⇥ Some thoughts on the MySQL/Oracle merger

Oracle is being strangely mum on the Sun acquisition… because they have to. But what about MySQL?

Oracle buys Sun… but where does that leave MySQL?

The pending acquisition of Sun by Oracle is rife with possibilities—from Java to Solaris, Oracle could conceivably get their collective hands on a host of very interesting technologies that are close to the hearts of many open-source advocates. More than that, Sun owns a rather significant software patent portfolio, portions of which have been licensed to OASIS (of which Sun was a founding member)—and patents can be a dangerous thing.

In the world of PHP, however, the acquisition is particularly relevant because of MySQL. Given that MySQL is so widely used throughout the PHP market (and MySQL’s less-than-stellar record in the licensing arena), there are rampant concerns that the acquisition could spell trouble for a system that so many people rely on… because it’s free.

Over the past few months, the topic of “what is going to happen to MySQL” has cropped up in a number of conversations that I have had (most recently while I was interviewing Marcus Whitney for the php|architect podcast); people mostly seem to be concerned about Oracle’s silence on the matter, which they take as an ominous sign that things are going to take a turn for the worse.

The silence, I am afraid, is nothing ominous—Oracle is required to keep mum until it has cleared all the regulatory hurdles required to complete the acquisition, and the European Union has not yet completed its investigation in the matter. Therefore, until the merger is approved (or rejected), it’s unlikely that Oracle will make any significant announcements, or even that either Oracle or Sun employees will be allowed to discuss the matter beyond a “we don’t know what’s going to happen yet.”

But should we worry?

The conspiracy theorist in me very much wants to tell me that the way MySQL entered in this picture is highly suspiciously. First, Sun buys MySQL—in what is practically an all-cash transaction (you don’t buy things with cash when you’ve got stocks to issue). Barely a year later, Oracle turns around and buys Sun. It almost looks like the two events are neatly connected—all of a sudden, even Oracle’s acquisition of InnoDB, a strange and impenetrable decision at the time, seems to fit perfectly as a piece in a puzzle of Faustian proportions.

The cynic in me, on the other hand, is resigned to the fact that life is so much less interesting than people often want it to be and thinks that Oracle’s acquisition of MySQL is entirely incidental to Oracle’s acquisition of Sun—in fact, I’d have to say that the conspiracy theorists have things backwards and it’s far more likely that Sun purchased MySQL in a bid to make itself more attractive to a potential acquirer, which may or may not have been Oracle.

In fact, it is perfectly possible that MySQL will prove to be unwanted baggage in the acquisition—specifically because it so obviously poses questions of antitrust.

Let’s not be naïve, though

Let’s be clear on one point, though: while MySQL may be entering the picture only tangentially to the acquisition, it’s naïve to pretend that it being owned by Oracle would not be cause for concern. Carlo Piana, an Italian lawyer and open-source advocate who is advising Oracle in the EU antitrust investigation notes in his blog that the concerns surrounding the acquisition are moot because MySQL is an open-source product and, therefore,  the option of forking makes it impossible for one company to stifle it.

That, however, is a little misleading, methinks. Forking is, indeed, an option, but it’s hardly a deterrent from a business perspective. It’s one thing to use MySQL because it’s open source, and quite another to use it because it is free. Most companies don’t care that MySQL is licensed under the GPL: they care about the fact that it is inexpensive, but that there is a commercial entity behind it on which they can rely should they require assistance that goes beyond their expertise. If you take that commercial option away, or make it overly unpalatable, adopting the product becomes significantly riskier, regardless of whether it is being actively developed by a community of volunteers or not.

Even in the PHP world, your worst-case-scenario (from a cost perspective) is that you hire one of a number of companies that have firm ties within the PHP project itself and can solve problems for you that go beyond what your developers can handle. It’s a basic safety net that every good business owner will want to have at their disposal “just in case”—even more so in the case of MySQL, where the available alternatives are not as easy to justify.

Thus, the GPL does not render the issue of the acquisition moot—if anything, it places it in sharp contrast against the backdrop of Oracle’s business, whose focus is on large clients and large installations. In this picture, Oracle could conceivably stifle MySQL in such a way as to leave its customers no option but to upgrade to the company’s mainstream DBMS offerings when they need more than basic database functionality.

…or maybe not: as Carlo Piana also notes, MySQL and Oracle are two very distinct products that, at least at the present time, do not easily segue into one another; therefore, Oracle would have to implement at least some level of business continuity between the two in such a way as to make the progression from MySQL to Oracle easier.

An opportunity for some?

For sure, the current situation represents an interesting opportunity for other database vendors—Oracle has no control over the EU investigation and, conceivably, they won’t be able to implement any MySQL-related plans until that concludes one way or another. This leaves a period of uncertainty in which investing in MySQL is no longer the clear-cut advantageous business decision that it used to be; even anecdotally, a number of Blue Parabola clients over the last few months have been asking us whether adopting MySQL was a good decision, given the fact that their growth might hit an Oracle-imposed roadblock. So far, our answer is that MySQL is still a good option, since Oracle’s influence is unlikely to affect the product so quickly that its clients will be unable to react in time.

However, this doesn’t mean that the doubts aren’t there and the door isn’t open to others who might be able to gain significantly from the situation. Microsoft comes to mind as a prime candidate here, because, with a slight tweaking of its licensing terms, MSSQL could really give MySQL a significant run for its money. Despite the fact that it’s not cheap (hence the primary problem), MSSQL is a great DBMS that works equally well in small and large installations—I am pretty sure that most MySQL users would be able to switch to it with relatively little difficulty, gaining significant advantages in the process (which could as well be gained by switching to Oracle—except that Oracle has, in my experience, a much steeper learning curve).

The problem with MSSQL, I think, is that the product doesn’t have a licensing progression that suits this particular purpose—the jump from the free Express edition to a $2,500-per-processor commercial license is too steep for most businesses that currently use MySQL (and, of course, you need to remember all those Windows licenses, too). Microsoft could certainly do better—for example, by creating a unified license that has a low entry cost and progresses nicely to higher levels, or by instituting a partnership program that limits the cost of MSSQL when combined with value-added propositions like support from its partners.

Edit: fixed a small typo.

Photo credit: Oracle Tux by BDegan