MySQL, JBoss most likely on 'the block'

Gartner Symposium/ITxpo, San Francisco, CA -  Beyond their similar roles as open source poster children,   MySQL and JBoss have two other things in common.  First, neither company wants to be acquired (both would prefer to go public).  In an interview last week regarding IBM's acquisition of open source J2EE-provider Gluecode, JBoss CEO Marc Fleury made it very clear that JBoss isn't for sale.  Here at Gartner Symposium/ITxpo, during an after hours conversation with MySQL vice president of marketing Zach Urlocker, it became pretty clear that MySQL's preference is to go solo as well.  But just because neither company has stuck a "For Sale" sign in the ground outside their offices (or posted an auction on eBay), doesn't mean they're not for sale.   Everyone has a price, and the more I think about it the more I realize that the two companies are currently irresistible acquisition targets. The question is: Who will acquire them?

Here's the logic.  For vendors, particularly software vendors, it's impossible to stay in the market without some sort of open source story.   Just look around the industry at companies like IBM, Microsoft, BEA, Sun, Computer Associates and Novell. They're the biggest companies in the industry and all of them are betting on open source in some way, shape or form.  Moving forward, this will be standard operating procedure for all software companies.  

But, being the in-vogue thing to do is actually not the big reason JBoss and MySQL will get acquired.  The reason they'll get acquired -- a reason that many companies get acquired -- is market share.  JBoss, for example, reports that it has 34 percent of the J2EE market, according to one study.  According to at least one summary of IDC's 2003 market share findings, MySQL has a negligible one-tenth of 1 percent of RDBMS market share.  But neither of these numbers tells the whole truth. That's because market share is based on revenues instead of the installed base; and when it comes to actual installed base, very few research outfits will venture a guess on market share since the software is freely downloadable and/or comes built-in to certain disributions of Linux.  There's no way to really tell who is using it and who isn't.  All I can say is that, over the past five years, the majority of organizations that I've come into contact with, particuarly start-ups, are running mission critical applications on MySQL. MySQL has way way more than one-tenth of 1 percent of the installed base.

Last week, IBM broke ranks and bought Gluecode.  What made this deal unusual was the fact that Big Blue already sells a commercially available version (Websphere) of what Gluecode offers as open source (a J2EE server).  What's even more perplexing is that, as long as the company was going to do something like that, then why not buy JBoss? Such an acquisition could have propelled IBM so far ahead of the rest of the J2EE bunch (if JBoss' 34 percent market share stat is accurate, IBM would be near 70 percent) that it almost seems like a no-brainer.  Sure, JBoss would have cost more.  But IBM is flush with cash.  But now that IBM hasn't applied that thinking to JBoss, someone else undoubtedly will, and the same goes for MySQL.  Somewhere on some whiteboard in some M&A specialist's office, these two companies are listed.

In the J2EE and database markets, there are plenty of companies that desperately need a boost in one or both of these markets.  Oracle, for example, would love to be in the 30 percent market share range when it comes to J2EE servers.  Sun too.  Or, what about Red Hat or Novell? Which database company wouldn't want the additional market share that could be gained by acquiring MySQL?  As ludicrous as it sounds, even Microsoft, which seems to be warming up to Linux in ways that no one ever thought possible, could make a play. 

But if there's one company that's really looking to restore its lustre, that company is Computer Associates.  This is a company that -- in its previous incarnation -- would acquire other mid-life companies and milk them for all they had.  Virtually everyone I ask sees the old Computer Associates as more of an investment bank  than a technology company.  The new Computer Associates wants to be a technology company, and judging by the way it has open-sourced its Ingres database, it also wants to be an open source company.  But the Ingres move hasn't raised enough eyebrows in database land to cause any sort of CA-reviving exodus.  For that, CA will need something else.  It will need a J2EE/RDBMS one-two punch with some buzz, and it has no chance of getting there if it goes the build route instead of the buy route.  That's why my bet is on CA to acquire both. But even if CA doesn't acquire them, my sense is that someone will, and long before the premium on either goes up because of an IPO.   The games began when IBM made its play for Gluecode last week. They're far from over.