IBM Makes Another Run at Windows: Can It Do Better Than OS/2 Did?


IBM has announced the "Microsoft-free" Virtual Linux Desktop as an alternative to Windows. Positioning Linux as a hosted offering is probably a better plan than running at Microsoft with a more traditional desktop strategy. On the other hand, general users have not been excited about Linux, UNIX, or even Windows Vista on their corporate desktops and they have more say about what they get than ever before. The only platform gaining significant share is the MacOS. Hosted solutions have similar issues to thin client offerings in that they depend heavily on a network connection and don't work when one is not available. Apple's approach to competing with Windows by obfuscating its UNIX core remains the only proven approach to displacing a Windows desktop.


A number of lessons came out of the OS/2 battle with Microsoft; this battle occurred when IBM was the largest and most powerful software company in the world, and Microsoft was still the smaller challenger. That dynamic is clearly reversed now. I was in IBM during that time as one of the folks trying to spin out the IBM software company, and we studied this battle very deeply.


None of this means Windows isn't vulnerable. It is more vulnerable than ever before, but to take it out requires an approach that addresses the user's needs first and IT's needs second. None of the challenging platforms, including Windows Vista, do that well enough to drive displacement.


Windows Opportunity and Difficulty


Microsoft doesn't seem to realize at a corporate level that Windows is perhaps its biggest keystone product and that, were the company to lose it, the cascading risk to its other offerings would make Microsoft a fraction of its current size and power. Vista bounced in the enterprise and the entire economic model that founds Windows and Microsoft Office is now at risk. Should both go, they will take IE with them, and put Windows Server, Exchange, and most of Microsoft's other offerings at risk, as well. You have only to look at IBM in the late '80s or the auto companies today to realize things can happen surprisingly quickly. And executives who thought they were safe can suddenly discover they are anything but. The difference can seem as quick as flicking a light switch.


But just because there is opportunity doesn't mean a competitor is sure to win. Two others are chasing this opportunity: Apple and Google. Apple can't scale quickly enough and has hardware pricing issues at the moment, but has still managed to push Windows down below 90 percent of the market for the first time in over a decade. Google hasn't yet rolled its platform. It is supposedly testing it on a third of its desktops internally now, but its applications offerings have largely failed to gain the traction Google had hoped.


It is my impression that IBM, Apple and Google should be partnering rather than coming at this separately. Each has a portion of what is needed for a successful move but none of them have all that is needed. Let's focus on why OS/2 failed and then the new IBM offering.


Why OS/2 Failed


OS/2 failed for one reason and one reason only. It didn't gain enough support from either other OEMs or developers to sustain itself in the market without unsustainable levels of funding from IBM. Even when it was given away, it couldn't reach critical penetration numbers. The end result was one of the most expensive failures in IBM's history. When it first partnered with Microsoft in the early '80s to create DOS, IBM realized that a platform from Big Blue would probably not fly and that the solution had to be owned by a company that didn't directly compete with the other hardware OEMs. The OS/2 disregarded that lesson and thus couldn't gain similar economies of scale to what a vastly weaker (at the time Microsoft was leveraging.


Strangely enough, IBM's own PC company hated OS/2. Many of the internal units I supported while there refused to use it, indicating serious problems with the offering that were swept under the table and not addressed. All of this also contributed to the failure. IBM simply would not do what was necessary to ensure the success of the offering, and a lot of IBM customers, along with IBM, paid a very high price.


IBM's Virtual Linux Desktop Offering


IBM's offering is a bundle of hosted Ubuntu Linux, Lotus Notes, and Lotus Symphony (which is based on Open Office), wrapped with IBM management tools and promising 50 percent savings over Windows. The savings would most benefit IBM/Lotus Notes sites (down to about 10 percent of the available e-mail market), which do not have Microsoft Enterprise Agreements. Eliminating a Microsoft Enterprise agreement has cost associated with it that will likely substantially offset the savings unless all of the related Microsoft offerings are displaced, and that isn't reasonable for most.


In many ways, this hosted offering is very similar to a thin client offering in terms of server loading and benefits. Both Wyse and HP have had a number of similar offerings in the market for years, but they have generally failed to capture more than a small percentage of the available market, even though users have actually preferred some of the thin client benefits with Windows -- these are generally best where terminals used to reside and don't yet address notebook users very well. This last is largely due to the still limited wireless coverage.


Notebook computers remain the fastest-growing and most profitable PC segment, now shifting to netbooks (they have some forecasting the death of Apple again), which do anticipate prevalent network connections, but those connections aren't yet prevalent enough to allow a hosting or thin client solution. Since IBM spun off its desktop business to Lenovo, this offering would be in clear competition to Lenovo's less disruptive diskless ThinkCenter offerings (which we covered last month). And, given that this is Linux, it is likely to be an ugly migration in that few settings and tools would easily migrate from Windows, and the path back, should it fail, would be prohibitively expensive.


Both Ubuntu and OpenOffice are, in the current economic climate, at high funding risk. While IBM is capable of assuring this funding, it has not yet taken a position that it will. It does have a history of standing behind its offerings, but the economic collapse will put a lot of pressure on every company to focus on where revenue is, and new offerings like this often take years to ramp to profit. The biggest issue, however, is similar to the one that killed OS/2: the lack of third-party support either from OEMs or developers for this unique desktop offering in a climate that clearly will have developers also focused on where customers are, not where they someday may be.


Finally, Linux has simply not been any more popular with general users than UNIX was. Even Windows Vista has not been popular enough to drive it through IT, and users have generally resisted having to use two separate solutions, one for home and one for work.


Wrapping Up


Users drove PCs and smartphones, and are the main driver behind Apple's success against Windows. IT in early adopter firms couldn't drive Vista into the enterprise, and this IBM offering is long on IT value but short on user value. Thin clients with similar benefits and vastly more mature offerings haven't been that successful, even with Microsoft support, suggesting that this one will largely bounce as well unless it is substantially improved.


The opportunity IBM sees is a real one. Microsoft hasn't been this vulnerable to attack since the late '80s, but it still has to be an attack that addresses what users see as problems, not just replaces the ones they know with a new set. In the end, the opportunity is likely a fraction of the existing thin client market, which runs into the low millions per year in terms of desktops. But, unless IBM applies the lessons learned from the OS/2 failure and grasps why Windows, the MacOS and smartphones succeeded, the outcome for this offering will repeat the one OS/2 enjoyed.


IBM did build the PC market. The opportunity does exist to recreate that success, but you can't ignore users when doing a desktop platform except in terminal environments (and even here it isn't particularly successful). For this to be successful, IBM has to obfuscate the Linux platform, make it attractive for users both at work and at home, and address the increasing mobile need. And it still needs to capture developers while not destroying the appeal to IT decision makers. There is a reason Microsoft's initial "embrace and extend" strategy was so successful.


Eventually, we will move to a solution that looks a lot like what IBM is offering. The market is clearly moving in this direction, but much like Apple improved on UNIX, for Linux to play here, it has to become something much more similar to the MacOS or Android first.