Since Oracle has weighed in about Red Hat’s source changes, it’s time to take a look at the history of Oracle Linux. That takes us back to 2006, the world of enterprise computing, and into new markets. Specifically, Java and middleware.
Climbing the stack
In the early days of Red Hat Enterprise Linux (RHEL), Oracle was a boon for Red Hat. Being able to run Oracle’s stack on RHEL meant Red Hat could expand with existing customers, and opened doors to business for Red Hat that wouldn’t have been open in the first place.
Heading into 2006, Red Hat was doing pretty well for a company that sold free software. Its annual revenue for FY2006 was $278.3 million, growth of 53% from FY2005. (Note: Red Hat’s fiscal year 2006 started in March 2005 and ran through February 2006. As part of IBM, its fiscal year is now tied to the calendar year.)
But public tech companies are always being pressured to grow and diversify by investors. It’s not enough to have a profitable business, especially if that income is tied to a single product. Red Hat’s focus on Linux may have been sufficient for Linux enthusiasts, but it wouldn’t keep investors happy forever.
Also? Red Hat had $1.1 billion cash on hand. Maybe it wanted to do some shopping. Celebrate a little bit.
It’s 2006, you’re a young company with some cash in your pockets and eyeing a little something to impress your friends. A Java middleware company would do nicely. So Red Hat settled on JBoss, another open source company that would let Red Hat expand into Java middleware and the buzzword heavy world of service-oriented architecture (SOA). (You laugh at the SOA buzzword now, just wait until we’re nearly 20 years away and talk about “digital transformation,” “shift left,” and DevSecDataSREOps or whatever…)
Oracle doesn’t have customers…
The problem? Well, Red Hat wasn’t the only bidder on JBoss. Oracle also had its eye on JBoss, but lost out to Red Hat. This did not please Oracle’s Larry Ellison, who told Financial Times flat out that Red Hat was too ambitious:
Red Hat’s own acquisition last week of JBoss – another open source company whose products compete with Oracle and IBM – could shake up alliances in the Linux world and provide another reason to act, said Mr Ellison.
“Now that Red Hat . . . competes with us in middleware, we have to re-look at the relationship – so does IBM,” he said.
On Red Hat’s growing influence and its ambitions to reach beyond the Linux operating system, he added: “I don’t think Oracle and IBM want another Microsoft in Red Hat.”
There’s a popular saying in the tech industry about Oracle: “Oracle doesn’t have customers, it has hostages.” Oracle had started to mull over having a “complete stack” so that it could avoid messy things like competition or sharing customers with other companies.
Unbreakable Linux
The JBoss acquisition was announced in April 2006. By October, Oracle was ready with its response: " Unbreakable Linux."
Was Oracle going to develop and maintain its own enterprise Linux distribution? Oh, goodness no. Oracle was going to start with Red Hat and then undercut Red Hat on price.
Oracle is offering its Unbreakable Linux program for substantially less than Red Hat currently charges for its best support. “We believe that better support and lower support prices will speed the adoption of Linux, and we are working closely with our partners to make that happen,” said Oracle CEO Larry Ellison. “Intel is a development partner. Dell and HP are resellers and support partners. Many others are signed up to help us move Linux up to mission critical status in the data center.”
“Oracle’s Unbreakable Linux program is available to all Linux users for as low as $99 per system per year,” said Oracle President Charles Phillips. “You do not have to be a user of Oracle software to qualify. This is all about broadening the success of Linux. To get Oracle support for Red Hat Linux all you have to do is point your Red Hat server to the Oracle network. The switch takes less than a minute.”
This is a recurring theme over the years: Let Red Hat do all the development work of pulling together a RHEL release, clone it (or use another clone) and then offer to “support” the clone. In Oracle’s defense, they were committing at the time to doing backports and some other bugfix work you don’t usually find in other RHEL clones. Still, Oracle was basically trying to do two things with Oracle Linux (neither of which friendly to “the community”):
- Cut off Red Hat’s oxygen supply. By undermining RHEL subscriptions, Oracle was making it harder for Red Hat to grow and compete. Clearly, Red Hat was using its profits to do things that didn’t make Oracle happy.
- Trying to have the “complete stack” to ensure Oracle gets as much spend as possible. We see this again a few years later when Oracle tried its hand with Sun, and even more ambitiously by adding hardware.
Is it a fork? Well, it ain’t a spoon…
It’s funny how times change, and how they don’t. Very quickly after Oracle’s “Unbreakable Linux” announcement, Red Hat came back with “Unfakeable Linux.”
Let me, for a moment, complain about web redesigns. Red Hat has, of course, redesigned its website a number of times since 2006. The “Red Hat responds” page is now swept away from the Red Hat site and only exists on The Wayback Machine.
In short, Red Hat wanted to let the world know that RHEL was “unfakeable” and that with Oracle making changes to the code it meant that it was no longer RHEL.
Indemnification
Here’s a word you probably don’t think about much in 2023 around open source, but was a huge deal in the mid-aughts: Indemnification. Younger readers who didn’t live through this period, rejoice.
In and around 2006, everybody was obsessed with “indemnification.” Basically, who’s going to take responsibility if my business gets sued over patents or claims that open source code contained proprietary code.
You see, in 2003 the company formerly known as Caldera International had gotten a new CEO by the name of Darl McBride, a new name from a previous company (SCO) and a new business strategy: Sue IBM. SCO Group had filed a $1 billion, with a “b,” suit against IBM.
It claimed IBM had “devalued” its SCO UNIX by copying code into the Linux kernel. At the risk of spoiling the plot… they didn’t win. The strategy appeared to be to force IBM to settle or acquire SCO. It didn’t. You know the phrase “never start a land war in Asia”? “Don’t sue IBM over intellectual property” would probably be equally good advice.
But IBM didn’t need its armies of lawyers and patent chest to win this. SCO’s claims were pathetic and its research was even worse. Discovery found that SCO Group didn’t even have full rights or standing over the code they claimed IBM had copied.
In fact, SCO became a laughingstock. In the process, “the community” rallied around IBM (you know, the same IBM that people now say has always been Evil™) and the industry started building defenses against nuisance suits and egregious intellectual property claims.
But, at the time, proprietary vendors flooded the zone with Fear, Uncertainty, and Doubt (FUD) about these suits. Microsoft, the leading FUD-spreader of them all, tried to fill the world with fear of Copyleft code tainting precious intellectual property. There’s more than a little irony that Microsoft has gone from “Linux is a cancer, don’t touch GPL’ed code” to “why not use this AI thingy trained on open source code to inject code straight into your product?”
Red Hat and other vendors had to provide “indemnification” guarantees to customers, lest they fret too much about adopting this risky open source stuff. And Oracle’s indemnification deal, Red Hat wanted to point out, wasn’t as good as theirs and any modifications by Oracle were especially not included in Red Hat’s Open Source Assurance Program.
With its Open Source Assurance Policy, Red Hat focuses on the Customer’s business continuity in the face of an infringement claim. With Oracle’s indemnity program, you only get an indemnity so long as you give Oracle an unlimited one in return.
And, yep, we’ll be revisiting indemnification, SCO, and more soon.
No Kernel (separate patches) for you
This was in the RHEL 4 timeframe. Yes, friends, all this stuff has only brought us up to RHEL 4!
- RHEL 2.1 was released in 2002, support ended in 2009. (Not RHEL 2. RHEL 2.1)
- RHEL 3 came out 2004, was EOL in 2010. (ELS 2014)
- RHEL 4 came out in 2005, was EOL in 2012. (ELS 2017)
- RHEL 5 came out in March 2007, EOL in 2013, and its extended life cycle support (ELS) ended (drumroll) November 30, 2020.
That’s not a typo. Red Hat finally told customers “no more, already” for RHEL 5 in November 2020. (Note, “EOL” means end of full support, here. Also, early version of this post had errors in the dates but I was kindly corrected on Mastodon.)
Why? Because RHEL was no longer just an operating system. It had become a platform. Red Hat’s vision was that they would expand up from the operating system to middleware and beyond. The larger industry was saying “oh no you don’t.”
It’s crucial to understand what’s really being fought over when Red Hat and IBM try to protect RHEL as a product, and when the larger industry demands the ability to clone RHEL bug-for-bug.
This platform business meant that RHEL releases got slower, got bigger, had a bigger scope than prior RHEL releases, and much higher stakes. It made it harder for competitors to elbow in because of inertia. In contrast to Oracle, which is the sole owner of the copyrights to its cash cows, Red Hat primarily owns its brand and surrounding knowledge of building its product.
Competitor wants to fork the open source project behind Red Hat Satellite to compete with Red Hat? Go for it! Competition is good. Copying bad.
It’s in that light, of protecting the product, that Red Hat did something in 2011 that sparked a number of complaints and commentary about its business practices: It “obfuscated” the Linux kernel source code in RHEL 6 to throw a wrench in Oracle Linux and others who were trying to clone and support RHEL.
The code was still released, it just didn’t break out Red Hat’s patches against the upstream kernel. If you just wanted to recompile the kernel, no problem. If you wanted to understand the differences, the whys, and so forth, then problem.
It had the unfortunate side-effect of harming the larger kernel community, but it was aimed squarely at Oracle. Nobody tried to claim it wasn’t GPL compliant, but complain they did. That episode will be the topic of the next blog post. (Yes, there are a lot of “next blog posts” in the works. Starting to think I should just write a book…)
Oh, and the platform thing? That also had a part in Red Hat’s interest in CentOS and what is now CentOS Stream. The model that existed around 2012 and 2013 of developing open source infrastructure on top of Linux (think OpenStack, Eucalyptus, Apache CloudStack, XenServer, and such) got much harder, and some flexibility was required.
We will also come back to RHEL as a platform, and how that has shaped Red Hat’s actions ever since, even if the larger perception of RHEL has never caught up to its actual identity.
Oracle Linux and its impacts
To sum up: Oracle isn’t giving away hugs and puppies because it cares about freedom. Oracle Linux started as a jab at Red Hat for acquiring a company Oracle wanted and because Oracle wanted even more customer money than it was already getting. The pretense that Oracle is morally superior in some way to IBM is silly, even in this lone area.
When it has suited Oracle to align interests with IBM, it has done so. When it’s convenient to take jabs at IBM, it does that too. Which is fine, on a business level. It’s just disappointing to see many people quick to side with Oracle or against Red Hat without taking into account the full history and understanding of how we got here and the implications on where we’re going.