Don’t look back

While I was on my way to the Open Source Goat Rodeo gathering in Utah, Bill Snyder, another blogger over at InfoWorld, posted a story called “Open Source Databases: The 97 Pound Weakling?” Despite the odd headline, Snyder’s post has some interesting factoids and opinions he’s pulled from reports by The 451 Group and Gartner.

I’m not sure how Gartner came up with 1% market share for open source databases, but it seems to me to be looking back on the legacy business of OLTP applications rather than looking forward to how things are being done in the online world. I would guess if you poll a hundred CIOs with any online presence at all, you’ll get a very different impression of open source DBMS usage.

Interestingly enough, Donald Feinberg of Gartner has said that he expects 70% of IT organizations to be using open source databases within 2 years. Heck, maybe we’re at that level already. And more broadly, Gartner has said that they expect 90% of IT organizations to use open source either directly or in embedded form by 2012. So I’m not sure how these numbers jive with what Snyder reports. (And this is not to blame Snyder; it’s not easy to come up with an apples to apples comparison of open source market share given that open source is disrupting the old model.)

Even the Independent Oracle Users Group (IOUG) reports that about a third of Oracle users have MySQL in production. Why? Because they’re increasingly using MySQL for new web-based applications. And that trend appears to be increasing steadily.

While I can’t comment on Snyder’s assessment of revenues for Ingres, I can say that the revenues reported for MySQL by the451 group are far below what we posted last year. (Unfortunately, you’ll just have to take my word for that, Sun doesn’t break out more details on revenues.) And the download volume and interest among customers has only increased so far this year.

Snyder also seems concerned that Open Source databases don’t have enough tools to support them. That seems a bit odd to me, considering how large an ecosystem there is for MySQL. Heck there are more than 50 exhibitors at the MySQL Conference in Santa Clara next week, so there must be something going on. In fact, pretty much all of the major tools including Visual Studio, Eclipse, NetBeans, Embarcadero and Quest all support MySQL. And if you’re looking for BI or reporting there are plenty of choices in both open source and traditional vendors including the likes of Pentaho, JasperSoft, BusinessObjects, Crystal Reports and so on. There are also plenty of graphical tools from MySQL, including the newest entry MySQL Workbench as well as new tools from up and comers like Webyog.

Admittedly, if Snyder had said there are fewer traditional OLTP applications that use open source databases, I would agree that’s a factor that limits adoption in the short term. It seems unlikey that users of systems like Siebel, PeopleSoft, Oracle Applications or Great Plains will be able to migrate to an open source database any time soon. But more importantly the next generation of enterprise applications (both on premise and onDemand) appear to be running MySQL almost exclusively. That includes the likes of Alfresco, Hyperic, Digium’s Asterix, Omniture, RightNow, SugarCRM, Workday, Zimbra and many others.

In Snyder’s article it sounds like no one is using open source databases for business-critical applications. Well, nobody but the likes of Alcatel-Lucent, Associated Press, BBC, Cardinal Health, Google, Hoovers, HP, Lafarge, Los Alamos Labs, Macy’s, MSNBC, Nokia-Siemens Networks, Orbitz, Nortel, Priceline, RightNow, Sabre, Scholastic, Smurfit-Stone, Sony, T-Systems, Virgin Mobile, the Weather Channel, Yahoo and Zappos among others.

Maybe these aren’t all mainstream IT companies. But my bet is that the best practices being defined by Web 2.0 companies today will become the enterprise infrastructure of tomorrow. If open source web infrastructure can support 24×7 high availability online reservation systems, then why not Enterprise applications? I think this transition is already well under way.

While the revenues of open source DBMSs are still small compared to the billions spent on legacy closed source databases, I think there’s a market disruption at hand and user adoption is the foremost indicator of this trend. And if we are headed into a recession, which way do you think that trend will go? Will IT managers keep spending $40K per cpu for closed source servers and cut projects? Or will they try to do more with less by using open source? Let me know your thoughts on this.

BTW, want to guess what database runs InfoWorld’s blog system? Hint: it ain’t closed source. In fact, a couple of years back, Chad Dickerson, InfoWorld’s former CTO wrote that “to a large degree, InfoWorld runs on the LAMP (Linux, Apache, MySQL, PHP) architecture.” My guess is that like many organizations, InfoWorld has web applications built on MySQL running side by side with their existing Oracle DBMS legacy systems. And now that InfoWorld is an online-only publication, what could be more business critical?