The rumors have been flying ever since the story broke in an Australian journal at the end of January. Given the colorful dynamics of personalities like Larry Ellison and Marc Fleury, you could imagine all the fun going on in the blogosphere.
The obvious question is what’s behind all the silliness, and whether the rush towards open source is yet another incarnation of irrational exuberance. Besides, who really buys support from JBoss anyway?
There’s little question that Oracle has trained its sights on open source. In the past few months, it’s acquired several open source storage engines on which open source rival MySQL relied. The result is yet another soap opera on whether little MySQL can survive as its 16-ton gorilla rival tries sucking up its oxygen.
In fact, open source is not a singular strategy. There are many kinds of open source strategies and business scenarios.
For instance, it’s not unusual for vendors to offload non-strategic, orphaned, or obsolete technologies to the open source community, just as CA did with Ingres before recently spinning it off. And it’s not unusual for vendors to open source non-strategic technologies, such as user interfaces or tools, if the result is a more accessible onramp to their core platforms. In this case, think IBM with its Eclipsed Java tooling and WebSphere.
But with the rumors flying over Oracle and JBoss, this time something’s different. Like IBM did last year with JBoss open source rival Gluecode, this is about buying real product. Whether it’s a real market is the question of the hour.
According to a recent Forbes online article, the rush to open source is little more than dot com redux, with one key distinction. Unlike dot coms, this time customers of open source are also potentially putting themselves in harm’s way if they rely on the products. The Forbes article added that open source isn’t such a great deal for vendors either, because in most cases, customers download the tools for free and don’t ante up for support.
If the upside is limited, why are adults like IBM and Oracle buying? If it’s so risky for customers, why are they downloading open source like it’s going out of style?
It first helps to understand what kind of open source we’re talking about, because there are several different go to market models.
The first is the spontaneous community, made famous by Linus Torvalds of Linux. A moral authority gently controls the chaos over a technology that, arguably, nobody owns. Excluding Red Hat, the only vendors making money are those for which Linux is not their core business.
Then there’s the foundation model, a more formalized version of the spontaneous community, where a third party non-profit like Apache, ObjectWeb, or Eclipse, which morphed into one, acts as arbiter. Similar to the community model, except that vendors forgo whatever ownership or royalty potential because they feel it can contribute to their core business.
Finally, there’s the captive model, like JBoss, where the vendor still controls all or most of the open source project, but still forgoes conventional licensing royalties. The main difference here is that the open sourced technology usually is the vendor’s core business.
Once you distinguish what kind of open source model it is, you can better understand the risks and upsides for customers and vendors alike. For the community and foundation models, risk depends on critical mass. That is, are there enough household names that support the technology to spread out the risk? This applies as long as the vendor doesn’t make the open source technology their core business. The coast should also be pretty clear for customers as well.
The captive model is where the issues get more interesting because either there’s something unique about the technology, or the vendor is in poor position to spread the risk. And if the vendor is at risk, the pain or gain is shared with customers as well.
A subset of these cases is where the technology is mature, like Postgres, but the market isn’t. In that instance, you’re relying on more typical dynamics of startup markets, where the issue is the vendor’s ability to execute. Fortunately, at least the vendor doesn’t have to add heavy product development costs to the equation.
So if there’s all this risk, what’s in it for customers? In a word: price. While high-end platforms like WebSphere or Oracle pack plenty of punch for mission critical back ends, when it comes to exposing your business to the web, you’re not going to pay premium prices. You just want a product that does its job and that’s that. And if the technology is pretty standard like Linux or J2EE there’s relatively little downside if you have to migrate.
At the end of the day, for vendors, some trends are simply too important to ignore. When you see commodity platforms like Linux, JBoss, or Ajax gain critical mass acceptance out of nowhere, you want to catch the train before it leaves the station. You want the hearts and minds of renegade developers because down the pike, you could sell plenty of razor blades. At that point, the question becomes how much is it really worth it to you for the opportunity to sell the razors?
