The blogosphere is abuzz with last week’s news that Oracle has decided to re-package Red Hat’s version of Linux and sell support for it at prices substantially lower than Red Hat’s own pricing. Here’s open-source skeptic Nick Carr’s take:
Yesterday, Ellison announced that his company, Oracle, fully intends to eat the fruits of the labor of Red Hat, the leading for-profit supplier of the open-source Linux operating system. Oracle is taking the version of Linux developed by Red Hat and distributing it under its own brand, as “Unbreakable Linux.” And, in a stab at Red Hat’s very heart, Ellison claims that Oracle will substantially undercut the open-source firm’s prices for supporting the software. It seems like a claim that shouldn’t be hard to fulfill. After all, Oracle doesn’t have to pay those labor costs.
Once open source became a business, rather than a movement, the rules changed. Larry Ellison, whos’s nothing if not a non-sentimentalist, understands that, and he doesn’t particularly care what “the community” thinks. His attack on Red Hat would never be called neighborly, but it is, as Business Week’s Steve Hamm puts it, “a ruthless and brilliant act of capitalism.”
It’s also something more. It illuminates a much broader and deeper tension in the digital world, a fault line that runs not only through the software industry but through every industry whose products or services exist, or can exist, as software. The tension is between social production and the profit motive. Volunteer labor means something very different in the context of a community than it does in the context of a business. In the context of a community, it’s an expression of fellowship, of the communal value of sharing. But in the context of a business, as Ellison’s move illustrates, it’s nothing more than a cheap input. Many of the most eloquent advocates of social production would prefer it if this tension didn’t exist. But it does, and it’s important.
I don’t think Carr (or Ellison, for that matter) really understands the relationship between a company like Red Hat and an open source community like the people who develop Linux. I think there are two considerations that these guys are missing, which I’ll discuss below the fold.
In the first place, I think describing the work of a Red Hat developer as “volunteer labor” rather misses the point. Certainly, such a developer creates something of value and gives it away for free. But that doesn’t mean it’s an act of charity on Red Hat’s part.
I had a conversation with a friend of a friend this weekend who happens to work at Red Hat, and he pointed out to me that Red Hat has two huge advantages over Oracle when it comes to Linux support. In the first place, it’s very important to large customers to have a product-development roadmap. If they want a particular bug squashed or a new feature added, they want to be able to get a specific answer about when that will get done. Red Hat can offer that. Oracle (at least to the extent they’re just re-packaging Red Hat’s work) can’t.
Relatedly, if Red Hat technical support can’t fix a customer’s problem, their call will eventually get escalated to the ultimate expert–the guy who wrote the software in the first place. This is a service Oracle simply can’t offer unless they hire developers of their own. Of course, they can try to have their engineers study the code and become experts on it, but in my experience, the quickest way to truly become an expert on a code base is to be involved in developing it.
The other problem with Carr’s critique, I think, is that he’s too cavalier about dismissing the open source community. As I said before, two of Red Hat’s selling points are their ability to get bugs fixed and to put customers in touch with the guy who originally developed a piece of software. That’s just as valuable when the developer of a particular bit of code is outside of Red Hat–say, an IBM employee or a grad student someplace. Red Hat engineers who have good working relationships with those people are far more likely to be able to cajole the outside developer into putting the bug higher on his priority list. If, on the other hand, Red Hat is seen as taking the community’s work and giving nothing back, Linux developers outside of Red Hat aren’t likely to respond to Red Hat developers’ emails. That will reduce Red Hat’s influence over the evolution of the operating system, and it will reduce the value of the support services that Red Hat can offer to its customers.
Hence, being seen as a good citizen in the Linux community isn’t just a matter of warm-and-fuzzy sentimentality for Red Hat. Carr’s supposed dichotomy between “cheap inputs” on the one hand and “the communal value of sharing” on the other isn’t actually a dichotomy at all. In fact, the latter is a precondition of the former. The “cheap inputs” will dry up if Red Hat fails to observe the “communal ethic of sharing.” (Indeed, the growing popularity of Debian and Ubuntu suggests that’s already happening to some extent)
I think the fundamental problem with Carr’s critique (and Oracle’s strategy) is that he seems to regard development, marketing, and support as separate, clearly distinct functions. If that were true, then Oracle’s strategy makes a lot of sense. But in fact, development is largely a precondition to marketing and support. The company that develops a piece of software will almost always be able to offer better support than a company that merely repackages somebody else’s software. Ultimately, Red Hat sells expertise. If Oracle doesn’t do any software development themselves, they’ll have no expertise to offer, and so it’s hard to see why customers would want to go with them over Red Hat.
Incidentally, the Red Hat guy I talked to says he bought some RHAT stock on Friday after the Oracle news drove the price down.