Gartner: Agenda behind EU's Sun-Oracle probe

The European Commission's decision to further probe Oracle's acquisition of Sun Microsystems has raised both questions and speculation.

Oracle said in April that it would acquire Sun, a server maker and software company whose assets include the open-source MySQL database. The deal has been approved by the U.S. Justice Department and by Sun's shareholders.

But the European Commission, the regulatory arm of the European Union, announced last week that it was opening an in-depth investigation into the $7.4 billion planned takeover, saying that a preliminary probe raised the specter of threats to competition in the database market. Oracle's enterprise software lineup includes the company's proprietary database applications.

The Commission has until January 19, 2010, to make a final decision on the Oracle-Sun deal.

But are the EU's regulators really concerned about the open-source future of MySQL? Or is there a larger agenda at play here?

Certainly, the in-depth investigation didn't come as a shock, says Gartner database analyst Donald Feinberg. "Nobody was surprised that this happened because this is coming from the organization that slapped Intel with a big fine and that went after Microsoft with a vengeance."

I recently spoke by phone, in separate conversations, with Feinberg and his Gartner colleague George Weiss to get their views on the Commission's action. This condensed Q&A folds together their responses.

Q: Since the EU/EC said they were further investigating the Sun-Oracle deal, there's been speculation as far as their motives. One of the EC's stated reasons is they're concerned that once Oracle has control of MySQL, there may be less incentive to further develop the product. Do you see that as the EC's motive, or do think there's more to the picture?
Feinberg: I think their statement that they're expanding their investigation proves that not only don't they understand anything about open-source software, but it proves beyond a shadow of a doubt that there's another agenda. If Oracle did what you just said--let's say they took the product, put it on the shelf, and didn't develop another line of code for it, Monty and Maria DB would become the new MySQL overnight. That's what open-source software is about. It has nothing to do with MySQL and everything to do with the fact that an open-source software product can fork anytime. If Oracle did nothing with the product or tried to destroy it, they would not be successful. You're talking about a product that has momentum in the marketplace. Now the problem is measuring that momentum in the marketplace. But the fact of the matter is they cannot destroy the product.

Q: And this is simply by virtue of the fact that it's open source?
Feinberg: Exactly. If Oracle decided to stop developing the Oracle DBMS, that would be the end of it because nobody else has the source code. Nobody else could do it. But the beauty of an open-source project is it's open source. Anybody can take it and move in the direction they want to move in. If you want to have an open-source product like MySQL, and Oracle is not going to nurture it and develop it and enhance it, then it will fork off and go somewhere else. So the whole premise the EC has, that Oracle could damage MySQL under the assumption that it's the leader in open-source DBMS, is wrong. Oracle will not do that. That would be a really stupid move, and Oracle's not stupid. Clearly the EC is using that for their agenda to look for ways to stop this deal. They're looking for ways to stop this deal because of other European companies who are putting pressure on the EC. The Commission is doing this on the part of the EU, and everybody else out there, including IBM and [Hewlett-Packard].

Q: What is the motivation for the EC itself?
Weiss: We have a pretty common position in Gartner that there is either a misunderstanding or lack of knowledge on the part of the EC where it feels open source can be used as a competitive threat in the market. ... That commission is there to protect the European vendors and opportunities for European common market members. There are vendors with databases that would find Oracle an intimidating presence and may be competing with Oracle not only on the database level but also on the applications level.

Feinberg: It's a political agenda. And although it's pretty strong, for a lack of better term it is the re-emergence of protectionism by a governing body of some organization. The EU is looking for how it can protect the companies in Europe.

Q: Can you give some examples of the companies in Europe?
Feinberg: SAP. Bull. Software AG. Fujitsu is Japanese, but it's big in Europe. There are many European companies that will get stiffer [competition] if Sun is invigorated, which is what would happen if Oracle buys them. Any hardware vendor that competes with Sun today is going to have a tougher time in the future, and the EU is trying to manage that, and they're wrong because that's not their job.

Q: How does the EC gather information or advice to make technology decisions like this?
Feinberg: Honestly, I don't know. You're talking about MySQL, whose revenues are about 0.4 percent of the DBMS market, and there is no other valid measure. Downloads are not a valid measure. The majority of the downloads of MySQL are done at the university and college level by students who use it every day. You can't take downloads and extrapolate that to people using the product commercially. The only measure that is accurate and valid is revenue from support. Now I will grant you that there are a lot of people using it without buying support. So that number, 0.4 percent, is incomplete. However, change it to an order of magnitude of difference, let's say, 4 percent of the market. It's still nothing. The market today is $18 billion that we're comparing this to.

Q: What about the comment from the EC that they see Oracle's databases and MySQL on a fairly even keel as if one competes with the other in the same market?
Feinberg: Not close. Oracle's more in the back office. Data warehousing. MySQL is a minuscule use of data warehousing. Oracle is one of the major vendors. So in data warehousing, they don't compete. In transactions systems, they don't compete. The only place where they overlap is in some of the non-mission-critical applications and in the Web-enabled applications where MySQL is much stronger than Oracle. By the way, my understanding is that part of the reason for these unfair-competition laws is a fear of lack of innovation. If somebody has a monopoly, they don't have to innovate. They can just collect money. Oracle released 11g R2 last Tuesday, the new release of their DBMS, and it's loaded with new functions, some of which are pretty major. So the argument about no new innovation goes away completely. The question that should be asked is not if MySQL gives them an unfair advantage in the market, but does it change the competition between Oracle and IBM, Oracle and Microsoft, Oracle and Sybase? It doesn't affect any of that. I'm not going to say that MySQL isn't going to take away some IBM, some Microsoft, and some Sybase low-end applications. But it doesn't change the major competition in the market for this stuff at all.

Q: Whatever the EC decides about this merger, MySQL is still out there. Wouldn't it pose a competitive threat either way?
Weiss: The only thing the EC may think about is "What if all the users of MySQL are suddenly invaded by an Oracle salesperson who says pay up?" Let's say Oracle creates certain anticompetitive practices in which they try to corral all of these millions of users on the MySQL database to be Oracle customers. Through open source, the only thing you pay for is the subscription support that you either want or don't want, regardless of what Oracle would like you to do. That's why you rarely see a very large monetized open-source vendor in the marketplace. Red Hat may be one of the largest. Maybe Sun could have been considered the largest because they have Open Solaris and Open Office. But in general, vendors have not been able to monetize open source into a billion-dollar business. If Sun couldn't do it, I don't know if Oracle could just because they happen to own another DBMS and then can create a multibillion-dollar business out of this at the expense of European markets.

Q: We've been hearing that IBM and HP have been using the confusion and uncertainty (about the eventual outcome of the planned Oracle-Sun merger) to their advantage. Have you been hearing the same?
Feinberg: Oh yes, absolutely, and not just IBM and HP. But all the hardware vendors that compete with Sun are trying to get Sun users to switch from their hardware platforms on the basis of the turmoil and uncertainty. I don't think it's a valid argument personally, but it's a good marketing technique. Part of the premise is that Sun is losing money like crazy. They had negative 40 percent growth last quarter. There's no question in my mind that this is going to affect that and make it worse. And at some point Oracle has to ask the question, do they still want to buy the company?

Q: Can they still back out at this point?
Feinberg: I don't know the terms of the agreement. It might cost them some money to back out. But of course, Oracle can back out. What happens if Oracle backs out is that Sun's in big trouble. ... With the negative growth they've had, with the uncertainty in the market, with the lack of confidence in Sun, if Oracle backs out of the deal, Sun's left hanging with nothing. So I think it would be devastating to Sun. And who would benefit? All the hardware companies out there, including the European hardware companies.

Q: When the EC says it's conducting an in-depth investigation, what further information can they gather that they don't already have?
Weiss: I don't know because anything that could have been documented out of the DOJ as far as due diligence could have been made available to them. There's lots of information with regard to the different forks that have occurred in MySQL. The only thing that remains is to aggregate this information and understand the working processes and business models that go on in open source and specifically in MySQL. It really would not take a whole lot of technical information; it would be more a business model understanding of how open source works.

Q: Is there anything Oracle can do or say to the EC in a situation like this?
Feinberg: The only thing that Oracle can say is: "Look at our track record. We bought InnoDB. It's an open source product. At the time it was used under a majority of MySQL installations and it still is. If we were doing something wrong with that product, people would stop using it." In fact, Oracle's enhancing it. They have not raised the price in the commercial end of it, and they're giving great support. They also bought a company called Sleepycat a couple years ago and acquired a DBMS called Berkeley DB, which is an open-source product. It's still sold by Oracle. It's enhanced by Oracle, and the customers are being supported by Oracle. So Oracle has bought two open-source products in the DBMS space and has not hurt either of them, but has in fact enhanced them. I suspect that most of the PeopleSoft people out there are getting great customer support today. Look at Oracle's track record of supporting their acquisitions going back to their first major acquisition. They bought a DBMS called RDB from DEC back in the early '90s. They still have people using RDB, and Oracle is still supporting it. It's not open source, but they didn't get rid of the product and didn't hurt the users.

Q: Do you think the EC is going to drag this out to the very end?
Feinberg: The only way they won't drag it out to the very end is if they yield to the market pressure, which they've got to be getting. I would suspect even Sun customers out there would like to see it done soon. So there's got to be market pressure for them to get it done earlier, and unless they yield to that, I think they will drag it right out as long as they possibly can. I don't think we'll see a solution to this until January or February. The only exception being if Oracle backs out, which they would probably do sooner rather than later.

Q: Do you think Oracle will stick it out?
Feinberg: My opinion is yes, I think they will. But again, you're giving into the political side of it, which is hard to predict. You have to ask yourself if Larry Ellison and Charles Phillips (Oracle's CEO and co-president, respectively) just get so fed up with this whole thing that they throw their arms up in the air and say it's not worth it. I would suspect the frustrations have to be running pretty high inside Oracle right now.

Q: If the EC doesn't approve this merger, what would mean for Sun and for Oracle?
Weiss: One scenario is that Oracle walks away from the deal. If MySQL is forced to be divested, or the EC exacts a certain toll from Oracle in penalties for monopolistic positions, Oracle will probably walk away. Sun is then thrust back again on its own looking for a suitor. We know Sun had suitors prior to Oracle. But it's unlikely anybody's going to help Sun out and acquire them like Oracle. That puts some difficult decisions for users on what to do if Oracle cuts Sun loose. Many of them are skeptical about what Oracle will do with Sun anyway. My sense from talking to the users of Sun equipment is that every day is burning a hole in their pockets in their willingness to purchase more Sun equipment. The longer it's postponed, the longer the users have to wonder whether this deal will go through and what the European Commission is going to do with the decision.