Why Sun bought MySQL
January 18th, 2008
Great commentaries about the acquisition can be spotted around already, but there is something I feel it has been missing in the move.
To me, one of MySQL’s biggest achievement is not their codebase or their community management (although impressive), but the invention and consolidation of the “unlock the GPL” business model: others had tried before to mix GPL with a reasonable revenue model (Cygnus comes to mind), but MySQL was the first that actually managed to sound reasonable and solid both to the eyes of the free software and open source world and to the eyes of investors and more traditional CTOs.
In fact, Sun’s new management creed has been basically a carbon-copy of MySQL’s plan: use the GPL as a flag for freedom hugging RMS in one hand, while forcing contributions’ copyrights to be donated to the for-profit entity on the other.
Most individual contributors don’t realize the difference between licensing their code under the GPL or giving their copyright to another entity that licenses that code under the GPL. For them, the result is the same: their patches get in the codebase (so they don’t have to keep patching the code at every new release, which is a massive maintenance cost), their name (and potentially their company’s) is recognized (ego massage is always good, and shines on their google ego search too… which is nowadays worth more than any CV) and their inalienable copy-left rights remain protected.
The difference is that if a thousands contributors each own their own copyright on the patches and all licensed them to you using a particular license GPL, you’re stuck forever with that license, unless you go knocking at every single one of those contributions’ rightful owners (most of which would be small private companies that are now dissolved or bought by other entities and really hard to track down) and convince them or their current managers to sign a waiver or a re-licensing agreement.
Which is precisely the situation where JBoss was for many of their codebases…. next time you hear Mr. Fleury telling you that he undersold, think about what RedHat really bought.
Sun, unlike RedHat, got it: their entire open source strategy is basically the same one that MySQL adopted; the GPL forces them to provide the codebase now and tomorrow under the GPL and this makes RMS and the rest of the open source and free software lovers happy, peaceful and collaborating, but owning the code allows the company to re-relicense the code however they please, for a fee, of course, to anybody that doesn’t like or fears the GPL’s terms.
It’s a really clever scheme: those who like the GPL have what they want for free and with the freedoms they care about, those who don’t get what they want for a fee (including the freedoms they care about… different than the GPL’s).
There is a problem with this, though: nobody knows if it really works.
I bet Sun has been trying to preach this revolutionary new way of ‘milking value’ out of this “age of participation” (as they call it), but they have not proven to their investors and stock owners that it actually works to increase revenue and/or reduce costs and/or increase the stock value (which is what they ultimately care about).
One of the necessary conditions for this scheme to work is that you have to convince people that they are not actually working for you for free, but that they are in fact getting something in return and something that compensates the bad feeling of giving away all rights on your work.
MySQL is one of the few companies that managed to be consistently growing, profitable and maintain a healthy and positive perception with their user and development communities under this “unlock the GPL” model. It’s all but an easy thing to do!
By associating the MySQL brand with their own, not only they will have access to their know-how and experience in managing such dynamics, but will also have a premium story to tell to their investors and stock owners that might end up considering this new management direction to be smart and farsighted and not just a blatant and late open source “me too”-ism.
So, first Sun quietly erases their stock history by changing their stock code-sign (no trace left in the NASDAQ database of the “5$ to 80$ back to 5$” swing that scared away any investor wise enough to click on a 5 years chart); then they buy into MySQL’s “unlock the GPL” business model and paint it all over their crown jewels (including their hardware silicon designs!) and roadblock any previous plan, even at the expense of hurting their existing relationships with other open source institutions. Finally, they acquire the original creators and best executors of the business model to both validate their model and to gain more experience to fix and tune-up their own execution problems.
I suspect that it really didn’t matter that MySQL made databases: I really don’t buy into the idea of owning a chunk of the open source stack of so many web sites gives you that much competitive advantage (other than market and PR buzz)… I think that framing the acquisition in terms of validation of Sun’s new business model dynamics makes much more sense.
Whether or not it will be worth the 800M$ cash they paid for it, it’s another story entirely.