MySQL getting too big for its corporate britches?

For anyone interested in seeing just how different and game-changing open source can be, there's really no need to look beyond MySQL, the open-source database leader. Jeremy Zawodny, formerly of Yahoo, and now of Craigslist, takes a hard look at the changing face of MySQL, reaching some surprising conclusions about MySQL in the process:

Nowadays MySQL has a much slower release cycle than it used to. It's still available in "commercial" and free ("community") releases. There's still a company behind it--a much larger one in fact. But one that also has a vested interest in showing how it works better on their storage appliances or 256 "core" computers and whatnot...

Meanwhile, all the cutting edge stuff (at least from the point of view of scaling) is happening outside Sun/MySQL and being integrated by OurDelta and even Drizzle.

Zawodny details the importance of these forks to MySQL ("The single most interesting and surprising thing to me is both the number and necessity of third-party patches for enhancing various aspects of MySQL and InnoDB"), and it's here that one sees the strength of the open-source model, but also the potential fragility of open source as a business, as I've written before. These forks provide a robust MySQL database...for free.

This is good, right? Well, it is, but perhaps not if you're MySQL (or, rather, Sun), the company. For all the benefits such forks and additions provide to MySQL, they absolutely depend on Sun doing the core development on the MySQL database, core development which becomes ever more difficult to fund if such peripheral projects siphon away Sun's return on the MySQL investment.

It would seem to me that the best way for this vibrant community around MySQL to become good for the corporate MySQL would be for the community to become so active and diverse that the MySQL database begs for standardization at the core again. Sun can provide that, making enterprise customers happy and, in turn, making Sun happy.

One thing is clear: Sun needs to immediately start releasing its own "fork" of the MySQL database, one that is tuned to enterprise requirements, and one that includes functionality/tools that customers can't find elsewhere. If it's fair for Drizzle, OurDelta, Percona, etc. to enhance and extend the MySQL experience, then it's fair that Sun do this, too. Only as Sun creates differentiated value will it ensure an ongoing, rising revenue stream that will enable it to fund MySQL development, development upon which these forks critically depend.

[astamer]

The popularity of open source database software such as MySQL and PostgreSQL notwithstanding, its reputation as a cost-saving strategy did not fully hold up in Aberdeen Group?s recent benchmark study on Protecting the Database (November 2008). In the areas of security, compliance-related audit deficiencies and ongoing operational costs, the performance reported by current users of open source database software was below that of the Industry Average.

The following links provide complimentary access to the research:

· Open Source Database Solutions: When Low Cost / No Cost, Isn?t - http://aberdeen.com/summary/report/sector_insights/5750-SI-open-source-database.asp

· Protecting the Database: When (Most of) the Eggs are in One Basket - http://www.aberdeen.com/summary/report/benchmark/5302-RA-protecting-database-security.asp

Andrew Stamer, Research Associate, Aberdeen Group

[mdcallag]

Matt,
Are you calling for an open or closed source fork? Why can't MySQL monetize the forks and patches done by the community? They have the expertise and the brand. What keeps them from selling support on this and from providing binaries with community patches to MySQL Enterprise customers? Their support is excellent and I can't imagine customers not wanting it for code produced by others.

[Matt Asay]

I'm not really calling for a fork to the core, but rather to more additions beyond the core that Sun can provide to enhance the MySQL experience for paying customers. I worry that the community forks are diverting revenue from Sun, making it harder for the company to invest in MySQL core development. Why can't Sun play this same game, but to the community's betterment?

[jrepenning]

If Sun forks, who would be left on the trunk? If Sun forks, who will deliver their core changes back to the trunk? I think Sun needs get some skin in the actual open-source, communitarian part. Whether they build their proprietary value on top of that, or in a fork, can't be clear until they clearly establish custodianship of the community.

[emacsuser]

A better solution is for Sun to repackage the improved MySQL bits from Drizzle, OurDelta, Perconaand and charge for support and upgrades ..

[Matt Asay]

This would be better if enough people paid for these, but I can tell you from 10 years of experience selling support for open source that it's a very weak business model. No one makes serious money doing that, including Red Hat.

[brian_aker]

Hi!

You say "it's fair for Drizzle ... to enhance and extend the MySQL experience, then it's fair that Sun do" like Drizzle is something that Sun is not involved in. We have seven people internally now who work on Drizzle for their job (which doesn't include the others who work on it in their spare time but are still Sun employees). Sure, Drizzle has a lot of folks who do not work at Sun who contribute, but Sun is still the largest single contributor.

Cheers,
-Brian

[Matt Asay]

Yes, Brian, but that's not being entirely forthright. It was not started as a Sun-blessed fork, and I'm sure that the business executives would have preferred to hear in some other way than through blogs. :-)

[malcolmspence]

the problem is that Sun spent $1BN on MySQL and so now they have to figure out a way to recover that cost, or as they say, monetize the product.

Red Hat has the same problem with JBoss. They all have to fork the code, add value to differentiate themselves and the original idea of open source (lots of users adding their feature and fixes) goes out the window.

Malcolm in St Louis

[arjenlentz]

OurDelta (http://ourdelta.org) has most accurately been described as a "distro for MySQL", similar to how Red Hat and Ubuntu create distros of Linux. After all, we know that a tarball does not make a product.

[Matt Asay]

Agreed on OurDelta. In my effort to be concise, I threw that in as a fork, but it's not. It's a fine distinction, but I do think it's different.

[Travis5650]

Forks are inherently expensive things for an infrastructure software provider to do, open source or otherwise. Especially when you have lots of customers in production. Customers want reliable bits that have been tested and certified by a professional QA entity. Customers pay for support, QA, and a reliable upgrade path forward. Forking would only weaken Sun's ability to ensure this for customers, unless of course they plan to invest wildly in QA infrastructure and practices(but Monty's recent comments didn't exactly describe this....yet). Unless Sun has a darn good strategy for monetizing, spending more for what they already have and potentially destabilizing their ISV and "to be upgraded" customers, would be a bad decision.

[ZUrlocker]

Let me add to Brian's comment and be clear: Drizzle is a Sun project. It involves folks outside of Sun also though its still a bit of an experiment.

--Zack