The Yin and Yang of Open Source Commerce, Part 2
By: John Terpstra
Wednesday, November 2, 2005 09:12:43 AM EST
URL: http://www.linuxplanet.com/linuxplanet/reports/6064/1/
Estimating the Global Number of Servers
In Part 1 of this article, attempts were made to identify the total potential market in respect of the number of individual consumers as well as the business consumer market. Part 2 makes use of the statistics available in an attempt to better identify sub-markets that can are feasible for rapid growth of commercial Linux.
The total number of servers installed globally can be estimated based on some insider market statistics. The values can be compared against IDC Market Research Information, however the comparison fails by a factor of two because of a significant use of computers that are sold as workstations that are actually used as servers in the <20 and 20-99 employee market segments (see Table 1). When the estimated use of workstations that function as servers is removed, the installed server base compares favorably with the IDC estimate of 21 million machines sold as servers [IDC Report: World-wide Client and Server Operating Environments Forecast and Analysis, 2002-2006: Al Gillen, Table 6].
| Table 1: The Global Server Market | ||||||||
| Company Size | # Employees/Company | Average Users/Server | US Number (2002) | Global Factor | Global Number | # Servers | % of Total Servers | % of Companies |
| 500+ | 3500 | 45 | 16,845 | 3 | 50,035 | 3,930,500 | 7.67% | 0.18% |
| 100-499 | 142 | 10 | 82,334 | 4 | 329,336 | 4,676,571 | 9.13% | 1.16% |
| 20-99 | 39 | 10 | 508,249 | 5 | 2,541,245 | 9,910,856 | 19.34% | 8.96% |
| <20 | 9 | 7 | 5,090,331 | 5 | 25,451,655 | 32,723,556 | 63.86% | 89.70% |
| Total | 5,697,759 | 28,372,771 | 51,241,483 | 100.00% | 100.00% | |||
Estimates of the installed desktop, laptop, and network client computer systems are difficult to obtain with accuracy. Based on various estimates, including the number of user licenses sold as claimed by vendors such as Microsoft, the total number of client and stand-alone systems is between 350-600 million. The real number is probably at the lower end of this estimate, given that a part of this sold infrastructure is clearly obsolete and that some licenses account for purchase of replacement systems.
Market demographics can be used to bring into question the behavior of the information technology industry as a whole. It also leads to a plethora of questions regarding the wisdom of the current commercial Linux businesses in their choice of target markets.
Market Schizophrenia
Microsoft is the dominant technology supplier across the entire information technology spectrum. Microsoft has approximately 60% market share in the installed server market, and approximately 90% in the desktop operating system market. Other sources provide estimates that are higher.
The heart of Microsoft's business server technology centers around a distributed network directory service called Microsoft Active Directory. Comparing the network directory service-centric focus with the number of sites that really need it, the question should be asked if there is an excessive preoccupation with this technology compared with needs of the market as a whole. In other words, 0.18% of all customers may need a directory service, yet 100% of Microsoft's server technology is predicated upon using it. Why is this so?
Technicians are easily excited over new technologies, and the more complex the subject may be, the more interesting it is to the technician, and the more confounding the impact on the uninitiated. Has the IT world allowed itself to be derailed with needless complexity? Linux does not need a directory service to satisfy the needs of 99.8% of the market place. Why does Microsoft Windows need it? While that question was a digression, it is a vital question all the same.
Since Table 1 shows that 0.18% of all customers (the percentage of 500-plus employee companies) account for 7.67% of servers sold, it is understandable that there should be strong competition in the enterprise market segment. But that does not adequately answer why it should dominate IT practices to the exclusion of any significant market presence for Linux vendors outside of this segment. Most of the IT world's needs can be met with simple, non-complex IT solutions. Should we not first strive to take care of our bases, before attempting to build castles in the sky?
The world is preoccupied with the need for Linux to achieve a dominant share of the high-end server market as the acid-test of being enterprise-ready. The mental assertion being that when Linux reaches maturity in the enterprise market place (500+ users per company) it will magically be ready for the market as a whole. This patently ignores the fact that small sites do not have the same complexity of needs as exists in the enterprise market.
There are approximately 50,000 enterprise-class companies in the world. Of these, a number have complex information technology needs, but certainly not all. If IT press reports regarding successful IT businesses are to be believed, the complex needs of enterprise class companies determine development objectives for the major IT vendors for all market segments. In practice of course, companies such as IBM and HP do produce products that are carefully designed to meet the needs of specific market segments--witness for example the HP OfficeJet all-in-one office fax, printer, and scanner product line.
Red Hat Software, SuSE, Caldera, Novell, TurboLinux, LinuxCare, and their peers in 1998-2000 were all focused on gaining credibility in the enterprise market place. It is a sad reality that every major IT company has its primary focus in the same puddle. IBM, HP, Dell, Intel, AMD, and Gateway are well aware that the bulk of profitable business is derived from a more mundane source--the small to medium business (SMB) market and the lower end of the small to medium enterprise (SME) market.
If the enterprise market is the only one that can be profitable, it is to be expected that a concentration of competitive activity should be observed. Perhaps it is necessary to consider a few additional factors so that the wisdom of the present fiasco can be better understood. Perhaps we shall see that the most profitable market has not yet been tapped.
The Slingshot Effect
One factor that led to the concentration of enterprise-only effort by Linux companies in the years 1999-2001, was the mention by Linux initiative leaders in companies such as HP, Compaq, and IBM that IT is just business. They indicated that the only justification to market Linux-based products and services is to increase overall company revenue. It was stated by more than one business unit leader that existing servers that are sold with Microsoft operating systems produce a higher per-unit revenue than if the same system were to be sold with Linux as the core operating system. Simply put, it is not in the interests of the major hardware vendors to actively promote the sale of business solutions that reduces overall revenue and at the same time antagonizes a major business partner.
Linux companies, on the other hand, sought to work with the large vendors in the hope of capturing large business opportunities that would rapidly catapult them towards commercial profitability. This was predictably a game of Russian roulette. In doing so, the Linux companies knew that Linux business units inside the large IT vendors were creating tensions. Not one company was willing to actively promote a Linux platform solution to the SMB/SME marketplace that would compete directly with Microsoft server-based solutions, even if it could be demonstrated that this was the most viable target market for a Linux initiative.
The determination of the Linux vendors to "make it big" resulted in nearly all development budgets being spent on chasing the needs of the enterprise customer. The result of this is that even today there are no Linux server solutions that are as simple to install as the Microsoft Small Business Server, and provide the same type of end-to-end business application solutions framework to the SMB/SME marketplace. On the other hand, after six years of intense development, Linux is a truly viable alternative to the high-end UNIX platform, and in many areas has replaced it. It is no accident that Sun Microsystems has released the source code for Solaris, except as a bid to quench the heat that Linux has created in their market stronghold.
Playing with the Big Boys
The Linux operating system vendors saw an urgent need to break into the market. They chose to ride to market on the coattails of the OEM hardware vendors. This automatically led to the decision to target the enterprise market space, a move that made it imperative to position Linux against traditional UNIX offerings, but in markets they did not already serve. For example, IBM and HP would rather capture UNIX business that was held by Sun Microsystems than to displace their own UNIX business with a lower-revenue Linux business solution. In hindsight, the fact that the primary hardware vendors were prepared to do this makes sense, and it was also a successful strategy for capturing UNIX business, which consisted of so-called "low hanging fruit."
This is a good moment to reflect on the Linux marketing strategies adopted by Compaq, HP and IBM, to illustrate the manner in which Linux business initiatives were clearly chosen so as to protect existing business units within these companies.
Compaq's high-end UNIX platform was the Tru64 Alpha CPU-based OS that they had acquired when they purchased Digital Equipment Coproration. The Intel CPU-based UNIX solutions Compaq were selling were based on SCO UNIX products (SCO OpenServer and SCO UnixWare). Tru64 and the Alpha CPU was a dying architecture and Compaq needed to replace it--Linux on Intel came onto the scene at an opportune time.
The decision to back an Intel-based Linux offering would keep the competition away from Compaq's major partner--Microsoft. This decision also pleased Intel because it helped to expand the market for Intel CPUs. This was clearly a simple business decision. Ultimately Compaq were unable to sustain competing activities with the result that in 2002 they were acquired by HP.
HP offer HP/UX-based UNIX business platforms. That is business they would not willingly cannibalize with a Linux solution offering. However, the major part of HP's revenue comes from Intel/AMD-based platform sales involving Microsoft business solutions. It is not in HP's interests to irritate Microsoft by helping Red Hat or Novell to sell a Linux-based business platform that will compete directly with a Microsoft solution.
Business is business--you make money by chasing the money, not by being altruistic. HP's behavior in the marketplace makes sense in this context. To top it off, they must keep share-holders at bay also. This is an ongoing challenge for HP, one at which they are being successful, according to IDC reports. On the other hand, we can expect that HP's shareholders would like to see the company more profitable with greater market share. How will they do that? What is the probability that HP will use Linux-based business to achieve this in the foreseeable future? Not very high, I suspect.
IBM has made a huge investment in Linux over the past six years. That focus has driven web-based Linux services, which is hardly surprising given the explosive growth of the web server market. IBM has driven a compelling argument for the use of IBM hardware and professional service sales into the enterprise market place and thereby created a market for Linux systems that could compete with Sun and HP offerings at the lower end of the enterprise customer spectrum.
IBM has been non-partisan with respect to Linux. They have supported all Linux vendors and continue to encourage business customers to purchase Linux-based solutions. Understandably, IBM has not specifically positioned Linux-based solutions into the SMB market because there really is no core infrastructure Linux-based solution that is suitable today.
I am sure that all three companies mentioned would agree that their Linux business initiatives have been financially rewarding and successful. But could they have been more successful?
Summary
It has so far been demonstrated that the global It market is largely untapped. Only 14% of the global population is an IT consumer today. Further, analysis of market statistics has demonstrated that the Linux businesses have so far targeted the top-end of the market very successfully.
The next part of this discussion will attempt to identify market opportunities and what may be need to create a more OSS friendly, profitable, and significant Linux presence that can redefine the IT industry.