Code

Where is MySQL Going From Here?

Where is the king of open source database engines, MySQL, and where is it going? For many years during the beginning of the 21st Century, MySQL enjoyed a stable and ever growing popular place in the open source database engine space. As its core feature set, its speed and its stability grew with each release through the 3.x and 4.x cycles, MySQL became the default database engine for a majority of the world's internet applications. Founder Michael Widenius built a loyal following due to his candid and open approach to delivering bug-free software without the usual and unmerited hype that surrounds most commercial applications. He did all this with the help of a diligent and refreshingly honest core development team that residing all over the world, a virtual company in every sense.

Most of us that develop web applications have grown used to the stable and evolving presence of MySQL as a key leg in the LAMP platform. OK, maybe the jump from 3.x to 4.x was a bit rough - but the picture remained the same. I don't think it’s an understatement to say MySQL could have been used to define customer loyalty. But cracks began to appear in this rosy picture when MySQL was first sold to Sun Microsystems Inc back in the first half of 2008. No one could tell what this would mean for MySQL. Was it a bold move by SUN or the beginning of the end for MySQL's happy days of prominence. This move had even more intrigue behind because commercial database giant Oracle had already purchased Innodb (Jeremy Zawodny) in the fall of 2005. Innodb is the row-level transaction safe ACID database engine used by MySQL and a key component in its feature set, which now seemed to be shaping up as a pawn in the battle for dominance in the internet database world. Would Oracle pull the plug on Innodb and cripple MySQL's transaction support?

Initially, SUN gained MySQL's market presence and the expertise of the MySQL development team to extend it even further, with the help of SUN's commercial and marketing reach. But this scenario soured, most notably with a buggy release of MySQL 5.1. The separation and loss of MySQL core expertise was underway. By the time Oracle came along in the spring of 2009 and purchased MySQL away from SUN, virtually all of the core development team that had years of knowledge in the MySQL database system was gone. Even Monty himself had moved on. Monty had initiated another engine called MariaDB (askmonty.org) as an "open source" community model and MySQL compatible database engine.

How will all this affect the millions of websites that are built around MySQL? The jury is still out. Competition and survival will play out under Oracle and MariaDB. I expect Oracle to struggle with MySQL releases due to its lack of core competency with MySQL and over time, migrate and replace key components with Oracle processes. They just can't build or maintain competency in two completely different database systems. One or the other or both will suffer. Will Oracle introduce fees for MySQL? That fits their model. Will MariaDB take off and eventually break all ties with Oracle's MySQL. If (or when) all that happens, we had better plan on a database storage migration and database code effort for all those websites. The storage engine migration is the easy part. The coding side of the picture should be relatively straight forward, as long as you use database abstractions throughout your systems and don't have all those hard-coded mysql_xxxx calls. The future will be here before you know it. Now's the time to prepare so you aren't caught off guard.