MySQL to Offer Proprietary Add-Ons
When Sun purchased MySQL earlier this year, reactions from the open-source community were generally positive, if surprised. MySQL was a pioneer in using a dual-licensing model for open-source software, and demonstrated that it could simultaneously profit and remain active in the open-source world. Sun's acquisition vindicated MySQL's business model, and seemed to indicate that MySQL would continue to reap the financial benefits of commercial software, along with the many other benefits of open source.
Earlier this week, however, MySQL dropped a bombshell: Future versions of MySQL Enterprise -- which until now, was simply a repackaged and supported version of the open-source distribution -- will include software that is unavailable in source form or in the open-source distribution.
Now, this won't happen for a while, and is apparently only relevant to features that are of interest to the largest enterprises, such as hot backups. These large companies are presumably paying for MySQL support contracts anyway, and will not feel the difference. So it isn't clear just how many companies need this feature but are unable to pay for it.
Regardless, there has been a great deal of anger toward MySQL -- and now Sun -- over this change of policy, as we can see from Jeremy Cole's blog posting from earlier this week. In follow-up comments to his posting, many MySQL users indicated that the announcement alone has convinced them to look at alternatives, such as PostgreSQL. (Never mind the fact that migrating from one database platform to another can be quite tricky, even when both databases are open source.) Others said that this vindicates the suspicions that they had when Sun bought MySQL.
One interesting argument was that the open-source community has performed a great deal of real-world testing for MySQL over the years, and that releasing software to enterprise customers without going through such community testing is tantamount to using these paying customers as beta testers. In other words: The open-source community performs a valuable testing service, and MySQL products without the community will be fairly buggy. While there is something to this argument, I do have to wonder how many of these small, non-paying customers would have tested MySQL's advanced backup systems were it included in the free version.
Perhaps the most interesting response was from Marten Mickos, formerly the CEO of MySQL, and now a vice president at Sun. He argued that MySQL is doing the same thing as EnterpriseDB and other vendors that sell proprietary extensions to open-source software. This might be true, but whereas EnterpriseDB has always positioned itself as a commercial company with strong open-source roots, MySQL was always seen as a company with one foot in each of the commercial and open-source camps. Indeed, Mickos wrote in a Slashdot posting that MySQL's profits are used to hire more developers for the open-source version of the software -- and admits that while their plans will not satisfy everyone, it's a "necessary evil" if the GPL'ed version of MySQL will continue to be developed.
Mickos has a much weaker argument when he says that "If we happen to develop a feature that we ship only to our paying subscribers, there is nothing stopping others (including yourself) from producing the same effect with GPL code." This is, of course, true for nearly any feature in MySQL; the nature of the GPL means that anyone can add features. But of course, no one other than Sun/MySQL can promise which community features will be incorporated into the official distribution.
It also remains to be seen whether the to-be-added commercial software will eventually be rolled into the GPL version, or what other functions will be released as commercial software.
While this decision appears to be the right one for MySQL as a business, it remains to be seen whether it is also right for MySQL as an open-source product, with a huge and enthusiastic community.
Does MySQL's announcement change your attitude toward Sun/MySQL as a company? Are you less likely to use MySQL as a result?