Sun's open source experiment

Sun’s acquisition of MySQL has been widely applauded. It is easy to see why. MySQL has a huge installed base, thanks to its status as the M in Lamp (Linux, Apache, MySQL, PHP), the platform underneath countless web applications. Also, Sun has long been in need of a database server to compete with Oracle. There is obvious opportunity for Sun in selling its hardware to MySQL users, and in selling MySQL to its hardware customers.

Closer scrutiny raises awkward questions. MySQL’s success is built on its excellent performance, its wide availability for free, and its tight integration with PHP. The original MySQL database engine was admired for its speed, but scorned for the lack of essential features like transactions, which is why the underlying storage engine is usually switched to

InnoDB, another open source product, which, as it happens, belongs to Oracle. That may be one of the reasons that Sun has hitherto backed a different open source database, PostgreSQL.

Oracle’s potentially awkward involvement is mitigated by an engine in development called Falcon. Sun must get it right, or the acquisition will make no sense.
Another problem for Sun is that most of MySQL’s users are individuals or small organisations who like the fact that it is small or free. These customers are not a big win.

An additional intriguing issue is that Sun is synonymous with Java. Will Sun try to migrate Lamp users to SAMJ ­ Solaris, Apache, MySQL and Java? Sun has a tricky task ahead when it comes to positioning these two software stacks.

Despite these niggles, the MySQL acquisition still has plenty of potential value for Sun, even if it turns out to be as much about brand as technology. It is interesting to ask why an enterprise company like Sun is paying big money for an open source product that began as one person’s fix for the inadequacies of another open source product, mSQL. The Lamp stack boomed because despite its humble origins it was more fit for purpose than over-complex, resource-hungry alternatives from companies such as Sun and Oracle. Part of the appeal of Lamp is that it is lightweight and nimble compared with Java.

All of this leads to several conclusions. First, MySQL now has the backing of a major company, making it more attractive to enterprises. Second, we must hope that Sun handles MySQL with a light touch, as integrating it too tightly with the Java platform could damage its appeal. Third, this acquisition demonstrates not only the value of open source, but also the value of lightweight, agile software. If Sun preserves both characteristics, then this is a great deal for its customers.