A Bridge Too Far: IBM's OS/2 Gambit



Sponsored by







INTRODUCTION

Many people who have observed IBM's marketing adventure with OS/2 Warp have come to the conclusion that IBM was either inept in its marketing activity, or else not sincere in its strategy. Some have even postulated that some IBM'ers were compromised by operatives of rival companies -- perhaps even from Microsoft itself. Somewhere between soap opera and cloak-and-dagger seems to be where most people will place this baffling episode of IBM's otherwise stellar return from its near-death experience in the early 1990's.

But as with most things technological, the truth is somewhat more complex than what "conventional wisdom" would have us believe. The cast of characters includes not only the fickle computer-industry media, IBM software and management executives, Microsoft acolytes, and the public. There is also an element of IBM management from other -- dare we say "rival" -- divisions within IBM. Not only the IBM PC Company but even the leadership of IBM Microelectronics was apparently involved in the boardroom intrigues. Let us examine how this came to pass.

IBM IS A HARDWARE COMPANY

Despite the fact that IBM sells billions of dollars worth of middleware, operating systems, and e-business applications each year, IBM is just what the name stands for: International Business Machines. IBM is a hardware company that bundles software and services with its equipment. They sell some of the very best software in the world, such as the DB/2 database and Lotus Notes and Websphere. They sell some of the finest operating systems on the planet -- AIX Unix, OS/400, OS/390, and of course OS/2 on the desktop. But the purpose of all of these software products is to sell hardware, or to at least make selling hardware easier. The fastest and most reliable computing machines in the world will not sell if there is no software available to operate the machines, to process instructions, and to monitor activities.

Software has therefore always been considered a "necessary evil" by most departments at IBM. In order to take advantage of the world-class excellence of IBM hardware, the IBM software must itself have world-class reliability and superb features. But most IBM executives have never viewed IBM software as standing by itself, and indeed they would not want it to. They would view such a situation as a case of "the tail wagging the dog." At IBM, software exists to fulfill the needs of hardware divisions -- it's as simple as that.

UPSETTING THE STATUS QUO

IBM's innovative desktop PC took the world by storm. Its openness had allowed Intel and Microsoft to take the market-share lead by a unique combination of commoditization and monopolization. But in the early 1990s, IBM was still under U.S. Government supervision from a 40-year-old antitrust case. IBM could not afford to risk attracting any attention by appearing to monopolize the fast-growing PC market themselves. IBM's previous attempt to recapture control of the PC market by introducing a proprietary bus architecture known as MicroChannel had failed, branding IBM as a dinosaur that was out-of-touch with the openness that most techies found appealing. IBM learned its lesson and never again attempted to force-feed a closed architecture into the PC hardware or software markets. But how could IBM recapture control of desktop standards without introducing proprietary architectures?

Like a police dragnet cast ever-wider as the culprits slip away, IBM moved out of the bus-architecture business and moved toward the motherboard business. IBM Microelectronics was tasked with providing alternative CPUs that would aid IBM in producing an Intel-free motherboard. During this period from approximately 1992 through 1996, IBM experimented with "Blue Lightning" CPUs, and also began cutting fabrication deals with AMD and Cyrix. IBM Microelectronics also was deeply involved in a parallel architecture path known as PowerPC, part of the IBM-Apple-Motorola triad. The goal was obviously to provide a new, Intel-free hardware architecture to wrestle control of the PC hardware industry away from Intel. But rather than playing the "heavy," IBM was offering openness (having learned a hard lesson from the MicroChannel debacle). IBM published the PowerPC chip and board specifications, hoping to jumpstart a new generation of Intel-free personal computers.

But what software would people run on this new wave of desktop machines?

MICROELECTRONICS AND SOFTWARE JOIN FORCES

As deeply as IBM's hardware people were offended by the growing clout of Intel, IBM's software people were even more disgusted with the decrepit, brain-dead software being pushed by Microsoft. So the software execs at IBM were only too happy to help with the PowerPC project. IBM's goal became not merely a thrust to sell software on the Intel platform and put a dent in Microsoft's desktop monopoly. IBM was trying something a level beyond that -- IBM was working toward an entirely new hardware/OS duo that would combine IBM's PowerPC chip with IBM's OS/2 operating system to become the next generation desktop platform. In other words, IBM was maneuvering to put both Intel and Microsoft out to pasture in one swift move.

From the executive boardroom, this move looked brilliant. Since PowerPC could run Apple software, IBM would gain entry to Apple's lucrative entertainment and education customer base. Since PowerPC could run IBM's AIX version of Unix, IBM would be able to offer a one-size-fits-all desktop and server computer, with an easy upgrade path from Apple, OS/2, and legacy Windows NT systems. By offering a single hardware platform for all these OS's, IBM could make a pitch to companies struggling to integrate multiple hardware from disparate sources. The key ingredient to this plan was to provide a viable stepping-stone or bridge platform to entice DOS and Windows users off of Intel PCs and onto PowerPC. That key element of the plan was known as OS/2 for PowerPC.

A BRIDGE TOO FAR

A famous World War II battle was lost because Allied forces misread their instructions and advanced too far ahead of the main thrust of the attack. They literally went one bridge too far down the road toward the enemy, leaving themselves without support and contact with the rest of their comrades. IBM repeated that error by failing to recognize the opportunity to push OS/2 as a software product in its own right. They failed to recognize the unique opportunity to put a brutal whipping on Microsoft's mediocre, unreliable kludges on the Intel-based desktop. Instead, IBM went for the whole enchilada by attempting to develop a version of OS/2 specially tuned to work efficiently and reliably on the PowerPC platform, in hopes of providing a complete spectrum of operating systems for this nascent would-be Intel-killer.

However, OS/2's architecture does not translate well to match the machine internals of PowerPC. Instead of admitting defeat, IBM's software division insisted, quarter after quarter, that they were only a few weeks from achieving the "holy grail" of a fast, reliable OS/2 system native to the PowerPC platform. IBM Microelectronics swore up and down that OS/2 could achieve up to ten times better performance on PowerPC than on Intel, which would allow IBM to offer a far superior hardware/software combo than any rival PC maker. In fact, this new platform would be far superior to IBM's own PC Company products. How do you think the IBM PC Co. felt about that??

SIBLING RIVALRY

The leaders of the IBM PC Co. were no stooges. They realized that IBM Microelectronics would soon be dictating architecture design to them, and obsoleting everything that the PC Co. was selling. Not only that, but if OS/2 ever became truly successful, the IBM PC Co. would have to learn a whole new operating system -- raising training costs and adding new support issues. IBM PC Co. executives did not want to get caught in the "squeeze play" between IBM's software division and IBM Microelectronics. So the IBM PC Co. stalled and dragged their feet, never fully supporting OS/2 preloads on IBM PCs. This negative attitude continued even through early 1995 when IBM CEO "Sweet Lou" Gerstner temporarily ordered all IBM PCs to be preloaded with OS/2.

IBM's PC salesmen never developed a positive attitude of salesmanship toward OS/2. Why should they? If OS/2 succeeded in taking a large share of the market, the IBM PC Company would have to deal with an angry Microsoft. They would also have to redesign every IBM PC to become a PowerPC machine eventually. Everything the IBM PC Co. offered would be discarded, and the executives in that division did not want to let that happen. This is why IBM never really took advantage of its PC division in the race to get OS/2 onto the desktop, even when the order to do so came from the top. Since OS/2 was part of IBM Microelectronics' plan to eventually replace the IBM PC Co., why would the IBM PC Co. play along?

OS/2 GETS SQUEEZED

By mid-1995 the PowerPC dream was almost over. IBM executives finally realized that no amount of continued resources (money, time, and people) was going to get OS/2 on PowerPC to become a speedy alternative to the rival platforms. Suddenly the IBM PC Company was no longer being squeezed between IBM Software and IBM Microelectronics. Instead, OS/2 was getting squeezed out of the way. IBM suffered great embarassment at having to ship PowerPC machines with Apple or WindowsNT operating systems preloaded. Rather than retreat one level and push OS/2 on the Intel platform, IBM retreated from OS/2 as a strategic platform altogether.

Sure, the leaders of IBM's software division knew that OS/2 was (and still is) superior to anybody else's desktop OS. But IBM did not want to push OS/2 on the Intel platform, legitimizing Intel as a realistic alternative to other IBM offerings such as AS/400 and AIX-based machines. So it became more convenient to knuckle under to Microsoft's Windows monopoly and let Intel-based PCs keep their well-deserved reputation for being cranky, unreliable kludges. If Intel could not be marginalized by PowerPC, then maybe Intel would at least be marginalized by its association with bugginess in the minds of long-suffering Windows users. The Intel-based OS/2 that was merely supposed to be a bridge to get people using OS/2 on PowerPC no longer had a strategic role at IBM, if there was no longer a desktop PowerPC system to build a bridge to.

CONCLUSION

IBM's grandiose vision of an all-encompassing PowerPC platform to fight the Wintel duopoly did not come to fruition because it was simply not technologically feasible to make OS/2 run fast and sure on the PowerPC platform. IBM could not afford to offer a PowerPC system as a mainstream platform if its power would only legitimize rival software offerings from Microsoft and Apple. IBM could not afford to continue pushing OS/2 on Intel if its power would only legitimize rival hardware offerings and eat away at AS/400 and minicomputer sales. Since IBM's gambit depended on OS/2 running natively on PowerPC, both PowerPC and OS/2 prospects were damaged by the failure to deliver this product in a timely and effective manner.

As a result of IBM's miscalculation, Motorola today is hurting, Apple continues to be viewed as a marginal platform, and IBM's software chariot has turned back into a pumpkin, leaving IBM's software division once again as a neglected stepchild. Only the people who work with OS/2 up close and personal realize that IBM did not have to bet the farm on a totally new hardware/software platform. We who know OS/2 know that it could and still can stand on its own as a shining example of excellence and power on the Intel hardware platform. But neither the IBM PC Company nor IBM Microelectronics perceive any benefit to a revitalized OS/2. For failing to accomplish an impossible task, OS/2 pays the price for having been ordered to advance a step too far in the battle of the desktop.







CREDITS





[ Top | Home Page | Faq ]


Copyright © 2001, Tom Nadeau.
All Rights Reserved.

E-MAIL: os2headquarters@mindspring.com