How Microsoft's Enterprise Desktop Stifles Linux and How to Fix it

Advocates of open source initiatives seem puzzled by their inability to penetrate the Enterprise desktop market. How can you blame them? From a technical perspective, the main Linux distributions provide a superior operating system and user experience at a lower cost. As you will see, that's the same strategy Microsoft originally used to capture the market.

The root cause of Microsoft's hold looks insidious. Even the best market strategists I know have failed to recognize the inapparent reason Microsoft maintains a commanding lead in the desktop space. I have explained how to break Microsoft's hold so many times that I am puzzled by the glaze over executive's eyes.

I saw how Microsoft achieved its market position in the enterprise. I watched from the firing line and even participated. So, here's how Microsoft did it and how to undo it.

Linux and Open Source Desktop Strategies

Open source advocates have put some strategies into play to erode some of Microsoft market share in the installed base. They have failed to cause much more than a ripple since Microsoft continues increasing its market when people buy new computers. OSS advocates have provided two application suites that test Microsoft's vulnerability. Those include the Mozilla Firefox Browser and the Openoffice.org Productivity suites.

I have listened to and worked on the business case many times. It goes like this: "First, we replace Microsoft Office and Internet Explorer with free products of comparable or better quality. The users become accustomed to those applications. Then the switch to Linux, OpenSolaris, Mac and/or FreeBSD becomes a non-event. People will have a similar experience with the majority of their applications.

"Once we have the products, we start with government and migrate them to Linux. That will put us back into a Posix compliant environment. Enterprises will follow as we start stressing standards".

The Open Document Format (ODF) represents the most formidable standard open source advocates have used. Entire countries, regions and state and local governments have chosen this standard. Still, the goal of any significant migration to Linux has remained illusive.

Where the Strategy Breaks Down

Microsoft embedded itself in the enterprise with something other than Internet Explorer and Office. The loss leader in their product line comes with Microsoft Office and requires a back office component to work. Until someone replaces Outlook, the opportunity to expunge Microsoft from the enterprise will remain illusive.

At companies such as Ericsson, I saw massive migrations to Exchange server prior to the year 2000. Convinced by a new CEO that they had to fix their year 2000 problems, Ericsson, a company with 100,000 heterogeneous desktops, switched from their own mainframe email solution to Exchange. They left 30,000 users without access to the corporate directory, meeting management, etc.

I orchestrated the building of a client to run on Solaris workstations only to face rebuke. The Exchange administrators did not what a CDO proxy running in their infrastructure. So, they got the company to buy their engineers Intel laptops to run Windows just so they could have Outlook. At the time they spent about $4,500 per user. You run the numbers.

I offered a free, open source Outlook clone which used a small service to work with Windows Outlook in a corporate environment. Instead, Ericsson and other companies threw more money at the problem and made a lot of engineers unhappy.

Once you have control of the lines of communication in an organization, you own it. If you empower an executive vice president to come between the CEO and the rank and file worker and mid-management, the CEO becomes ineffective. The same with Information Technology. Whoever owns the communication lines controls the organization.

Microsoft's Grip

Microsoft has a hold on corporate messaging with Microsoft Exchange. Despite the many problems with Exchange we wonder if enterprises will ever move to another solution? Each time another solution looks to make a breakthrough, Microsoft changes Exchange to make corporations more dependent on it. For example, Microsoft took the global directory out of Exchange, put it into the enterprise and called it Active Directory. Now, after a long partnership with Blackberry, Microsoft will incorporate new technology in Exhange with SP2 to allow Windows CE appliances do he same things that Blacberry does.

In 1996, two representatives of Digital Equipment Corporation (now HP via Compaq)

staged a small seminar at a consulting firm for whom I worked. During the

half-day program, the moderator introduced us to DEC's five business

initiatives. One of those initiatives stuck in my mind as odd. I never forgot

that sense of oddity. Having worked with everything from large IBM mainframes to

the infamous Osborne Computer, the moderator's claim that DEC and Microsoft

would corner the corporate marketplace with NT seemed pretty off in left field.

At the time, UNIX and Netware combined to own 80% of the network nodes in the

world.

Microsoft and Unisys had just pulled off a major win by having NT declared an "Open System" by a US Federal procurement arbitrator. Now, Microsoft could move into Government procurement. Armed with that sound byte, the time to move had finally arrived for Microsoft to capture market share.

Then I saw an article by David Herdman, published on September 1, 1996, in

Digital News & Review, which led me to believe that DEC and Microsoft

had serious intentions about controlling the Network Operating System market.

Here is an excerpt from the article:

DEC's research shows that by 2000, Windows NT and UNIX will each have 40%

of the server OS market, compared to the present market share of 5% for NT and

50% for UNIX. The current and anticipated moves by many organizations to NT

servers and workstations support the findings of this research. The cost of

software is one of the reasons for the migration towards NT. For example, core

tools of Oracle for Digital Unix cost $1,675 per user, while the same tools for

Windows NT cost $250 per user. Though some of the advanced UNIX tools are not

available for NT, for average identical functionality, NT software costs less

than a sixth of the price of corresponding UNIX software. The current form of

Windows NT Server is far more suitable for desktop machines than for use as an

enterprise platform for large machines that many organizations depend

on.

In reading Herdman's article, I wondered why DEC (once the number two computer company behind IBM) would undermine its

own UNIX operating system. Then, I began to believe that DEC saw its future in

NT, not in UNIX. Microsoft could benefit from having a premiere UNIX company

help them extract UNIX as competition. If correct about that assumption, I saw

how Microsoft could sway perception away from UNIX and toward themselves.

As

recently as 1997, Novell appeared as the insurmountable leader in

Network Operating Systems for the PC. Microsoft had surprisingly made major

gains with Windows NT. I recall reading another article by Laura DiDio in the

January 13, 1997 issue of ComputerWorld which said:

Microsoft's Windows NT Server network operating system (NOS) is expected to

continue to take market share away from Novell's NetWare in 1997, although

NetWare currently commands between 55 percent and 60 percent of the NOS market,

according to International Data Corp. (IDC). Windows NT Server's links to the

Microsoft Office desktop application suite and Back Office server suite provide

the impetus for future market share gains. Some users are concerned that

Microsoft may come to dominate the NOS market the way it now dominates the

desktop. Windows NT currently holds a 20 percent market share, while IBM's OS/2

Warp Server and Banyan Systems' Vines account for 15 percent and 3 percent,

respectively, of the NOS market

[Note: Microsoft had a 20% market share and OS/2 15% at that time.]

So, with all the projections in the media, I wanted to know which strategy

Microsoft would use to beat UNIX and Novell. I decided to take DEC's offer to

put me in direct contact with some of their marketing executives. I met with

two key members of sales management who convinced me that Exchange Server would

make the difference in the NOS war. They asserted that by capturing the lines

of communication in the enterprise, they could control the enterprise.

Not having an attachment to keeping old seminar notes, I had to dig around

until I found something resembling the 1996 presentation made by DEC to our

firm. I found <SEC-DOCUMENT>0000950135-97-003873.txt : 19970918 in the

Edgar database at the following
href="http://www.sec.gov/Archives/edgar/data/28887/0000950135-97-003873.txt">location.

Therein lay the following information, a matter of public record. The underlines

belong to me. Remember, this text comes from Digital Equipment Corp.'s annual

SEC filing:

THE ALLIANCE FOR ENTERPRISE COMPUTING

Windows NT was designed for client/server environments. Under the Alliance

for Enterprise Computing, DIGITAL and Microsoft are working together to

provide products and services to support a single, integrated client/server

environment that extends from the desktop, across the enterprise and out over

the whole world through Internet communications.

DIGITAL has built the largest Windows NT service and support organization

in the world. We now have approximately 1,600 Microsoft certified systems

engineers and solution developers--a number that soon will grow to 2,500. We

developed software that supports the development of applications that will run

across Windows NT, DIGITAL UNIX and OpenVMS systems. We introduced the first

Windows NT clusters. And our x86 and AlphaServer systems own the benchmarks for

Windows NT performance. In tests conducted by Pro/E: The Magazine, a DIGITAL

Personal Workstation running Windows NT outperformed all workstations tested,

regardless of Central Processing Unit architecture or operating system. In fact,

it outperformed a similarly priced Sun system running UNIX by 171 percent.

Microsoft, for its part, is developing a 64-bit version of Windows NT for

initial release on Alpha. But the relationship between the two companies extends

beyond technology. It includes joint engineering, marketing, sales and service

programs. Working in partnership, we've helped customers build user-focused

enterprise systems and networks, and we have installed more than 500,000

Microsoft Exchange seats and have another 500,000 under contract.

Working with Microsoft, its partner in the Alliance for Enterprise Computing,

DIGITAL is implementing the largest Microsoft Exchange network in the world.

Running under Windows 95 and Windows NT, Microsoft Exchange provides a complete,

highly functional mail and messaging system that includes address book, auto

signature, remote mail and public folder functions. This last provides a forum

where authorized members of a workgroup or product team can share information

and ideas.

That year, DEC projected the sale of 2.5 million seats and sold 25 million. Imagine that. As a Microsoft Certified Professional, I had my plate full handling referrals. My little firm became a Gold Partner when that meant someting.

The Corporation has established a number of services alliances with

companies in the information technology industry. To support customer's

migration to Windows NT-based platforms, the Corporation has trained, and

Microsoft has certified, a professional services workforce of approximately

1,600 engineers dedicated to providing comprehensive systems integration and

service solutions. Under the Corporation's alliance with MCI Communications

Corporation ("MCI") and Microsoft, MCI delivers Internet and Intranet products

and services to MCI subscribers based upon the Corporation's Alpha servers,

Microsoft's Windows NT operating system, and Microsoft Exchange and Internet

Explorer software products, and backed by the Corporation's support and systems

integration services. The Corporation is a leading provider of mail and

messaging solutions in the global enterprise environment, having announced

during the fourth quarter of fiscal 1997 that it had passed the one millionth

Microsoft Exchange seat under contract. The Corporation also has been designated

a preferred service provider by Computer Associates International,

Inc.

As

I study the changes in the market, words from an interview with two DEC

executives keep running through my mind: "By capturing the lines of

communication in the enterprise, (one) could control the enterprise". By

controlling the perception which depends on communication, Microsoft made

Windows - fumbling, stumbling Windows - the de facto standard for desktops in

the minds of executives in the business world. Microsoft owned the consumer

marketplace with Windows 95. Now, they own the corporate world with NT technology.

Among Microsoft's many achievements, they have convinced people with the

barest of technical knowledge that Exchange should be their choice in messaging.

This has to stand as one of the greatest feats of manipulating perception in

modern times. One might even suggest that penetrating the enterprise must rank

with Microsoft's garnering the PC market by selling IBM on DOS, fifteen years

earlier.

To understand how pervasive Microsoft has become in the large corporations

that UNIX once dominated, consider this quote from CIO magazine:

http://www.cio.com/archive/071598_wars.html

There seems to be a sense of inevitability among vendors and IS executives

that Microsoft will eventually achieve its goal of making NT the dominant

network operating system. Microsoft is cultivating a strategy akin to IBM's

during its heyday, says Jeff Tash, president of Hewitt Technologies, an IT

consultancy based in Waltham, Mass. "It used to be said that nobody ever got

fired for buying IBM," Tash says. Microsoft, having won the desktop operating

system wars and now reaching for dominance on servers, wants a similar rarefied

reputation, he says. Such a reputation, partly based on potent marketing clout

that Unix vendors can't match, helps Microsoft sell NT to technically

inexperienced non-IS executives who have gained significant IT purchasing

responsibility in recent years. According to a September 1997 Windows NT

Adoption study by International Data Corp. (IDC), a sister company to CIO

magazine, 48 percent of departmental server purchases today are made by

executives outside central IS departments. Consequently, line of business

executives sometimes purchase NT servers without the CIO's knowledge.

"Before you know it, you've got hundreds of these NT servers all over the

place," says Thomas J. Bittman, vice president of platforms and operating

technology at GartnerGroup Inc. in Stamford, Conn.

Of course, Microsoft has been working vigorously to change that perception

(of scalability). In May 1996, during Microsoft's much-hyped Scalability Day,

Chairman and CEO Bill Gates declared that 95 percent of all businesses could

now run its enterprise applications on NT (see "Windows NT Scalability,"

CIO, Sept. 1, 1997).

Compaq has established a number of service alliances with companies in the

information technology industry. To support customers' migration to Windows

NT-based platforms, Compaq has trained, and Microsoft has certified, a

professional services workforce of approximately 2,300 engineers dedicated to

providing comprehensive systems integration and service solutions. Under

Compaq's alliance with MCI Communications Corporation and Microsoft, MCI

delivers Internet and Intranet products and services to MCI subscribers based

upon Compaq's Alpha servers, Microsoft's Windows NT operating system, and

Microsoft Exchange and Internet Explorer software products, and is backed by

Compaq's support and systems integration services. Compaq is a leading provider

of messaging and collaboration solutions in the global enterprise environment,

and is the leading integrator of Microsoft Exchange, with more than 400 major

accounts worldwide. Compaq also has been designated as a preferred service

provider by Computer Associates International, Inc.

So, what can we take

from this information as we ponder whether Microsoft can surrender an appropriate share of the enterprise desktop market? First, you have to wonder how UNIX and Novell could lose such a large market to the then fly weight Microsoft.

The Strategy to Which No One Would Listen

I orchestrated the building of an Outlook clone that ran on Linux, Solaris and AIX and HP-UX by late 1999. Exchange administrators at Intel, Boeing, Ericsson and others refused to let it into their infrastructures. So much for opening your source code if people don't have the skills to look at it and see it's safe and secure.

I orchestrated the building of an open source replacement for Exchange server that would run on IBM mainframes running Linux. It spoke to other Exchange servers in a peering fashion. Outlook clients saw it as an Exchange server. IBM signed us as their first Influencer Partner, Austin Ventures wanted to do a "B" round financing as did other VC's. Sun wanted to license the technology. We couldn't find anyone who could evaluate the code.

I contacted Research in Motion to discuss an agreement with them to work with us to port their Enterprise server to Linux. They didn't want to disturb their relationship with Microsoft. Today, some people call Exchange SP2 the Blackberry killer.

I had discussions with Apple about porting the Exchange replacement to OS X server. They wanted to do that at first, then decided their best opportunity existed in consumer goods - think iPod. Sun already had a mail and calendar server and didn't want to change horses in mid-stream.

Essentially, a solution for the Microsoft Exchange problem has existed for five years. The client existed before Evolution had a screen shot for people to see. In fact, Evolution couldn't send and receive mail when a product already existed that worked in tandem with Exchange forget an extension that needed someone to turn on web services for Exchange.

I offered the mail client to the Openoffice.org project. They decided to build one from scratch instead of using the expertise available to them. In case no one has noticed, Openoffice.org and Star Office haven't flown off the shelves of retail stores at any price. I purchased a Star Office 7 from CompUSA during Christmas of last year for $19 after an instant rebate and a rebate from Sun. I didn't see any takers standing in line. Sun still wanted $79 for the product from their online store.

People want Outlook and Exchange and that drives enterprise desktop sales of Microsoft Windows. You won't find enough difference between Openoffice.org and Microsoft Office to justify most people upgrading to the latter. But give them a messaging solution in the enterprise space and see what happens. Don't make that solution a radical departure - keep it close to what they have used. Evolution, by the way will not do it for the Microsoft enterprise crowd. That's not saying Evolution is bad, I use it happily.

What Else?

Having witnessed the emergence of NT first hand and knowing Microsoft's strategies for capturing the enterprise market, I have an argument for taking the market back. I wonder how many Linux users wonder the same thing today that I have for almost six years. If we have had the technology all along, what stopped us?

My answer to that question is simple: No one believed we would make it this far. I wonder if anyone is listening now?