Fast and 'free' beats steady and paid on MySQL
MySQL, the lovable little database engine that could - for reasonable values of could - is starting to feel the pain of being an open source project distributed by a large company.
With a slower release cycle, community contributions are having a hard time making it into the mainline codebase, and an illicit market for patches and forks is emerging.
Drizzle, a slimmed-down version of MySQL started by MySQL director of architecture Brian Aker, promises the need for a database "optimized for cloud and net applications".
Translated for engineers, Drizzle will allegedly take better advantage of multi-core CPUs, so your 8-core Amazon EC2 instance will serve a web app database just a bit faster. That is, when it's production ready.
Not all contributions are so sweeping. There are so many smaller patches to MySQL and related software like InnoDB that the OurDelta project has sprung up to aggregate them all into a single build.
Running an OurDelta build in production is a bit like straddling a rocket engine that's eerily marked "use at your own risk". It's really only a last-ditch effort for solving a performance problem.
MySQL wonk Jeremy Zawodny recently attracted some publicity when he wondered out loud why all this was necessary. It appears that since being acquired by Sun Microsystems, MySQL's process has been slowed by a 30,000-person bureaucracy, and the open source community has the patience of a six year old.
Any sufficiently large open source project will have forks, so it's no surprise that it's happened to MySQL. If OurDelta or Drizzle gain significant ground over the mainline build, then MySQL will fit perfectly in Sun's target market: large companies with money to burn. The rest of us will use what comes with our Linux installation.
This is the same lesson that Debian hasn't yet learned from Ubuntu: fast and good-enough always beats slow and correct. ®
Ted Dziuba is a co-founder at Milo.com You can read his regular Reg column, Fail and You, every other Monday.