
Having had the chance to sleep on Oracle's Unbreakable Linux announcement I woke up with a nagging feeling that all was not right with the plan. There are a number of potential problems I could see with it, while there was a bigger issue - one that I could not quite put my finger on - as to why this would not work.
Fortunately Red Hat has responded to the list of potential problems far more articulately than I could first thing in the morning. According to Red Hat, the Unbreakable plan will fork the RHEL code base and cause customers hardware and software certification problems.
Under the heading Unfakeable Linux Red Hat has outlined its response. Here are some of the salient points:
Q: Oracle says their Linux support includes the same hardware compatibility and certifications as Red Hat Enterprise Linux. Is this true?
A: No. Oracle has stated they will make changes to the code independently of Red Hat. These changes will not be tested during Red Hat's hardware testing and certification process, and may cause unexpected behavior. Hence Red Hat hardware certifications are invalidated.
Q: Oracle says their Linux support includes the same software compatibility and ISV certifications of Red Hat Enterprise Linux. Is this true?
A: No. Oracle has stated they will make changes to the code independently of Red Hat. These changes will not be tested during Red Hat's software testing and certification process, and may cause unexpected behavior. Hence Red Hat software certifications are invalidated.
Q: Will Oracle's product result in a "fork" of the operating system?
A: Yes. The changes Oracle has stated they will make will result in a different code base than Red Hat Enterprise Linux. Simply put, this derivative will not be Red Hat Enterprise Linux and customers will not have the assurance of compatibility with the Red Hat Enterprise Linux ecosystem.
These are all issues that occurred to me, and Red Hat also comes up with a few more, namely: Unbreakable Linux will not enjoy Red Hat's security certifications, Oracle's security updates will naturally be delayed if they follow Red Hat's, Oracle will have a different build and test environment, Oracle cannot guarantee API and ABI compatibility with RHEL, and many of Red Hat's customers actually get their front line support from a hardware vendor, with Red Hat as backup.
Of course, Red Hat would say that.
Ingres CTO Dave Dargo has another view point, and it is one worth listening to since in 15 years at Oracle he established the company's Linux program office and defined the company's strategy on Linux. " Sometimes you just have to call bullshit on something and Oracle's announcement on Linux is as great a candidate as any I’ve seen lately," he writes although he's more polite about it than I am.
One of the many things Dargo questions is the suggestion that moving from Red Hat to Unbreakable Linux Network will be easy. "The claim is that it takes less than a minute to switch from Red Hat’s Network (RHN) to ULN. It’s going to take more than a minute, and a fair amount of cost, to get through the legal agreements and process of switching over," he writes.
Dargo also questions "why Oracle’s first real support network is for someone else’s product.Where’s the Oracle Database Network and Applications Network and PeopleSoft Network and Siebel Network?" he asks. "It’s amazing that they can provide all that for a mere $399 for a competitor’s products, but not for their own $200,000. product."
Then there's the matter of price. Oracle states that it is undercutting Red Hat on price, and while that might be true Dargo also points out that if you add in the cost of an Oracle database license and support for one year the actual saving for a four processor system comes to $500 on $200,000, a 0.25% cost saving. Is that worth the hassle of moving from one operating system support vendor to another?
Dargo also raises another big question, one that would explain why it does not matter if Unbreakable Linux actually works or not, which is I think the issue I was finding hard to pin down.
"If Oracle is tremendously successful in taking Red Hat’s business then, ultimately, Red Hat won’t be around," writes Dargo. "Oracle will then either need to acquire Red Hat or staff up to include the same resources that Red Hat has in building, distributing and supporting their product. Is this their plan, to get Red Hat’s valuation low enough to acquire them?"
The suggestion that Oracle could buy Red Hat started this whole thing off. I wrote at the time that Ellison's comments to the FT explained why Oracle would do no such thing. It's worth taking another look at them.
"I don't see how we could possibly buy Red Hat - IBM would just say, Larry congratulations, we're going our own way. They could just hire Red Hat people and they'd be in business straight away. So I don't see how anyone can buy Red Hat, not at anything near these prices, because anyone who feels like taking the code - they have no intellectual property," he said back in April.
The company had just acquired open source database vendor Sleepycat, however, proving that there was some value in open source acquisitions. "The price is reasonable and you're getting a high quality development team," he explained.
Is Unbreakable Linux just a ploy to drive down Red Hat's share price to make it more affordable for Oracle? If it is, it's working: Red Hat's shares were down 16.3% to $16.33 in after hours trading, wiping $685m off its market capitalization.
Is that reasonable enough for you Larry?
I really enjoy the viewpoints on this blog. For similar takes on Unbreakable Linux 2.0 (inclusing patent ramifications), please visit my blog at
kevinclosson.wordpress.com
You want an unbreakable Linux? Here are the steps:
1) Install Linux on any computer. Without the computer ever touching a network.
2) Shutdown and unhook said computer, placing it in a safe.
3) Bury the computer in the foundation of a skyscraper.
No software is truly unbreakable.