Open Season on Sun / MySQL

· by Steve · Read in about 4 min · (842 Words)

The acquisition of MySQL by Sun Microsystems for an eye-watering $1bn is old news by now, but I just thought I’d recommend listening to a special edition of Open Season which talks to the executives involved (as usual, hosted by the leaders of Alfresco and MuleSource) - if you’re at all interested in the state of ‘commercial’ open source, Open Season is always a good listen, even if most of us aren’t within a million light years of the sort of enterprise business these guys are involved in.

My take on this: it makes perfect sense. Oracle historically tried to acquire MySQL but I don’t think that was ever a good match - there’s the fact that they’re in the same general business, and thus the destiny of MySQL, if acquired by Oracle, was really to become either a ‘light’ variant of Oracle or some other subservient product, which of course would never sit well with the MySQL guys. Just look at SleepyCat, which you hardly hear anything about anymore since Oracle acquired them, their technology is buried on the Oracle website now - still heavily used maybe, but it’s lost all its identity. There’s also the fact that Oracle has always seemed like a ‘user’ of open source to me, and not so much a contributor, despite having a dedicated OSS area. Being a long-term proprietary software vendor, it’s not really in their interest for open source to do anything but prop up their own offerings, so there’s inherently some conflict there. So I see open source as being a tacked-on, subsidiary effort for Oracle, and so a headline open source project like MySQL becoming part of it never felt right.

Sun, however is rather different. Like IBM they’re historically a hardware vendor, which means inherently open source is a more natural play for them. Sure they had / have Solaris, but that’s not conflicting really - server software and operating systems are inherently interdependent, and of course Solaris is open source already. Sun have finally committed wholesale to the open source model, albeit with some initial reluctance but now they really seem to have grasped the nettle, since it really does seem to be helping them - Solaris is gaining ground again for the first time in years. Better, they don’t have a database offering yet, and no vested interest in ‘merging’ MySQL into anything else they already have, so MySQL has a good chance of making a smooth transition.

Some people might think that Red Hat would have been a better buyer, but I don’t think so. It wouldn’t be a good idea for Red Hat to gain too much of a hold on headline open source projects (they already acquired JBoss after all, and that didn’t go that well) - the more players in the market the better. And I happen to like Sun - despite their many errors over the years, Solaris is still a very good product and open sourcing it was a good move, and I still have a lot of respect for Java, because I prefer solid frameworks over the latest programmer gimmicks.

In the wider context it’s good news because all the people that read the FT and Forbes will see this $1bn figure, and perhaps it will encourage them to take open source more seriously. Databases are a largely solved problem now, very much like operating systems (well, server operating systems anyway - there’s still more that can be done on usability in desktop systems I’m sure), and I’m tired of seeing the ‘big 3′ (Oracle, SQL Server, DB2) dominating most installations. There are lots of quality open source databases out there now (MySQL, PostgreSQL, Informix, Firebird, Ingres) that aren’t considered nearly often enough in some circles. Database software needs to be looked at as mostly a commodity now - unless you have very specific needs there’s very little practical difference between any of the players’ offerings now. Some have nicer GUIs, some have specific tools for specific markets, but underneath it all they’re pretty much doing the same thing, with most of the differentiation being in marketing and product attachment (e.g. the software vendors that attach their software to SQL Server, which I find utterly reprehensible given how easy it is to write DB-independent software these days). Having used all of the ‘big 3’ as well as several open source offerings I pretty much can’t tell the difference any more, from a code perspective (beyond some SQL-standard defying specifics which tend to end up in stored procedures) - DBAs of course see more of the difference because they’re using the vendor tools more, but to be honest any DBA worth their salt can transfer their knowledge easily because all the underlying principles are exactly the same. The quicker we think of these bits of software as the ‘gravy’ which just binds the more interesting parts of our applications together, and not as something we should be paying through the nose to get access to like it was still 1985, the better.