Enterprise Unix Roundup: Oracle, Enterprise Linux Pioneer?

IBM and Oracle agreed to offer more support for Linux applications running on IBM System z mainframes. The deal doesn't extend to Oracle's Unbreakable Linux at this time -- though not from lack of interest on IBM customers' part.

Brian Proffitt

They call it Unbreakable Linux, although if the rumors are true, Oracle may need to call it something else. Unbearable? Overbearing?

The reason behind the jibe? Well, it seems Marten Mickos, CEO at MySQL AB, has let it be known that one of his biggest competitors, the aforementioned Oracle, might be planning to pull an end-run and release and support the open-source MySQL database themselves — just like Oracle is doing with Red Hat Enterprise Linux.

Mickos doesn't seem to concerned, and he's a pretty unflappable guy. And while such an attempt tends to rub my sense of fair play the wrong way, the truth is that Oracle may be starting a trend that will ultimately be good for the enterprise ... even if the company is doing it for purely selfish reasons.

Short term, MySQL isn't worried about Oracle encroaching on its turf because the company believes that such a move from Oracle would only endorse MySQL. This is probably the case. When I got over my initial reaction of Oracle = Arrogant, though, it occurred to me that this kind of play will have long-term benefits as well.

The crux of the argument is simple: Someone should have thought of doing this before.

"This" means what Oracle is doing right now — taking pre-existing free/libre/open source software (FLOSS), changing it to something it can support, and releasing it as its own product — can be done simply because it's FLOSS. There's no rule that says you can't do this. In fact, the rules encourage such sharing of innovation.

Discuss this article in the ServerWatch discussion forum

Under no circumstances should it be thought that Oracle is doing this out of any sense of altruism, and therein lies the rub. Oracle wants to knock Red Hat off the top of the enterprise Linux mountain — no ifs, ands, or buts. And, if Mickos is correct, it will want to do the same to MySQL. Even with selfish motivations, Oracle may have stumbled on the formula that only FLOSS allows.

One-stop shopping.

The idea that an enterprise customer should, in the near future, be able to go to one vendor and get an entire stack's worth of products and support from that single vendor is something people have been aiming for for years. Even in the open source community, various Linux distributions have formed partnerships and alliances with other FLOSS vendors, to get their product offerings aligned.

Oracle, through sheer chutzpah, did them all one better: It reached out and took what it wanted. In other words, it out-FLOSSed the FLOSS companies.

This is because, up until Oracle, everyone wanted to play by the old rules. The ones where you made business deals and got permission to use code. Larry Ellison and crew figured out something that even the new FLOSS companies weren't willing to: They don't have to play by the old rules.

The old rules work for proprietary software. The new rules are where you take what code you need, build a product, and ship it (ideally with good support). What's interesting is that Linux distribution companies have been doing this all along. A distribution is just a collection of many different FLOSS apps, all wrapped up in a neat package. When these distros were put together, there weren't a lot of business deals. If a distro needed a text editor, it would just build packages for whatever editors it wanted: emacs, vi, gedit, kate ... No one has a meeting or issues a press release when GNU GCC is included in a distribution.

Oracle seems to have figured out how to apply this take-what-you-need model to the stack — that catchy name for a collection of servers designed to interface the customer with the data in whatever particular manner the customer needs to interface. Instead of an application distribution, Oracle is building a stack distribution — without any "overhead" of a partnership arrangement.

What may be embarrassing for the FLOSS folks is they didn't figure this out on their own. But before heaping too much praise on Oracle, there is something to be said for politeness and fair play.

Oracle may have made the next intuitive leap in the evolution of the software business, but it must also to be able to provide some really solid support to customers to make this work. They way it is building its stack distribution has precluded getting much help from the actual product sources. True, it has avoided the need to share revenue with any partners, but it has also burned a few bridges with the FLOSS community. This may prove key in the months to come.

Whether it turns out to be good or ill, however, Oracle may have started something big within the enterprise. Sooner or later, another vendor is going to figure out how to put stacks together without a lot of fuss and without ticking everyone else off. These kinds of integrated solutions will be a huge draw in the enterprise space.

More to the point, such customized solutions would be available only from FLOSS vendors, because only they can freely share their code. Such on-the-fly solutions would be impossible for proprietary vendors because of the sheer hassle of getting closed-source applications to efficiently work together — not to mention all the potential license headaches.

In its effort to dominate, Oracle may have paved the way for FLOSS to really succeed in the enterprise.

Brian Proffitt is managing editor of JupiterWeb's Linux/Open Source channel, which includes Linux Today, LinuxPlanet, and AllLinuxDevices.

This article was first published on ServerWatch.com.

Comment and Contribute


(Maximum characters: 1200). You have characters left.