM plans to milk the host cow for some time. As proof of its commitment to host-based technologies, IBM will spend more on research and development this year for its IMS database than it did 30 years ago when the product was first introduced, according to Mills.

Desktop hopefuls

Despite IBM's technical strengths on host systems and its financial success on both host and distributed platforms, critics say they believe the company needs the equivalent of a hit single on the desktop -- something that applies a glossy finish to its gray enterprise product lineup. During the past decade, IBM has tried and failed with several key products including OS/2 Warp, and several desktop applications including Lotus SmartSuite series.

"In severe contrast to the success they have had the last couple of years with its host-based stuff, the market is still waiting for [IBM's] first big winner to be delivered in the PC arena," says John Dunkle, president of Workgroup Strategic, a consultancy in Portsmouth, N.H.

IBM officials and some analysts say the company already has that hit single in its Notes groupware product with 25 million user seats and counting. And although analysts say Lotus is still shipping more Notes servers per quarter than Microsoft is with Exchange, the gap between the two has narrowed from what seems like 1 mile to only a few inches during the past year.

What promises to keep Notes a half-step ahead of its archrival is the long-awaited 5.0 version of the product now due "to come down the chimney and not out of the pumpkin," which is Lotus President Jeff Papows' way of saying it will be released in late December and not, as once expected, later this month.

The continued success of Notes with Version 5.0 is important to IBM, not just to fuel its economic growth and to fend off Microsoft at the departmental-server level, but as a vital anchor to its overall e-business server strategy. Any further extended delays or major bugs shipped in the final version could cost IBM critical momentum.

"I think Notes [Version 5.0] is to IBM ... what Windows 95 was in importance and criticality to Microsoft," Papows says.

Another possible desktop hit could come from IBM's WorkSpace On-Demand server, which serves up Java-based customized desktops to users from central servers.

Being in start-up mode has given IBM software an agility that is unusual for a large company. Still, IBM has problems no true software start-up has, including a huge installed base to keep happy and a fuddy-duddy image to overcome. These are complex challenges to overcome in the brave new world of e-commerce, but at very least, unlike most start-ups, IBM software doesn't have to worry about venture capitalists who want their money back trying to sell it to Microsoft.

Ed Scannell (ed_scannell@infoworld.com) is an InfoWorld editor at large and veteran IBM reporter.