Open Source in the Enterprise

By Jesse Freund

Open Source is not only a tour de force in the engineering world; it has become a cause celebre among many engineers. It seems that only its growing popularity can threaten open source's underdog appeal. Indeed, last year, research firm IDC proclaimed that Linux officially went "mainstream," and the numbers certainly back up the assertion. Revenue from Linux products (servers, PCs, and packaged software) is soaring from just over $10 million in 2003 to more than a projected $35 million in 2008, a 26% annualized growth rate.

All of which begs the question: Where does open source belong in today's enterprise? That's a tricky question. While full-featured and mature open source solutions exist to support most business functions, the answer typically involves strategic issues rather than technical hurdles. If it were just a matter of building from the ground up, it would be possible to use Linux together with the Apache Web server, JBoss' application server, and the MySQL database to create a pretty viable architecture -- what could be termed a complete open source "stack" for an Internet-based application. But few large organizations build architectures from scratch, and fewer still want to face a project that involves lengthy integration issues. There has also been debate surrounding security, and whether Linux is more or less secure than Windows. Let's examine some places where open source is making inroads into today's enterprises.

The middle

After Linux, Apache is probably the poster child for open source in the enterprise. From its beginnings as a small homegrown Web server, Apache has been embraced by corporate America as a viable business tool. Similarly, other open source projects are also starting to migrate from front-end servers through middleware and toward the back-end database. In the middle, application servers play a vital role by allowing businesses to reduce database costs. And open source application servers, like Jboss, promise even more cost savings. But selecting an application server is no light decision, and the effort of integrating one is no small feat, so for the time being at least, open source application servers face entrenched business obstacles, even if they offer enterprise-ready technology.

The back

When it comes to enterprise databases, the question isn't whether open source solutions like MySQL are ready for the enterprise. MySQL is. Many large companies have MySQL running side by side with Oracle, Microsoft, and IBM solutions. The question is: Is open source ready to be the sole database solution? In other words, can it handle all of the business-critical tasks? That's trickier. While supporters of open source point out the advantages -- the lack of licensing encumbrances, cost reductions, and relatively high stability -- detractors cite the lack of XML support, inadequate business intelligence features, and substandard auditing and compliance solutions. In the end, companies like Oracle, IBM, and Microsoft invest billions of dollars in their database products every year, an expense that open source developers simply can't match, so proprietary databases are likely to be perceived as more feature-rich and stable for the foreseeable future.

The unsung hero

Lost in all this is-it-or-isn't-it-ready talk is a place where open source is having a large impact on today's enterprise: applications libraries. Open source applications libraries -- the tools used by engineers to create components of larger applications -- are undoubtedly saving companies large sums by reducing the time it takes engineers to build internal applications. From Web services to XML parsing to email processing to GUI development, an open source tool exists to speed most aspects of application development. While difficult to measure in revenue figures, today's open source application libraries certainly reduce labor costs and speed development time.
 
Conclusion

In the end, the mere existence of an open source solution doesn't justify the high cost and higher risk of integration. And it doesn't always make business sense to throw out an existing proprietary product, even if that product doesn't deliver every desirable feature.

Jesse Freund is a Contributing Writer at Business 2.0 and a frequent contributor to Wired.