How to Make Money Off Open Source

Before they tackle the subject at their LinuxWorld panel, Database Center Site Editor Lisa Vaas got together with executives from MySQL and Sleepycat to find out how they're achieving an oxymoron: namely, making money off open-source code.

When it comes to open-source products and the viability of the companies who traffic in them, end users basically just want to know whether the company from which they get their software is going to be alive and kicking within the next few years. Or, as MySQL CEO Marten Mickos put it to me recently, "Nobody wants to buy from a loser."

Software vendors, though, increasingly are interested in open source as a business model.

Case in point: One such vendor wrote to me recently, in response to an editorial I wrote that compared two open-source databases: MySQL and PostgreSQL. He claimed to have "the fastest RDBMS on the planet"—a speed-demon database that runs "15-106 times faster than the commercial databases, with the functionality of a PostgreSQL and speed in excess of TimesTen." (TimesTen is a database and software maker that markets super-revved performance that delivers "real-time" capabilities.)


The maker of this speedy database was pondering whether to put this product out using an open-source business model. Before making a move toward open source, though, his company first needed to analyze the impact on its business model, he said. While open source is a compelling means of distribution, he said, he worried that his ideas would get ripped off by the competition.

Plenty of software makers, big and small, are thinking along similar lines. Thats one reason why people who are indeed making money off open-source software will get together at LinuxWorld to present a panel discussion, "Making Money on Open Source." On the panel will be MySQL CEO Marten Mickos, head of a company thats making money off its namesake open-source database, and Margo Seltzer, chief technology officer of Sleepycat Software, a profitable-from-Day-One open-source database maker.

First question: What about the problem of intellectual property theft?

Thats an easy one. Just because you open up your code doesnt mean copyright and patent law disappears. "Many people think the only way to protect intellectual property rights is to lock it up in a safe and apply as many patents as you can," Mickos pointed out. "Our message to the world is the opposite: Get real, it doesnt get better being locked up in a safe, and patents arent the way to protect it. You can open up your code and still be protected by copyright."

MySQL works with a dual-licensing structure: Theyre happy to give out the software free of charge to those who are happy to do the same. Theyre also happy to sell it for a fee to those people who dont want anything to do with such reciprocity. There are enough people who want to use the MySQL database without opening up their own, proprietary code to keep the companys revenues consisting mostly of license payments.

Sleepycats dual-licensing structure is similar. Sleepycats open-source database product, Berkeley DB, differentiates itself by the fact that its a library that links directly into an application. If the application into which Berkeley DB is being linked is open source, the library is open source and can be redistributed without licensing fees. If the application that links to the library is proprietary, Sleepycat retains all intellectual property rights and has the ability to grant a license so as to generate licensing revenue.

As a library, unlike other open-source products, Sleepycats Berkeley DB doesnt do anything in and of itself. Database functionality is realized only once the library is bound with an application that makes calls into the library.

According to Seltzer, the products position as a library gives the company the leverage to make a distinction between people who embed it in open-source products vs. proprietary products. "The GPL is wonderful for the open-source community, but not necessarily sustainable commercially," she said. "If youre a pure GPL or a stand-alone application, its very difficult to be open source and generate licensing revenue. The fact that were a library and you have to link with us before you do anything useful turns out to be useful in making the distinction clear."

The intricacies of open-source licensing raise far more questions than these, so consider this just a primer. To get more detailed information, Id suggest starting at the GNU GPL site. For a real education on open-source licensing and how to make money off this model, though, stop by the panel on Wednesday morning to grill Mickos and Seltzer, wholl be joined by other experts, including Stuart Cohen, CEO of Open Source Development Labs.

What other questions do you have about open-source licensing? Write me at Database Center Editor Lisa Vaas has written about enterprise applications since 1997.

Editors note: This story has been revised since its initial posting to correct erroneous references to PostgreSQL being on the LinuxWorld panel.