Oracle, Open Source and Rumours of War
Oracle has acquired SleepyCat Software and is rumoured to be in talks with Zend and JBoss to round out the slate. Here's the (always) brilliant Redmonk commentary from Stephen O'Grady. Mike Olsen's commentary is obviously relevant, as the CEO inside the deal in as much as he can talk about it. The press ran amok with it, with too many articles to helpfully list here. The fear and uncertainty commentary (ZDnet, Newsforge) is unhelpful.
Here's an extra few observations to build upon Stephen's commentary:
- SleepyCat's dual license is "if you're
free-and-open-source-software you can use our software for free and if
you're charging for a license we want our share." This is the same as MySQL world the GPL as the public license, and driving their
OEM embedded business through a different license. Here's the exposure:
while MySQL can continue to drive their "free software" business
without Oracle's permission regardless of Oracle's position on future
licensing, their embedded business presumably depends upon the
ability to OEM license SleepyCat.
I know the MySQL tech leadership and business leadership well enough to
know they would have been thinking about this for some time. While I'm
sure Marten is watching the competition closely, my bet is this was
expected and MySQL has a plan. So Marten's ever so quotable "It would be like trying to kill a dolphin by drinking the ocean" isn't bravado — it's just everyday business in a competitive world.
- Culture comes from the top. If there's a strong executive sponsor at Oracle that wants to get engaged
with the open source world (as there was at Novell with executives like Chris Stone
and Al Nugent), then the injection of the DNA may actually take hold.
Miguel de Icaza and Nat Friedman were the Ximian DNA, but without strong sponsorship from the executive levels, that DNA would never have taken. Stone and Nugent were their long enough that the Ximian team could get
things started, and the SuSE team probably had it much easier.
- The Zend rumour (if true) is actually pretty simple. They're the PHP developer tools company. And while Oracle would love everyone to be writing big expensive applications in Java with a long entrenched life time against the Oracle database engine, they have probably discovered that developers script a ton of Oracle access in PHP. Make those developers happy. They're the ones that will make the move to MySQL, bringing it through the back door. If IBM is sleeping with the premier tools company around PHP, buy it outright, and break up that play at the same time.
- The rumoured JBoss acquisition has two equally compelling features, depending upon whether you're following the (historical) SAP playbook or the IBM one.
- First, the IBM game. Seven years ago (maybe eight), IBM showed up in the Apache community. They slowly began to play by the rules and earn a reputation. At the heart of their Websphere engine was an HTTP server, which is pretty fundamental. Why build your own when the Apache license allows you to embed the public one. And to be clear, the economics says you want to amortize as much of the dev/test/support/maint as you can of the complement called an HTTP server. And when they began SELLING Websphere on top of Apache, then everyone thought the open source community was going to freak. And it didn't because within the Apache world IBM was playing by the rules — making sound contributions, etc. So Oracle could be getting rid of the overhead of an Oracle complement that's costing too much. Acquiring it makes more sense than leaving it to the vagaries of IBM after the Gluecode acquisition.
- The SAP playbook says that to drive sales into the mid-tier (because you already own the Fortune 1000) you have to give the customer the complement (SAPDB, now MaxDB from MySQL) to sell them the core (R3). The rumoured Oracle JBoss acquisition may be a play to give the customer a less expensive app server to drive the core revenue around the database.
It's just business. And who knows, Oracle may even be waking up to the business tactics of open source software as IBM did 7-8 years ago, and will begin to contribute appropriate to their own lines of business and to the benefit of their customer engagements.