MySQL, Oracle, and the EC

[unixronin]

Monty Widenius, the creator of MySQL, makes this plea to help save MySQL from falling under the control of Oracle, an event which is looking increasingly bad for MySQL should it come to pass.

My letter to the EC:

I am opposed to the acquisition of Sun Microsystems by Oracle.  This is partially, but not entirely, because Sun currently owns MySQL, one of the two preeminent open-source database engines, and I firmly believe that Oracle gaining control of MySQL would be a conflict of interest that would threaten the future of MySQL.  MySQL has become a very capable database which threatens the low end of Oracle's market for its eponymous flagship product.  I question Oracle's dedication to maintaining the Solaris operating system, but more importantly, I suspect that one of Oracle's major motivations for acquiring Sun may be to eliminate MySQL as a competitor for the low end of its business.

Oracle's record following its acquisition of InnoDB show the likely future of an acquisition of MySQL by Oracle:

• InnoDB bugs have been fixed by Oracle only when required by contractual obligation

• New features announced before the acquisition have been delayed by as much as three years

• The open-source community has been excluded from continuing InnoDB development

• Patches that would have improved InnoDB performance have been neglected

• Oracle proprietized InnoDB into a closed-source product, InnoDB+, forcing Sun to fork InnoDB just to be able to apply performance patches and bug fixes

MySQL's fate is likely to be no better.  The probable future of MySQL if Oracle is allowed to buy Sun is that it will either be proprietized into a closed-source "Oracle Lite" commercial product like InnoDB+, or killed altogether.  For Oracle to gain control of MySQL would allow Oracle to strengthen its position in the database market into a near monopoly, leaving PostgreSQL as the only remaining enterprise-capable open-source database engine.  Thus I see an acquisition of Sun by Oracle as directly anti-competitive.  My concerns over this possibility are only made worse by Oracle's recent "astroturfing" among its large corporate customers, who do not use MySQL and would be unaffected by the acquisition, urging them to send letters in complete support of the acquisition.

It is my feeling that if Oracle is allowed to acquire Sun Microsystems, it should only be upon condition either — at minimum — that Oracle make strict guarantees to maintain the open-source status of both MySQL and OpenSolaris, keep them in active development, and generally not treat them as "red-headed stepchildren"; or, better, Oracle should be be required to divest itself of MySQL — NOT terminate MySQL — and transfer it to another entity that will maintain it properly and maintain its open-source status, as well as guaranteeing the future of the OpenSolaris foundation.

Comments

[ithildae]

Monty is hardly an impartial bystander in this. His goal is to reengage the licensing engine that he sold to Sun. There is some speculation that he wants to un-GPL the code so that he can generate greater income from the proprietary extensions his company can create. (The speculation is covered on Groklaw, quoting from Florian Muller's letters to the EC.)

I think the Oracle purchase is a good fit. MySQL does not compete with OracleDB. The two products fill entirely different needs in the IT infrastructure. MySQL competes more directly with SQLServer, a Micro$oft product. Larry would love to eat some of Bill's billions. I don't see Oracle trying to kill MySQL.

The EC blocking the purchase makes little sense to me. True, there are some Europeans that desperately want to regain control of MySQL, and lock in that revenue stream, but they sold those rights a few years back, for lots of money. Micro$oft is a huge opponent of the purchase, but they have a vulnerable product that a real database company could seriously damage. SAP is also an opponent to the deal, no idea why except that they are a Micro$oft partner.

I am not enthused with Oracle running an important element of the LAMP stack, but I am really not happy with what I have heard from Widenius and Muller about what MontyAB wants to do with it. Just because it is not an ideal choice does not mean it is not the best choice.