EU Oracle-Sun Investigation: Hobbyists and Hecklers at Work
On the 3rd of September the European Commission has decided to conduct a phase two inquiry, which includes forwarding surveys to Oracle’s competitors and customers as well as organizing “crowded“ private hearings.
While lobbyists of both sides are arguing to convince the EC that the deal may – or may not – limit competition in the database market, Monty Widenius asks hecklers for help to save mySQL, and Oracle makes commitments.
Florian Mueller - known EU campaigner and strategist, involved also in the software patents war – shared with me some of his current ideas, and I took the chance to speculate around the events.
Florian, before starting, do you have anything to disclose about your clients?
I got involved with this because I use MySQL in a .NET-based development project and feel affected. I contacted Monty about this process and received his support (see the announcement). In addition to Monty and I being in close contact, there are contacts with other parties opposing Oracle’s plans to acquire MySQL, but that’s the normal thing: you remember that back on software patents we also had a broadbased alliance.
EU wants to ensure that open source database alternatives stay available. Your position paper states that:
Incumbents are most likely to desire the acquisition of a disruptor that (i) has a track record of steadily climbing up the curve and (ii) begins to make its presence felt in lucrative segments.
Florian, is this the case, in your opinion?
Absolutely. I’m not allowed to talk about high-end MySQL customers who request confidentiality, but Mlex reported a couple of weeks ago that Deutsche Boerse AG, the German stock exchange, chose MySQL over Oracle. Or think of MySQL Cluster, a carrier-grade solution competing at the ultra-high end with Oracle RAC.
I believe a number of similar cases exist, but I can hardly buy the idea of MySQL being a credible challenger for Oracle, and I am not alone. Beyond any consideration about viable and sustainable business strategies involving GPL code, EnterpriseDB is the only open source credible alternative to Oracle. It is a matter of fact that MySQL never addressed Oracle’s market segment, and the idea of MySQL and Oracle competing in the same market is still a speculative hypothesis.
Florian, your paper mentions the SAP-MySQL agreement. Considering that SAP developed and made available its own database, the ‘database-tax’ seems more an internal problem/decision. Do you see it differently?
SAP’s own database was made available as an open source program (“SAP DB”) free of charge but customers continued to choose mostly Oracle and other proprietary databases.
SAP’s decision to partner with MySQL back in 2003 – namely two years before MySQL became a proper relational DBMS – and its subsequent decision to quit the partnership in 2007, apparently has little to do with the idea to replace Oracle or other proprietary SAP certified RDBMS.
SAP open source strategy is getting more and more mature, blurring the boundaries between open and proprietary systems.  Exchanging few email with Erwin Tenhumberg - Open Source Program Manager at SAP – I learned that open source integration is (rightly) matter of customer demand, as he explained talking about the open source office suite OpenOffice.org.
Office integration in an SAP context is a quite complex topic. Therefore, it is first not trivial and second not economically feasible to simple modify every single piece of office integration technology at SAP, so that it also supports OpenOffice.org and ODF in addition to Microsoft Office. Instead, we have to focus on the most important scenarios first. Otherwise SAP might waste a lot of money and engineering resources and still not make customers happy. Therefore, I’m very interested in more customer and partner feedback, so that we can spend the resource where they are needed the most. The more detailed and specific the customer feedback is, the better it is.
It all boils down to economic decisions. SAP doesn’t offer its customers a SAP open source integration using an open source RDBMS because probably it doesn’t make business sense. Not yet, at least.
Free and Fremium competitive forces are central in Florian reasoning. Since the whole paper to me seemed an advocacy for a specific business strategy (premium subscriptions). Florian, can you elaborate it further?
It’s about much more than that. MySQL’s business model has a service component, but a service business is not very scalable and is also limited in other ways. There are two key revenue streams that relate to MySQL’s ownership of the code: selling commercial licenses to other technology companies who embed MySQL as a building block into their own products and don’t want to be affected by the GPL’s “copyleft” (“share-alike”) obligation for commercial and other reasons; and premium subscriptions, with some key functionality such as the query analyzer being made available only to subscribers and being a key reason why subscribers pay a premium subscription fee that his much higher than what they would pay for a pure service.
The position paper doesn’t actually advocate any particular business strategy. It describes how MySQL’s revenue model works, especially its most scalable parts. What it advocates is to sustain that business model because otherwise there wouldn’t be a basis to finance further innovation.
MySQL never publicly disclosed data about the percentage of revenues deriving from commercial MySQL licenses’ sales, but as far as I understand from him dual licensing is less relevant today. While I have no access to this data, the EC could investigate and makes its own decisions based on the merit of real numbers, though.
MySQL doesn’t sell info-products addressing MySQL users’ needs, but any third party could create and sell similar digital assets for a fee, as well as managing forums for customers. Selling cloud-services based on MySQL à la ScaleDB – one of the innovative open source startups recently arrived on the scene – could be another option.
Possible business strategies for MySQL, as well as the Apache vs GPL debate raised by Matt Asay and more recently taken up by Greg Stein, are interesting, but I don’t want to go on analyzing any further the business side and its implications. The reason is that I am absolutely not convinced that business sustainability should be an EC concern.
Dual-licensing is both an opportunity and a cost, as nicely elaborated by Stephen O’Grady, and beyond MySQL customers alsoMySQL partners matter, as highlighted from Matt Aslett. Both points are relevant to MySQL’s sustainability, that is a separate thing from MySQL’s business sustainability, though.
Florian, what about a MySQL foundation raising capitals to mantain a free version?
There’s no reason to assume that the amount of development spending needed, which is in the tens of millions of euros per year and keeps growing, could be provided on that basis.
The European Commission now has to take a decision based on whatever Oracle proposes, which so far is simply the transaction as agreed upon between Oracle and Sun’s shareholders. The Commission can’t force anyone to do anything, it can only say Yes or No to a proposal. I doubt that Oracle can get the transaction cleared unconditionally. If Oracle at some point offers remedies (potential solutions), those will have to be evaluated and a key question will be whether commercial customers are comfortable that such proposals eliminate the concerns identified.
Theories of maybe a consortium of other companies being put together and having the funds and the expertise and talent on board to do the job are too speculative.
I trust your opinion about the amount of development spending needed, but I understand that your statement has to be read considering the actual product life cycle. Cost extimations probably were drawn thinking of MySQL as an RDBMS competing with more mature commercial products from other vendors.
Looking at MySQL just as an important part of the “open source innovation backbone“, maybe much less resources are needed to mantain it. A foundation funded by open source and proprietary vendors having an interest in its existence could be an option.
The EU in this case could play an active role, starting by gathering European companies who wish to become involved in such initiative. Let’s strive for open source sustainability, open source code production matters more than business models.
Roberto Galoppini, 
Stefano Maffulli 5:35 pm on December 17, 2009 Permalink
Setting up a non profit foundation may be an elegant solution to the problem. I think Sun should have enough experience dealing with communities (Java Community Process, for example). I believe that the non profit foundation would help also OpenOffice.org
Roberto Galoppini 5:43 pm on December 17, 2009 Permalink
Oracle or Sun maybe or maybe not interested in takin part in the MySQL foundation, it could be named FreeSQL or something like that to avoid all IP issues in case. I believe a relative high number of vendors could have an interest to support it.
OOo is different, very few players could be really interested and all in all the EU didn’t mention the office suite and without any pressure I can hardly imagine any change at Sun or Oracle.
A. Rebentisch 12:07 am on December 18, 2009 Permalink
So Monty sells MySQL to SUN and then wants it back for free when SUN is overtaken by Oracle. I find that a bit creative use of antitrust law for such objectives. The more critical technology of the merger is probably Java.
I am sure Monty could sell certification/support instead of licenses, regardless of the software “code” license. Florian’s relicensing request put the stakes a bit high.
Roberto Galoppini 1:32 pm on December 18, 2009 Permalink
Apparently Monty has been taken pretty seriously from the EU, and he is definitely one of the most important voices taking part into this antitrust investigation. So said, I believe that Monty’s concerns about the viability of his business approach shouldn’t be relevant in this case, though. What matters to me is code production sustainability, and nothing else.