As soon as I heard the news Monday night, the wheels began to turn for me. So many angles, so many implications. IBM acquires Gluecode. Apache. Derby. MySQL. JBoss. SMB. WebSphere. WebLogic. Open source. Services based pricing. Whatever else might be said about the deal, it can’t be said that it’s boring. While I’m not convinced that it’s the most important deal since Lotus, as one of my industry analyst colleagues in the linked story believes, it’s big. On to the Q&A.
Q: Why do this?
A: Many of the commentaries I’ve seen regarding the announcement have focused on the acquisition as a direct assault on JBoss, and with good reason. JBoss has quite a bit of momentum at the moment; at BrainShare, Novell was practically fawning on JBoss, and the HP deal from last year is apparently showing good results (coincidently, I had a call scheduled with JBoss yesterday afternoon). While I’m skeptical of some of the marketshare claims I’ve heard, I don’t question the fact that JBoss is a popular platform with a sizable installed base. That undoubtedly worries IBM; some believe dramatically so. But to me JBoss is only a part of the explanation; the bigger question is, why is JBoss so popular?
Q: Ok, I’ll bite: why is JBoss so popular?
A: Because the majority of applications developed – particularly at the departmental level IBM is targeting here – simply do not need the bells and whistles of a WebSphere or WebLogic. Nor are developers keen to hit up procurement for licenses for software they’re likely to underutilize. Far better to simply download something like JBoss – or now Geronimo – and get coding. In a piece from June of ’03, my esteemed colleague said the following in a piece entitled “Evolution and Extinction: The Application Server Market in 2003 and Beyond:”
Not every organization is a Charles Schwab or an eBay, but BEA and IBM tended to compete as if that were the norm, not the exception.
This is the problem, I believe, that is finally being addressed with the acquisition of Gluecode. When I was a systems integrator, there were multiple occasions where clients felt compelled to buy WebLogic or WebSphere, when for their particular requirements a simple Tomcat instance would have sufficed. With the rise of JBoss, some of these firms finally had an alternative that they could purchase support for. So while this acquisition is no doubt aimed at blunting the inroads JBoss has made, in reality its about the underlying needs that drove customers to JBoss.
Q: The bottom line, though, is that this is a deal about open source, right?
A: Well, the product itself is open source, so yes that’s clearly involved. But I’m more aligned with views like this one, which attribute the primary motivations as basic economics: a.) using Apache as a new channel, and b.) allowing IBM to address more adeptly down-market needs. Inasmuch as open source supports those two goals, it’s about open source. But it’s channel first, open source and community second.
Q: Ok, I get the need for something like Gluecode, but why Gluecode?
A: Well, as the JBoss folks will tell you the commits to Geronimo (the project Gluecode is based on) don’t compare currently to JBoss’ offering, so the answer is probably not technology. Two factors then, which are closely related: the people, and the ties to Apache. Folks like Geir Magnusson don’t grow on trees, and as IBM said on yesterday’s analyst call, they have experience selling Apache based products. So as is often the case with acquisitions of small, open source technology players, I’d attribute the acquisition to so-called soft factors such as people and community.
Q: You mentioned Derby (nee Cloudscape) and MySQL? What’s the deal there?
A: Well, one underexamined aspect of the deal is going to be the persistence mechanism. MySQL, as I understand it, is the current database platform of choice for Gluecode. IBM, however, has previously donated their Cloudscape code to Apache forming the Derby project. Despite the fact that Derby and MySQL are different – though both relational – types of persistence mechanisms, you think there’ll be any desire to see Derby phased in? Does the pope wear a…well, you see where I’m going.
Q: How about the SMB aspects to this deal?
A: Well, it’s interesting. IBM recognized a few years ago – along with a host of other enterprise technology vendors (MS at one point threw $2B at SMB marketing & sales) – that the SMB market was a tremendous and mostly untapped opportunity. Express was the result; today, the Express software catalog has almost as bewildering array of product choices as its enterprise cousin. To IBM’s credit, the Express products were not merely crippled versions of their enterprise offerings; a lot of care and attention went into the design of these slimmed down offerings. A little while back, yours truly was able to install the WebSphere App Server Express and DB2 Express, and both went in with less than ten clicks.
But the fact remains that tagging Express with an SMB label is a bit misleading; Express is really about medium sized businesses, not small. The metrics vary, but the typical Express shop is manned by between one and five thousand people – Mom and Pop shops need not apply. That’s where Gluecode is interesting. Not that your five person laundromat on the corner is going to build themselves a J2EE application on an Apache based infrastructure, but a.) resellers could, b.) slightly larger shops with some latent technical expertise could too. So while this isn’t an SMB play in the way that, say, Microsoft Office is – it’s far closer to that market than WebSphere was.
Q: Who wins from this deal?
A: Well, obviously the Gluecode folks (congrats guys). Although the terms of the deal were not disclosed, they look to have made good money in a short time. IBM looks good too, given the new opportunities Gluecode affords them (particularly after Gluecode gains some new bodies and energy). Paradoxically, I even expect some short term gain for JBoss, given that IBM has essentially just validated the idea of a lightweight, open source application server.
Q: Who loses?
A: Well, the application server market just got a lot more interesting, so BEA, Oracle and Sun should all expect some blowback. Longer term, JBoss is facing a new sales challenge; they knew how to sell against webSphere, and know how to sell against Gluecode as it exists now (see here for Marc Fleury’s take on the news). But everyone remembers when WebSphere was merely a speck in WebLogic’s rearview mirror. There is also likely to be conflict in the future between Gluecode and WebSphere; while they are not directly competitive now, Linux didn’t used to be competitive with Unix. WebSphere, like all commercial products in a market with a credible open source alternative, will have to continue to innovate to stay ahead.
Q: Last question: pricing? What’s up with that?
A: This is to me perhaps the most interesting aspect of this deal. WebSphere of course is sold along traditional enterprise application pricing lines; base per processor licensing cost with fractional but ongoing support and services costs. Gluecode’s revenue, like JBoss, or MySQL or countless other open source applications is strictly support/services based. You can download the bits for free, and pay when you need a helping hand. Personally, I’ve believed for a while that the days of per processor licensing are winding down. Maybe not tomorrow, maybe not next month, and maybe not for all software, but as more customer friendly support oriented pricing becomes more commonplace, the traditional commercial products will have to adapt, or be far enough out in front feature/function-wise to justify the additional cost (and complexity). In this respect, the fact that IBM is not altering Gluecode pricing structure seems to me to be a stalking horse, if you will – an experiment. In that context, it bears watching whether the WebSphere pricing trickles down, or as I expect, some of the Gluecode pricing trickles up. Food for thought.
