When you see every Web site as a database, rather than a flat HTML file, all sorts of things become possible. Portable, controllable identity becomes possible. Personalization becomes possible. Customization becomes possible.
In retrospect Amazon.Com was one of the first true Web 2.0 sites. It was bulit as a database from the ground up, a decade ago. It was very much ahead of its time.
What does this have to do with open source?
Open source allows for easy interoperability among database-based modules. When you can see the code, you can change the code, and make things work. With open source, everyone in the community is supposed to be helping push the codebase forward. (That's not the way things really work, but that's the idea.)
And it's against this backdrop that we have the Oracle purchase of Innobase. Oracle is the Microsoft of the database world. They have been rolling-up that world, looking to create a monopoly enterprise customers must deal with, and acquire the accompanying monopoly profits. To have them controlling the chief engine for writing to a mySQL database is, at a minimum, disquieting.
That's because, in the open source world, mySQL is an essential database. Anyone with an open source Web 2.0 project has likely invested a lot in mySQL already. Switching would be costly, prohibitive for most.
The mySQL folks reacted cautiously to the Oracle announcement. It was called an endorsement of the GPL. But the press release made careful mention that there are several engines -- MyISAM, Memory, Merge, and Cluster as well as InnoDB -- for moving data in-and-out of mySQL's Pluggable Storage Engine Architecture.
Cut through the PR-speak and it reads to me like concern, caution, watch it. Oracle could always buy mySQL, too.