CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Linux: The penguin expands its habitat

Linux recently has become quite a hot topic in investment circles and for Internet infrastructure players.

Linux, the open source operating system, recently has become quite a hot topic in investment circles and for Internet infrastructure players. Vendors in this segment have had some of the most successful initial public offerings (IPOs) in history.

From a fundamental perspective, users and industry pundits alike are saying Linux and the open source software movement could be the biggest threat to Microsoft's dominance in the operating system market over the next few years.

But can investors take a movement seriously that claims a portly penguin as its mascot? They should. For one, Linux as an operating system is growing faster than are any of its competitors. It is also is an ideal operating system for appliance computing, a high-growth segment that I've commented on the past. Most impressive is that Linux systems are the operating infrastructure behind an astounding 31 percent of all Internet servers--giving it the No. 1 unit market share for this important segment

Meanwhile, Dell Computer, Compaq Computer and Hewlett-Packard have all announced varied levels of support for Linux. Recently IBM has started to throw its weight behind Linux in a large way. This endorsement by the hardware heavyweights validates Linux' potential in the OS market, I believe.

To back up, what exactly is Linux? It's an operating system derived from Unix that runs on multiple computer processors, including Intel's Pentium line, the PowerPC (for both IBM and Macs) and numerous RISC processors. Compared with the commercial version of Unix (such as Solaris from Sun Microsystems) it still lacks certain features needed for high-end processing. Yet most users agree that Linux is more robust and scalable than lower-end operating systems like Windows NT. Probably the most significant feature, however, is that Linux is available as an open source code, which means users can modify the source code as they desire.

The Open Source movement is important for a number of reasons. First, surprisingly there are more software developers in residence at user sites (for example, companies with internal IS staff) than there are at software companies--and they like open source. Second, the source code is available for free. Finally--and this is what gets investors excited--Linux is a truly open software platform, providing a foundation for future improvements. Those improvements are continually made by the Open Source community, which claims programmers in the thousands.

The question for many investors still remains: How can anyone make money from free software? It helps to understand the four basic business models (that I've seen so far) that separate focused Linux players.

The first model I've witnessed is the Linux software distribution model, as exemplified by Linux distributor Red Hat. Since Linux source code is always being modified, upgraded and debugged, it can create some user confusion as all the modifications are generally available.

A distribution vendor's goal is to simplify this process (and choose which additions and modifications make sense) and provide other software services. Distribution companies (and there are also a handful of private companies working alongside Red Hat) control the different versions of the OS, document the versions and provide the actual software and customer support. The model has shown that customers are more than willing to pay $50 to $100 for "free" software when it comes with added support and is consistent from server to server.

The second model is the appliance model, which is the strategy used by companies like Cobalt Networks. As mentioned before, I am a big believer in appliance computing, with devices tuned for specific tasks. Given this model, low cost and simplicity are key. Tuning Linux for specific tasks addressed by an appliance is relatively easy, and costs are minimized.

Third is an expertise, or service-based model. Linuxcare, a privately held firm, is the best known company that uses this model, although Andover.Net, the developers of key Linux Web sites and expertise can also be included in this camp as well. While the idea of selling expertise or services for a fee (or for advertising revenue) is the "cleanest" model in terms of building barriers to entry, it is also the most complex--supporting multiple hardware systems with multiple versions of Linux is no mean feat.

Fourth and final is what I call the system model. VA Linux is a good example of a company in this space, following a traditional computer systems model (like Dell or Sun). It sells complete, configured servers that are pre-loaded and optimized for the Linux OS.

Successful companies that follow this model can boast Linux expertise that can match or exceed the capabilities of distribution and service companies (as they will have to provide support themselves) and offer a broader product set than the appliance companies..

It's too early to tell which model will emerge as the best fit for the growing market. I think there will be winners in each category, so I believe it would be wise to invest in a basket of names. I'd expect to see Linux vendors proliferate, but beware--it's easy for a company to call itself a Linux company even when it only supports the Linux OS. So for now, stick to those pure-play Linux investments.

The preceding comments should not be considered a recommendation concerning the purchase or sale of any securities mentioned therein.