Open Source Unix--Momentum Is Building.

Author and philosopher George Santayana once defined a fanatic as someone who redoubles his effort upon losing sight of his goal. Lately, that's all too often also a good description of some Linux boosters, who have lost sight of the goals of open-source software in their eagerness to attack Microsoft's computing hegemony. Linux, BSD, and other open-source software packages are not about destroying Microsoft; to expect that to happen is precisely the kind of thinking that led Don Quixote to take on the windmills. Instead, in the end, the open-source software movement is about giving more control of the developmental direction of software to the people who use it.

It's also about making Microsoft's software better. Redmond's oft-troubling anti-competitive tactics, which have earned it the enmity of competitors, partners, and the government, will not work on open-source Unix. There's nothing to buy, no one to pressure, no competitor to crush, only thousands of programmers around the world devoted to making software better. The only thing that will save Windows is making it better, too. It's fashionable to scoff at Microsoft's ability to deliver quality software, but no one has ever accused the software giant of failing to respond to a competitive challenge. They have always had the technical know-how needed; now they have the motivation.

Anarchy In Action?

Open-source software is distributed under a variety of licenses that generally guarantee the right to read, redistribute, modify, and use the software freely. The most famous example of open-source software is the Linux operating system, which has garnered support from major vendors such as Oracle, IBM, Hewlett Packard, Silicon Graphics, Sun, and Dell. Other examples include the other open-source Unix (OSU), BSD, and applications such as Sendmail, the Apache Web server, the Perl scripting language, and others that serve as the fundamental "plumbing" of the Internet.

Proponents claim that by supporting independent peer review and rapid evolution of source code, open-source licensing promotes software reliability, quality, and security. More to the point, it means that feedback from users can be acted on more quickly than is possible under the proprietary model of software development. This is the key to the success of Linux and its ilk and it is also the key to the opportunities and profits that await VARs and integrators in this burgeoning market. A canny VAR with in-house programming skills has in Linux and BSD a tool of unparalleled flexibility and more and more support behind it, above and beyond the already excellent distributed support offered by other OSU programmers around the world.

To identify and quantify the dimensions of the open-source Unix phenomenon, Survey.com and ENT Magazine established the Open-Source Unix Research Initiative, the first result of which is the largest survey ever conducted on open-source Unix in organizational computing. Based on the responses of over 2,200 decision-makers in private industry, government, and education, the results of the survey indicate a great deal of opportunity and some pitfalls. What follows is a brief review of the highlights of the survey, focusing on the areas where the specialized knowledge and capabilities of VARs and integrators can best be leveraged and outlining a profile of the best prospects.

Who's Who Of OSU

All respondents had purchasing authority or influence. 11% were corporate executives, almost 8% were line-of business professionals, and close to 75% were IT professionals. Almost 60% of these last were executives or management; other data we have seen emphasizes the importance to customers of communication at a high level--the vision thing--and this appears to be an even more powerful tactic for open-source Unix sales.

The mean revenue of responding companies was $3.5 billion. Almost a quarter of responding companies have more than 10,000 employees; the mean size of companies under that size is about 800 employees. The mean size of the smallest organization category (1-249) was 123 employees.

17% of responding organizations are IT-related firms, 13% government, 9% education, and the balance in a wide variety of industries, including business consulting/professional services, aerospace/automotive, Insurance/diversified financial, banking, finance/investments, and manufacturing. Some of these, however, are much better prospects than others, as discussed at the end of the article.

The Open-Systems Bandwagon 

The open-systems bandwagon is gaining momentum rapidly. Only 31% of respondents do not intend to deploy open-source Unix. By the third quarter of this year, 55% of all responding organizations will have deployed at least one Linux or BSD system. Adoption is faster among smaller companies: only about 25% of companies with less than 250 employees intend to wait until the fourth quarter or later, compared to almost 50% for the largest companies ([greater than]10,000 employees). Integrators serving the enterprise market have longer to prepare. The installed base of OSU in organizations already deploying it will almost double in two years.

About 6% of respondents who have already deployed OSU are using only BSD, about 73% are using only Linux, and about 21% are using both. The Apache Web server is the most common open-source application among both non-adopters and potential adopters.

Pros And Cons Of OSU

The main reasons chosen for not deploying OSU (in order of importance) are lack of application support, no perceived business benefit, lack of technical support, lack of internal expertise, and no perceived technical benefit. The factors most likely to change non-adopters' minds are (in order of importance) support for applications, improved vendor support, better support for drivers and peripherals, trained internal personnel, better user interface. Apparently, reaching non-adopters requires education about open-source Unix benefits and the assurance of good technical support. These are exactly the strengths of VARs and integrators.

Once customers are convinced, however, the picture changes radically. This is where the profits and opportunities are multiplying fast and thick. Among potential and current adopters, the top reasons for deploying OSU are (in order of importance) reliability and stability (tied for first), performance, security, and manageability, which are among the attributes of Windows that respondents rank lowest in comparison to open-source Unix. Not surprisingly, current adopters rank Windows even lower than potential adopters. Fanatics may assume this is because of the "how do you keep them down on the farm" syndrome, but it's more likely that organizations that have already adopted OSU did so because they were more fed up with Microsoft Windows.

Applications On OSU And Its Victims

The applications potential or current adopters are most likely to host on OSU are (in order of likelihood):

* Database server

* Firewall

* Web application server

* Web server

* E-commerce

Of the seven vendors whose support for OSU is most important to respondents, four are database vendors: Oracle, IBM, Sybase, and Informix. SQL Server, look out!

However, potential adopters, being farther back along the adoption curve, tend to prefer "first-wave applications" such as web server, file, and print server to E-Commerce; the latter is more the province of current adopters, who have already used those first wave applications as proof of concept. OSU growth on eight different enterprise server applications (including those listed above) will exceed 100% over the next two years. The number of organizations using OSU as the primary OS for database servers, for instance, will grow by 193%.

Any proprietary Unix and then NetWare and Windows NT 3.5X Server are the most likely operating systems to be replaced as part of a first implementation of open-source Unix. In organizations that have already deployed it, the most likely to fall victim to the open-systems juggernaut are, first, any Windows NT version and, then, proprietary Unix. For Windows 2000, things look somewhat rosier. In fact, Windows NT/2000 will continue to gain ground or hold its own in many areas, especially in larger companies. NetWare is one of the operating systems most dramatically impacted by OSU, which appears to be taking a great deal of server share away from it.

Not Yet In The Groove

About two-thirds of respondents who have already deployed OSU purchased it separately and either added it to an existing system then in use or to a new system. The remainder either purchased it separately and added it to an obsolete system no longer in use (not as many as folklore has it), or purchased it as part of a complete system. This last group, the one of interest to the channel, accounted for only a little over 14% of current adopters; only 7.5% of organizations who have already deployed OSU got it from a VAR or systems integrator. There's nowhere to go but up. VARs and integrators serving the federal government and non-IT-related private industry will find it a little easier.

The opportunities among individual industries vary widely. The best opportunities appear to lie among engineering/construction/architecture, manufacturing-consumer packaged goods, aerospace/ automotive, Web design development, and business consulting/professional services. The industries not to waste your time on (or to go after if you want less competition, and don't mind doing a lot of education) appear to be oil/gas/chemicals, insurance/diversified financial, utilities/energy, state or local government, and process manufacturing.

For VARs and integrators who want more detailed information, the full report is available for $995; the executive summary can be found at www.survey.com, or, for more information, email [e-mail]. The full data set in SPSS or Excel format, enabling analysis of the data in detail, is available for $5,000 (including the report).

Dave Trowbridge is the senior analyst of Survey.com, a market research firm specializing in data-intensive reports on information technologies, where he monitors operating systems, portals, and various aspects of business intelligence