MySQL and Oracle

Monty Widenius, the creator of MySQL, needs your help to save MySQL from Oracle. In short, Sun bought MySQL. Oracle wants to buy Sun. Get the picture? Oracle has already let InnoDB languish after acquiring it. There's no doubt that they would like to do the same thing with MySQL. After all, many companies look at the price of Oracle versus the price of MySQL and Oracle loses another sale. Oracle does not want MySQL to succeed.

Frankly, this might be a strange blessing in disguise as it could increase adoption rates for PostgreSQL. That being said, I support open source software, even if it's software which represents the Platonic ideal of a "love/hate relationship."


What if it represents the Platonic ideal of a hate/hate relationship? :-)

I can't fathom any reason why people should attempt to foment government interference in a business transaction because they have some half-baked notion that it will kill their pet project. The bulk of MySQL is GPLed code, and it will survive, whether or not Oracle decides to support it or not.

There is an easy historical lesson here: Remember what happened with XFree86? Due to disagreement with new licensing terms, seemingly overnight, a fork was created, and now nearly every OS distribution ships with instead.

If it comes down to it, we can fork MySQL, too. Even mighty Oracle can't make the existing source releases disappear. So what's the big deal?

==== What is the goal of MariaDB?

To provide a community developed, stable, and always Free branch of MySQL that is, on the user level, compatible with the main version. We strive for total interoperability with our upstreams and our own community.

MariaDB will be kept up to date with the latest MySQL release from the same branch.


Monty has already forked it!

To respond to the commenter regarding "Why Bother, we can always Fork", I guess the issue is that Oracle could keep MySQL breathing just enough to suck all the oxygen out of the mysql/derivative ecosystem, which would basically see MySQL limping along just well enough to keep its dominant niche but never get good enough to challenge Oracle's bread and butter products.

There's only so many developers with free time to donate to open source, forking is always seen as the last option. Xorg was in response to what was a long term lack of direction and poor management on the part of XFree86. Literally this project was limping along for years. How long can Oracle keep MySQL limping along in a similar fashion? Quite a while is my guess.

I would say this is a good test case for open source software. Can software that is "open" survive being sold to a third party (the party in this case being SUN - let's be clear here, Oracle didn't buy MySQL - MySQL sold out to SUN, and now Oracle wants to buy SUN)? If Oracle manages to "kill" MySQL this way (which I don't see easily happen), then it can happen again. Linus might be willing to sell Linux for the right price, Larry might be willing to sell Perl, etc.

MySQL introduced the concepts of SQL and relational databases to the masses (of web developers). However, today there are more and often better (non-relational) storage alternatives for those applications.

So I see this as just another chapter in that story: SQL is returning to it origins of enterprise (and legacy) systems. And even there, the "nosql" solutions are becoming more popular every day. I'm not saying that relational databases are bad by definition, or going away any day soon. It's still very solid and powerful abstraction for many cases, but the days of "everybody needs a SQL database" are fading away, along with the once ubiquitous MySQL.

Now it would be good time to verify that Perl has good support for the non-relational storages. I'm talking about drivers etc. We got the DBI right, and it was definitely one of the "killer apps" for Perl.

@Ovid: Interesting. I didn't know about those projects, thanks.

I'm not saying the relational model (or SQL as a language) is broken and must go away, but I see that more and more of the common web applications (and that's definitely the success story for MySQL) have different requirements from those that MySQL, or any other of the mentioned SQL databases, originally solved for us, in practice.

So what has changed since? Obviously the hardware performance, but also the semantics: the common problem is no longer (for example) to centrally store, fetch and format cooking recipes from a database, but storing the sessions, environments and states of systems built in JavaScript - running multiple versions on the DOMs of thousands of different client browsers concurrently - and other dynamically typed languages on the backends.

PS. It's a different subject, but I do think we may have pushed the concept of Web too far (with the scenario described above). It's not done right and we need better abstractions, better tools, and more fundamental rethinking than just implement "nosql" or whatever is the latest meme.

Here's a blog post (in German) of a former MySQL consultant (who now employed by $WORK), pointing out that Monty has its own agenda:

He also points out (outside of the blog) that SUN has poured a lot of knowledge into MySQL 5.4 and the upcoming 5.5, and that we at $WORK see the performance enhancements Oracle poured into InnoDB while they owned it.

Leave a comment

About Ovid

user-pic Have Perl; Will Travel. Freelance Perl/Testing/Agile consultant. Photo by Warning: that site is not safe for work. The photographer is a good friend of mine, though, and it's appropriate to credit his work.