I feel like I am coming late to this topic. All the pundits have already had there say and the blogosphere has been rampant. I have been talking with a lot of MySQL folks lately, encouraging them to at least test PostgreSQL as an alternative. MySQL folks are nervous. They don't like the opportunity Oracle brings to the table. This morning I was asked quite bluntly, "From your perspective what is the future of MySQL?".

It is an interesting question because in my perspective MySQL has always been a "just good enough" second class citizen in the database world. I don't run into a lot of MySQL migrations because the Command Prompt's customers are generally already PostgreSQL folks or asking us to implement something anew. Trying to be objective I think MySQL will live on in various incarnations but I do think its glory days are over. It will be a supported (for a short time) but second class citizen from Oracle. In the Sun purchase Oracle was primarily after two things with Sun, Solaris and Java.

During Innotech yesterday I was speaking on the open source panel and one of the participants stated that they were nervous about the fact that MySQL had been bought twice in the last two years. I did mention that I didn't think MySQL was going away and that Oracle is a smart company and there is a lot of mind share with MySQL. However, Oracle is not interested in the 1000.00/yr business. That is the majority of MySQL revenue. It was estimated that MySQL AB was only doing 50M a year when they were bought by Sun. 50M a year is petty cash for Oracle. I only see two outcomes (both with variations).

Oracle will completely change the MySQL model to make it more profitable and thus alienate its majority user base (small websites) or maintain it long enough to allow MySQL to kill itself. MySQL is already killing itself through the various forks that have permeated in the last 9 months. I see a large possibility of mass migration from MySQL by non web applications. Yes there are a lot of them. Why? Because one way Oracle can make money from MySQL is to continue to charge for "linked" software against MySQL.

If you are building a web app as long as your web language is open source, you are good with the GPL. If you are building a monolithic app in say C++ you have a serious problem because the nature of the GPL guarantees that your C++ app will have to be open source (assuming you are linking to MySQL libs versus something like ODBC). The majority (by far) of the world still isn't using Open Source. MySQL does have a strong following in the appliance world and if Oracle opts to start charging for MySQL closed applications the way they charge for Oracle the appliance world is going to run, not walk to other technologies.

The obvious choice is PostgreSQL because of the BSD license and the maturity of the software. In conclusion I expect that MySQL in two years likely won't exist except on the most tertiary level. Most new projects will be developed in either PostgreSQL, Firebird or one of the forks (MariaDB, Drizzle).