No, Virginia, MySQL Has Not Gone Closed Source

There's a lot of talk this week, complaining that MySQL is moving away from its open source roots. While I make no bones about wanting to take every one of MySQL's paying customers (because after all, we have a better commercial offering ;-), I have to say in fairness that the talk is not really true or fair.

In actuality, as Dana Blankenhorn points out, MySQL started moving away from pure open source last year, when they created a split between the Community and Enterprise editions, and delivered closed-source software as pay-only services from their web site. This split recognized that they serve two different customer basis (as Zack Urlocker said just today): those who wish to pay, and those who wish not to pay. To my mind, there's nothing wrong with this hybrid approach! Indeed, it's the same approach that we've taken since our inception 3 years ago, and that we refined when we launched Postgres Plus last month.

I do believe it is fair, however, to ask for transparency and clear language. In Zack's blog, he wrote, "...just to be clear: (1) MySQL 6.0 backup capabilities are open source, (2) Add-on modules will be for paying customers." Perhaps I'm splitting hairs, but I'd rather see #2 say "commercially licensed" or "closed source." That's what it is, and there's nothing wrong with it.

PostgreSQL: More Traffic than MySQL?

MarkMail, a mailing list archive service, found that Postgres gets far more traffic in its mailing lists than MySQL. Since, January 2000, MySQL has 340,000 messages with about 3000 new messages per month. Meanwhile, Postgres has 583,000 messages in that same period and 7,000 new messages each month! You can read all of the details and see the charts here.

MySQL's Heliocentric Universe

As you’ve no doubt heard, Sun announced today that it is buying MySQL – for about $1 billion! While it’s only been a few hours since the announcement, I thought I’d put down some initial thoughts.

Market Validation
First, this transaction clearly validates the strategic – and commercial – importance of open source databases (OSDBs) like MySQL and Postgres/EnterpriseDB. It also validates the claims we’ve seen recently from key analyst firms, who have observed a significant increase in production use of OSDBs during the past two years, and who have begun to guide enterprises toward using OSDBs for a variety of production applications. Indeed, Gartner issued a report on this topic just last week.

By all reports, the $1 billion purchase price is 10-to-20 times trailing annual revenues, and the acquisition is by far the biggest ever in the open source software space. If there was doubt left in anyone’s mind of the impact of open source in the market, or of the potential value of open source companies, this puts that question to rest.

MySQL is Not Postgres!
We’re very happy for our friends at MySQL, who produce a fine database for a different set of customers and markets than does EnterpriseDB. MySQL is very strong in the website and other architecturally-simple applications, while Postgres/EnterpriseDB was designed from the ground up for enterprise-class applications, including high-volume OLTP. Truth be told, we almost never compete with each other in customer deals. Sun recognized this difference in their conference calls today, re-affirming their commitment to Postgres for those customers who require it.

A Heliocentric Universe
In their public communications today, Sun emphasized that their business model is all about customer choice. Jonathan Schwartz noted that they distribute products from competing vendors. Therefore, he posited, MySQL simply fits in neatly with all the rest. While there’s some truth to this, the fact of the matter is that MySQL’s sphere of influence is now Sun-centered. The remaining platform vendors — including IBM, Red Hat, Microsoft, Novell, Oracle, and even HP — have to be looking at MySQL in a whole new light today...one that recognizes that the database is owned by a competitor. No matter how many ways I try to look at it, MySQL is now SunDB. And what becomes of MySQL’s partners, who will now be integrated into the Sun infrastructure? At a minimum, this will cause confusion and uncertainty in the short run.

Why Not an IPO?
MySQL announced months ago that they were planning to go public, rather than be acquired. Truthfully, I’ve never believed that would happen. Their business model (1 out of every 1,000 users pays, and only a small amount at that) has gotten them nicely to somewhere in the $50-75 million. But that’s with more than a million downloads a month. Without a biiiigggg parent company and its associated distribution channels, how much bigger can they grow, and how quickly? Could they meet the requirements Wall Street places on superstar companies? I’ve never believed it would happen. I think they did the right thing by selling.

How Does This Impact EnterpriseDB?
Very positively!!!

• It is another huge validation for the open source database market, in which we operate
• We are left as the leading independent OSDB, growing by more than 250% last year
• We are friends – not competitors – of all the platform vendors, including Sun
• A valuation benchmark has been set (although there are no certainties about the future)
• Bloggers, analysts, and press are all over the story, and we’re included frequently

What do you think? I’d like to hear thoughts.

-- Andy

P.S. While Sun is clearly investing in the “M” in LAMP, I wonder how they feel about the “L”? ;-)

Windows is to Linux as MySQL is to PostgreSQL ?

I just read the first "Open Source Barometer" report from Alfresco, written by Dr. Ian Howells, their Chief Marketing Officer. The report analyzes market and usage trends in operating systems, application servers, databases, portals, and browsers. It's a terrific read for anyone in the open source world. Concise and well-organized, it presents useful information in a straightforward way.

What makes the report really stand out, though, is that it differentiates between usage scenarios. In other words, it recognizes that different products are used for (a) casual community use, (b) commercial evaluation situations, and (c) production deployment. The statistics are enlightening.

In operating systems, Windows is widely preferred over Linux for casual use, while Linux is the preferred deployment platform. Specifically:

Community Usage: 60% Windows; 35% Linux
Evaluation Usage: 42% Windows; 43% Linux
Deployment Usage:29% Windows; 52% Linux

No big surprise here. We certainly see the same trends with our customers.

But the interesting insight for me is that the same trend exists between MySQL and PostgreSQL...

Community Usage: 62% MySQL; 9% PostgreSQL
Evaluation Usage: 50% MySQL; 23% PostgreSQL
Deployment Usage:40% MySQL; 28% PostgreSQL

In other words, PostgreSQL is 1/6 of MySQL’s size for casual use. But in deployment, PostgreSQL rivals MySQL for the top spot.

Interesting stuff...