Dogs and Cats living together? (JBoss and MS)

This news made the rounds - Microsoft and JBoss announced a partnership to make JBoss and Microsoft infrastructure work better together. story 1  story 2  story 3  story 4 

The Technical stuff

Interop! I love it. Since it's a new partnership, the technical details on just how all this interop will happen, are slim. But we get the idea that the focus is on Directory services, Web services, Management, and database (SQL Server). All good opportunities. And that ordering is how I would rank them in order of importance. We'll keep you apprised of progress.

The Business Side

Of all the stories examining the business angles, I think Red Herring gets it "most right." The other articles talk about a "philosophical divide", or even a religious issue. It's not. It's business. This deal defines co-opetition. Microsoft is way too big a company to try to live in a homogeneous world. Microsoft's had some solid successes and some spectacular failures over the years, and if you look carefully, you'll see the successes have been where Microsoft built solid platforms that others could benefit from. The failures are the exact opposite. Eg, Passport - too much conflict there to be an agreeable platform.

JBoss says better than 50% of their customers use Windows. So there are obvious partnering opportunities between JBoss and MS. Microsoft wants those customers to succeed, and so does JBoss. The cooperative opportunity is clear, the only interesting question the JBoss/MS announcement is, how will the pair handle the competitive aspects? But that is nothing new.

Microsoft has been pursuing this approach for years. Forever, in fact. Look at database - it's gotta be one of the biggest "add-on" software categories - you know, something that is not an operating system. In June, Gartner estimated the database market for 2004 to be about $8 billion (wow), and that better than 40% of all new relational database licenses are for Windows. That doesn't mean 40% of the $8b goes to Microsoft, it means 40% of the $8b spent, goes to database licenses that eventually get used on Windows. The $8b doesn't include the OS license, of course. I'd guess the OS license to be much much smaller - maybe 1/10 the cost of the RDBMS. It's probably less, but that's the right order of magnitude. Which means, 8 * .4 * .1 = $320M in OS revenue. Of course, not every new RDBMS license implies a new license for the underlying OS. So the .1 number is high, and the actual ratio is smaller yet. Anyway, the point is, out of that huge pile of RDBMS license money, Microsoft makes a tiny fraction.

Of course, $250M is not a tiny sum of money. But I'm talking about ratios here. It's 0.03 of the total $8b. By the way, I have to say that none of this is based on Microsoft's private information. I'm just doing back-of-the-envelope SWAG calculations based on Gartner's number, and based on my understanding of how much commercial RDBMS licenses costs in relaiton to how much Windows Server costs.

Microsoft works really hard to get that fraction, with strong partnerships with IBM, Oracle, Sybase, and other companies, to make sure their database products run great on Windows. The old win-win approach. At the same time, Microsoft sells SQL Server, which is competitive with those other database products. That competition may distract, but does not detract, from the partnering imperative for the larger RDBMS market. The success of Windows as a platform depends on it.

Microsoft takes approximately the same "platform" approach in virtually every product that could be called a success. The way Microsoft builds Windows into a large platform is by aggregating lots and lots of those segments. RDBMS is a really big one, but there are lots of others that are much smaller, where the same sort of math applies. Microsoft provides underlying platform value, and is able to service the segment in exchange for some small fraction of overall segment revenue. Lots of little win-win deals. Add em all up and you get a big number.

The JBoss/MS deal is just part of that approach. You might say, the deal is a commentary on JBoss, more than anything else. ( JBoss is now big enough that Microsoft needs to partner with them. ) But that's not really fair, because JBoss has been doing strategic deals with large partners for years now. JBoss has been bonafide for years, and this deal doesn't "make" them.

I guess the timing was just right for this one.

Anyway, I am glad to see this deal happen and I look forward to demonstrating the details of technical interop made possible via the partnership. By the way, JBoss already runs great on Windows. With SQL Server. I run that combo on my home network!