Oracle's Unbreakable Linux Unleashed on an Uncaring World

Last Friday an IBM spokesman described the company's decision not to certify its products on Oracle's Unbreakable Linux. For those who may not be familiar with the Oracle initiative, it announced last October that it would offer a clone of Red Hat Enterprise Linux (RHEL), branded as Unbreakable Linux, at a lower price than Red Hat charges for RHEL.

I blogged about this initiative at the time and expressed skepticism about its prospects. I specifically addressed the issue of Oracle's challenge in getting other vendors to certify their products on Oracle's Linux:

Finally, what about the rest of the industry? Just because Oracle claims that its product is compatible with Red Hat doesn't mean that other ISVs and service providers are going to automatically support Oracle's Linux. It is a slog to get these players to certify an application, and, so far, no one has offered any reason that they would particularly want to do so on Oracle Linux. If customers demand it, they'll certify, but I don't see them leaping to take on another combinatorial QA effort to help Oracle's initiative-of-the-moment. Unlike the trade press, industry participants have long memories and jaundiced views regarding these type of announcements. I doubt any of them have formed Oracle Linux task forces this Monday morning.

If you read the statement by IBM, it essentially echoes what I noted -- if customers demand it, IBM will do it, but they're not going to go out of their way to help out Unbreakable Linux.

And if you think about it, that makes perfect sense. Obviously, there's the issue of taking on another platform certification. This is no small effort; beyond the cost of QA, there are the downstream costs of training support and sales, updating marketing materials, and so on. For a company like IBM, that's an enormous expense. So, on the face of it, the initial issue for IBM is "why make a big investment for a platform with no discernable market momentum?"

Beyond that, however, is a larger strategic issue. Oracle wants to cut Red Hat out of the picture so that it can control more of the platform software stack -- from OS through database and app server all the way up to enterprise application.

But guess what? There's a number of other companies that want to have that primary customer relationship, too, because overall account margin migrates to the entity with the primary relationship with the buyer -- everyone else gets relegated to the position of supplier. And IBM is the quintessential, prototypical account management company in high tech. You think they're going to jeopardize that by introducing an OS from a competitor that wants to own the same territory as them? Are they going to pitch their products running on an OS from a competitor that would immediately take the opportunity to insert itself into the purchasing decision to pitch its products?

While many have written off Oracle's Unbreakable Linux strategy because it hasn't got much traction so far, it's far too early to know with certainty how the story will end -- these kind of enterprise platform struggles play out over years. However, we can know for sure that no big vendor that competes with Oracle is going to break a sweat putting its products on Oracle's Linux.

Oracle's strategy has to start with pushing its own products on Unbreakable Linux as a way to get some presence in data centers. It can then reach out to second- and third-tier ISVs to motivate them to put their products on Unbreakable Linux. If Oracle is assiduous enough in this effort, it could create enough of an ecosystem around its product that end users would start demanding first-tier apps from companies like IBM be available on the platform. At a guess, however, the kind of channel partner slog to slowly build a platform ecosystem is unlikely to appeal to a company with Oracle's typical attention span.

While this story has been obsessed over as an open source subject, it's only tangentially about open source or Linux. Linux is just a tool in this vendor conflict, which is all about what company is going to have the primary customer relationship. Don't imagine it's anything else. 

As a coda to this discussion, a comment at the end of a blog posting that carried the discussion really tickled me. The commenter wrote:

Larry should stick to what he knows. He hasn’t even finished wrecking PeopleSoft yet.