I predict this thing lasts no more than 2 weeks before someone finds a way to break the "unbreakable"

Seems to me like you are inviting trouble with a name like that.

Even so, this would sell like butter. Can't you hear CEOs masturbating over this?

Unfortunately, yes.

PHB: "We need to get this in here! It's UNBREAKABLE!" IT Guy: "Sir, I assure you, it's breakable." PHB: "That's not what the shiny pamphlet says." IT Guy: Facepalm

Basically, it's a more-current version of the Linux kernel added to their distro (which is a clone of RHEL5).

Surprise surprise... the newer kernel has more performance optimizations and is somewhat faster/featureful.

There are a few new features - see here for more on one of the data integrity ones:

http://www.c0t0d0s0.org/archives/6919-About-this-data-integr...

Yep. And RHEL MRG already includes a 2.6.32 kernel on top of regular RHEL 5. This probably even uses the same .config file.

So they have a two pronged approach:

1. Encourage customers to adopt this kernel to run Oracle workloads. Saying it's the only "recommended" kernel for Oracle software helps greatly. Oh and it's 75% faster.

"The Unbreakable Enterprise Kernel is now the only Linux kernel Oracle recommends for use with Oracle software."

2. Still "support" Red Hat's older kernel for other ISV workloads (really Red Hat is supporting it anyway).

"Oracle Linux continues to include Oracle’s existing Red Hat Compatible Kernel. Customers can choose to run the Red Hat Compatible Kernel if they prefer strict Red Hat compatibility over a system optimized to run Oracle and other enterprise software."

Unbreakable Linux didn't really present any huge advantages before. Now it does. Why wouldn't you choose it now?

Can't the other distros (and the kernel developers) just take these patches and apply them upstream if they are so wonderful? I'm not sure how having patches to a GPL'ed kernel gives you a competitive advantage..

RHEL doesn't follow mainline, so Oracle will almost always have a more modern kernel. As another commenter pointed out, most of the improvments are just from this one point.

That sure sounds like biting the hand that feeds you. When you are (essentially) a re-distributor of RHEL5, why on earth would you bash on RHEL5!?

Because they compete with Red Hat and want to steal their customers. And now they can claim they have a superior product because they are using a modern kernel. Red Hat needs to release RHEL 6 ASAP.

To paraphrase some brand of beer, the difference is breakability.

It's nonsense and exactly what branding is for, a linux kernel is a commodity with a price near zero, in order to differentiate a commodity product you need to brand it.

Branding is about perceived differences and having the consumer identify with your brand. Not that you are repackaging the same goods as your competitor, anyone who knows that is already lost to your campaign and as such is irrelevant.

If anyone's interested with PostgreSQL tests on this kernel: TPS is 20% lower than the stock RHEL 2.6.18-194.11.4.el5 kernel.

I wanted to start flaming oracle for hurting the IT all around, but somehow, I lost my words. My response to this would be a solid facepalm.