Monty Widenius and David Axmark are typical techies. When they needed a low-cost database for high-speed data warehousing they took one look at the commercial alternatives and decided that they could do better themselves.

The result was MySQL, a slimline relational database that could be used for a wider range of applications than Axmark and Widenius originally intended. So, like good techies, they released it under the open source general public license (GPL) so that everyone could use it.

That was in 1996. Significantly, it was the first low-cost relational database that could be used as a web site back end. Although it lacks many of the features found in major commercial databases, this is also one of its selling points. It makes it quick to set up and relatively easy to use. It also offers a high level of performance on comparatively modest hardware.

Furthermore, as the first waves of Internet web site developers have become more proficient – or have graduated from college – they have taken their preference for MySQL into the workplace. The open source database is used in a number of organisations, such as the North American Space Agency (NASA) and Internet portal Yahoo!

Most significant of all, Axmark and Widenius have started to include some of the enterprise features that users have been crying out for as optional modules. These include support for transactional integrity and roll back, so that if a transaction has to be abandoned before it is completed the database is returned to its original state.

But if the software is given away for free, how will MySQL AB actually make any money? New CEO Marten Mickos aims to develop the business by selling training, support and services to MySQL users. It also sells licenses for organisations that re-sell MySQL in their products.

Sales are undisclosed, but the company claims to be profitable and boasts more than two million MySQL installations. It also received its first round of venture capital funding of €2 million in November 2001.

But despite the popularity of MySQL, analysts do not believe that it will ever challenge the database market leaders.

The lack of many features and lingering corporate suspicion of open source software means that it is unlikely to be adopted for high-end, mission critical applications where the big money for service and support is made. MySQL will therefore struggle to break out of its departmental and web site back-end niche.

In addition, there is doubt that the open source commercial model can ever be particularly profitable. "The problem is that no one has proved that a business can be profitable and remain viable long term by selling support and services around open source software," says Meta Group analyst Charles Garry.

However, if the company continues to develop new features without compromising performance – and winning new users at the same time – it will certainly cause many headaches for executives at IBM, Microsoft and Oracle, even if Oracle founder and CEO Larry Ellison still sleeps soundly.