Reflections on Open Source Commerce, Part 2

Tuesday May 27th 2008 by John Terpstra

Linux on the desktop has yet to gain any real market presence, despite the unrest over Microsoft Windows Vista. Author John Terpstra continues his look at the challenges facing the desktop and server markets, with a close examination of current market dynamics.

Linux-based business-critical servers have found a stable home in the data center and in infrastructure computing applications in organizations of all types. The low-hanging fruit has been harvested; most Linux servers have been installed to handle workloads that traditionally would have run Unix-based operating systems. This application area is rapidly maturing. The major opportunity for ongoing growth is in developing countries.

Linux on the desktop has yet to gain any real market presence. Despite the unrest over Microsoft Windows Vista, the companies that focus on Linux as a business have yet to deliver a go-to-market proposition that is compelling for the consumer, for the retailer, distributors, and original equipment manufacturers. The consumer segment is the fastest growing potential market for Linux-based desktop solutions. The rise of the OLPC and the ASUS Eee PC, together with Microsoft's reaction to them, is proof that the consumer market is opportunity rich.

But will Linux be the platform that delivers just good enough in time to create a paradigm shift from the desktop and laptop to the new-school ultra-mobile, wireless enabled, consumer device that will work transparently the world over providing previously unimaginable access to the all the information that will be sage-guarded, housed, processed, and delivered to you over the grid? We do not know. But we must consider what will become of those who will not, or can not, change their computing practices.

There will always be a transition market, and there will always be a residual market. This is perhaps the area that should be the target for Linux and open source solutions development, look at is as a training and preparation ground for the disruptive change that may follow. One question still begs an answer: How will all of this be delivered to the end user, the consumer?

p>Sharon Linsenbach believes that new technology companies are again seeking the help of the channel to leverage the resellers' relationship with the customer. She presents some interesting perspectives on the importance of the mechanism vendors might use to get goods and services into the hands of the consumer.

Linux distribution vendors have not made a particularly compelling inroad on the consumer desktop. Some would argue that Linux should never be a consumers' choice, instead they want to see the focus on the business desktop. The trouble is, no Linux distribution has so far been able to deliver the substance of what is required for large-scale corporate adoption. What distribution model will work to get Linux solutions into the hands of consumers, both the in the home market and also the corporate consumer?

The simple fact that few would deny is that every Linux distribution available today is more or less good enough for the majority of consumers. Why then is there no rush to get Linux-based laptops and desktop systems to market? What is the real obstacle that is responsible for holding back the flood-waters of a huge shift away from Microsoft Windows Vista? Is it because it is too difficult to re-educate the consumer? What is holding business back?

Let's speculate about one reason why OEMs are not rushing to ship Linux-based personal computing systems. Perhaps the sale of PC and laptop hardware is not profitable and that the only reason OEMs still sell them is because Microsoft pays market development funds that are the last remaining incentive to do business. I think you would agree that is definitely not the case, but if that were the case there clearly would be no attractiveness in a Linux-based solution.

Alternately, perhaps the commercial Linux distribution vendors just do not understand the market. An OEM would not want to change a comfortable relationship with Microsoft for a multitude of relationships with vendors who do not have a commercially viable business solution.

There are many channel dimensions. OEMs are just a small part of the bigger picture. Retailers sell products to consumers, but to the OEM they sell customer opportunity. This opportunity is expressed through a value that is placed on shelf space, key traffic areas within the retail store, staff competence, staff knowledge and support mechanisms. Getting an OEM to produce a Linux-based PC platform is merely the first step in working the channel to get products moving.

Consumers are not consummate open source technology buffs. Very few will find delight in performing an after-market installation of Linux on a PC or laptop system. The only way to convert them to use Linux is to provide it in a compellingly easy to use packaged form. If there is not way to do this, we shall have to find another market or else keep Linux in the closet. Maybe we should just take satisfaction that Linux is a great learning tool for want-to-be computer scientists. Be aware however, that if there is no market for computer scientists who are skilled in Linux, the future for these people must be somewhere else. Where would that be?

By comparison with the markets and channels to them for desktop computing solutions, the server market is well understood within the Linux world. The channels that service the server market usually employ a direct sale to the consumer model. S erver solutions are commercially viable because workload types can be easily identified , defined and satisfied with well understood technologies and applications.

The Linux Foundation recently commissioned a white paper that has been published by IDC ("The Role of Linux Server and Commercial Workloads," by Al Gillen, Brett Waldman, Elaina Stergiades. April 2008). This paper recognizes that over the past decade Linux has predominantly displaced UNIX in the server operating environment (SOE). The paper correctly identifies the fact that over the next decade Microsoft will aggressively defend its position in the SOE market and thus we can expect an all-out effort to displace Linux-based solutions with new Microsoft offerings.

Given the lack of financial muscle to preserve the Linux market share in the SOE space, we can expect some erosion of market share. The importance of capturing a larger desktop presence in order to defend the SOE market can not be overstated. In the absence of significant desktop growth for Linux, over the next decade we can therefore anticipate some loss of employment of open source developers by the Linux vendors.

This is a good point to reflect on the assertions and questions raised so far by examining some market statistics. The following are an update on what was provided in the Yin and Yang article.

Table 1 presents a summary of US Bureau of Census information showing the distribution of businesses by number of employees. The growth of the US work force between 2002 and 2005 was 3.48%, the number of large businesses ( more than 500 employees) grew by 3.75%, the total number of firms grew by 5%.

Table 1: US Business Distribution by Size—1988, 2002 and 2005
YEAR DATA TYPE TOTAL Number of Employees
<20 20-99 100-499 >500
1988     Firms 4,954,645 4,444,463 430,640 66,708 12,824
Establishments 6,016,367 4,516,707 581,622 244,697 673,341
Employment 87,844,303 18,319,642 16,833,702 12,761,379 39,929,580
2002     Firms 5,697,759 5,090,331 508,249 82,334 16,845
Establishments 7,200,770 5,147,526 692,775 332,508 1,027,961
Employment 112,400,654 20,583,371 19,874,069 15,908,852 56,034,362
2005     Firms 5,983,546 5,357,887 520,897 87,285 17,477
Establishments 7,499,702 5,409,151 679,382 331,999 1,079,170
Employment 116,317,003 21,289,196 20,444,349 16,911,040 57,672,418

In the Yin and Yang article world GDP was used to factor US businesses demographics to arrive at a world business and server usage estimate. Table 2 provides the most recently published GDP information. A column was added to report the ratio of area GDP to the global GDP. The factor for the US economic GDP contribution has remained almost static, while other countries show marked changes. These figures imply contrary to expectation that Indias' contribution to GDP declined over the past three years. It might be an interesting exercise to explore what the cause may be and to what extent this may bias the assumptions made in Table 3.

Table 2: Global GDP for 2007 (Est)
AREA 2004 2007 Estimates
US$ Billion Percent Global Factor US$ Billion Percent Global Factor
Total 55,500     65,820    
US 11,800 21.26% 4.7 13,860 21.06% 4.7
EU 11,700 21.08% 4.7 14,450 21.95% 4.6
China 7,260 13.08% 7.6 7,043 10.70% 9.3
Japan 3,750 6.76% 14.8 4,417 6.71% 14.9
India 3,320 5.98% 16.7 2,965 4.50% 22.2
Rest of World 17,670 31.84% 3.1 23,085 35.07% 2.9

Table 3: The Global Server Market
Company Size Employees per
Average Users/
US Number (2002) Global Factor Global Number # Servers % of
% of Companies
>500 3,326 45 16,845 3.0 50,535 3,735,098 7.20% 0.19%
100-499 193 10 82,334 4.0 329,336 6,356,185 12.25% 1.23%
20-99 39 10 508,249 5.0 2,541,245 9,910,856 19.10% 9.47%
<20 4 3 5,090,331 4.7 23,924,556 31,899,408 61.46% 89.12%
2002 Total     5,697,759 4.7 26,845,672 51,901,546 100.00% 100.00%
Company Size Employees per
Average Users/
US Number (2005) Global Factor Global Number # Servers % of
% of Companies
>500 3,300 45 17,477 3.0 52,431 3,844,940 7.07% 0.19%
100-499 194 10 87,285 4.0 349,140 6,773,316 12.46% 1.24%
20-99 39 10 520,897 5.0 2,604,485 10,157,492 18.69% 9.24%
<20 4 3 5,357,887 4.7 25,182,069 33,576,092 61.78% 89.34%
2005 Total     5,983,546 4.7 28,188,125 54,351,839 100.00% 100.00%

From Tables 1 and 2 has been derived an estimate of the number of servers that are used in global infrastructure computing. The findings previously reported for 2002 are shown together with the 2005 numbers in Table 3. The average factor shown in the total line equals that shown for the US in Table 2. As was done previously, it has been assumed that compared with global business size distribution, the US has a disproportionate number of large (more than 500 employee and 100-499 employee) businesses, so the factor is therefore lower in this area. It is therefore also assumed that the global ratio of 20-99 employment businesses is lower in the US than elsewhere, thus the factor for that was increased to compensate. Based on the assumptions shown, the total number of infrastructure computing servers has grown from 51.9 million to 54.4 million over three years. It is left as an exercise for others to confirm or dispute these estimates.

Between 2004 and 2008 it has been estimated that world internet usage has grown from 974 million users to 1.355 billion. The indicated growth rate calculates at approximately 8% compound growth per year. The usage growth column in Table 4 shows that most internet usage growth has taken place outside of the US.

Table 4: World Internet Usage Statistics
World Regions

Population (2008 Est.) (Million)

Population % of World Internet Usage, Latest Data (Million) % Population (Penetration ) World Users % Usage Growth 2000-2008
Africa 955 14.3% 45 4.7% 3.3% 903.9%
Asia 3,776 56.6% 512 13.6% 37.8% 348.1%
Europe 800 12.0% 374 46.8% 27.6% 256.1%
Middle East 197 3.0% 34 17.0% 2.5% 923.7%
North America 337 5.1% 243 72.2% 18.0% 125.2%
Latin America / Caribbean 576 8.6% 127 22.1% 9.4% 603.4%
Oceania / Australia 34 0.5% 19 56.5% 1.4% 151.6%
WORLD TOTAL 6,676 100.0% 1,355 20.3% 100.0% 275.4%

The Netcraft web server survey reported that on October 24, 2005, there were 74.4 million publicly accessible web servers globally. The April 2008 report claims that responses have been received from 165.7 million web servers. This is a significant growth in the number of web sites that are being hosted, however one needs to be careful not to conclude that this parallels the number of physical servers in use. Growth estimates for these are best obtained from IDC server shipment statistics.

It is clear from the world internet usage statistics that there is a significant rate of growth of potential desktop systems users. The growth rate in this area far exceeds the rate of growth of the installed server market and is therefore of primary interest as a means of gaining increased Linux adoption.

Linux has matured rapidly as a platform. It has gained a durable and significant market share in server applications, but penetration into the desktop space remains elusive. At this time it is not at all clear how Linux adoption will gain sustained dominant market share any time in the next decade. It will be interesting to see how the obstacles to desktop adoption will be overcome.

The attractiveness of faster, more powerful, more slick desktop systems has been almost lost. The consumer is content enough with solutions that are good enough, though clearly not perfect. Linux is good enough for the average consumer's desktop, but there are barriers to adoption that need to be overcome. This is becoming a critical hurdle. Some of the obstacles have been commented on in this article in the hope that a bright enterprising new-comer to the game will find a compelling and entrepreneurial solution that will finally deliver Linux-based solutions into the consumer's hands.

This article was first published on
Mobile Site | Full Site
Copyright 2017 © QuinStreet Inc. All Rights Reserved