Submitted by Today SUSE, the company behind Rancher, NeuVector, and SUSE Linux Enterprise (SLE) and a global leader in enterprise open source solutions, announced it is forking publicly available Red Hat Enterprise Linux (RHEL) and will develop and maintain a RHEL-compatible distribution available to all without restrictions. Over the next few years, SUSE plans to invest more than $10 million into this project.
The spicy bit here is that the CEO of SUSE, Dirk-Peter van Leeuwen, worked at Red Hat for 18 years before joining SUSE.
Excellent.
Dang… take that oracle.
But in 5 years will there be SLES available on the suse website and the suse-rhel fork on another website ran by a separate organisation that suse altruistically set up? Or will the fork be integrated into the opensuse/sles world and what shape will that take?
I doubt doubling up development efforts to own your competitor would be a long-term solution…
Oracle?
https://www.zdnet.com/article/oracle-takes-on-red-hat-in-linux-code-fight/
The only way I see this making sense is if SUSE is going to be providing a migration path from RHEL to SLE.
Like you, I doubt they’re going to be supporting two enterprise distributions indefinitely.
You’re likely correct that they won’t maintain 2 separate distros long term, but I could see them taking it the other direction also: deprecating SLE, and providing a migration path to the “new” RHEL-based distro.
My guess would actually be that they will deprecate both of them and provide a migration path to their fancy new buzzword-compliant immutable container-based atomic updating “ALP” system that they’re heavily developing and promoting. They’ll obviously keep the current SLE 15 around in maintenance mode for an extremely long time, but everything from SUSE and openSUSE is indicating that they’re betting big time on the industry moving away from traditional Linux operating systems to to the previously mentioned buzzword soup.
They’re not wrong. Containers make services more portable.
The container tooling is still kind of crap though, and one of the bigger selling points of migrating to OpenSuse is nicer tools to auto generate images for physical machines and containers.
You may not be wrong. I notice that they said they were going to “fork” RHEL. That could be quite different from the identically cloning that everybody else is talking about.
It would be quite an interesting move if they created something that was RHEL compatible now, allowing people to switch but which diverged longer term and became more something of their own.
Having to be “bug-for-bug” compatible with RHEL seems like a poor way to differentiate your own offering. I don’t think SUSE just wants to sell simple support for their flagship enterprise distro ( currently SLES ). Like Red Hat, they need to control it.
Given the staffing, expertise, and infrastructure that SUSE already has to maintain though, they are better positioned than anybody to pull off a sustainable RHEL clone. It may be worth the investment simply to stick it to Red Hat, to devalue their offering, and to raise awareness of the SUSE brand–especially in North America.
Suse’s Liberty Linux service offering already supports RHEL and CentOS.
https://www.suse.com/products/suse-liberty-linux/
Not the most recent RHEL versions though: “RHEL/CentOS versions supported: 8, 7, 6 (Reactive LTSS), 5 (XLSS; upon request”.
If I understand, this new announcement means that SUSE will take on the task of creating and maintaining an up-to-date RHEL clone and adding support for that to the Liberty Linux offering.
Naw, that’s too much liability. Suse is paying CIQ/Rocky Linux to do the dirty work of cloning RHEL, and then Suse will sell support.
From the article:
Yes, embodying “the core principles and spirit of open source”. LOL
Suse will also be competing with CIQ who also sell support for Rocky Linux (https://ciq.com/support/rocky-linux/), so this also looks like Suse paying off CIQ. CIQ gets paid to assume the liability of cloning RHEL, and they get paid because Suse was going to undercut them like CIQ is doing to RH.
$10M is a nice cash out for a bad business model. LOL
Flatland_Spider,
We’ve talked a lot about this and I understand your position and the reasons behind it, so I don’t think we need to rehash that. But I am curious: what is your feeling of the GPL? Is it a bad FOSS license in your view and in-hindsight, given that it permits freeloader cloning? How would you change or fix that?
Hypothetically we could try to impartially gauge the level of contribution in FOSS and factor it into a revenue sharing model. Note that I think it would be too complicated in practice, but it’s an interesting thought experiment…. Everyone’s share of compensation could be proportional to their output using a formula for lines of code and bug count. Everyone, Redhat included, would take a share proportional to their actual contributions whatever that may be. So if one contributed 5% of the code, they’d get a 5% share of sales, which seems fair in principal,
Ironically though, even if it were possible to fairly split revenue by contribution, I don’t think redhat actually wants this. Consider that they’ve been existing on both sides of this FOSS equation. While it’s true they contribute to FOSS, they’d owe a lot to the FOSS projects they’ve been including in their distro without any forms of compensation, the value of which could realistically total billions of uncompensated FOSS sales over redhat’s history.
This is such a complicated subject! Haha.
Alfman,
It is a complicated subject.
I don’t like the GPL because it doesn’t go far enough to ensure the code remains public and available. It’s not as dangerous and people make it out to be, and it’s a very capitalist license.
Places like Google have been larger freeloaders since the GPL’s definition of distribution doesn’t take into account distribution inside an organization.
(Oh yeah, from earlier. Here is a BSD like anti-capitalist license someone threw together. https://anticapitalist.software/)
I also don’t particularly like corporateware. You know software where the software is owned by a corporation rather then a community project which people can take and use.
(Which is funny because I’m typing this on a Fedora laptop downloading OpenSuse Leap ISOs for testing. My distros of choice greatly benefits from corporate sponsorship. LOL)
People and companies need to contribute in some form or fashion to keep the commons lush and vibrant. I try to pay my way by submitting bug reports and running upstream versions. It’s not much, but it’s what I can contribute.
FOSS should probably have state level funding. There should be government grants or budget dedicated to keeping FOSS viable. FOSS has enabled the last 20 years of development (good and bad), and it’s a significant driver of economic growth and innovation.
Flatland_Spider,
Yes, very much so.
A license like the AGPL addresses that concern specifically.
I hadn’t read it before, it’s a very interesting idea though. There’s definitely food for thought there.
Well, we live in a corporate world and most of the time unavoidably so.
Yes, I know you’ve been saying this, I just wanted to understand how you felt about the GPL given that it’s a bit at odds with your vision of how things should be. It sounds like you are in favor of a license that did more to address freeloaders.
I think it’s too little too late to change. Both the licenses and the projects using them are already so entrenched that they’re not going anywhere. Major corporations can actively avoid software under the AGPL or anti-capitalist software license depriving them of market share indefinitely.
I agree. This is not about donations necessarily, but government contracts should stipulate open source requirements for all software contracts. After all it is our money paying for it, we should be entitled to the source. It’s just indefensible that we the public are paying for proprietary software year after year and getting nothing for it. Contractors who want to take government money should be obligated to open source. This would be a boon for FOSS financing while increasing the benefits to the public.
Holy crap! What a move! good
Indeed. The enemy of my enemy is my friend.
Can anyone explain this to me?
Why not invest $10M into SLES? by promoting it as a true alternative to RHEL? Because I expect that maintaining and truly supporting RHEL takes as much as effort as it does to SLES. So they doubled their capacity needs while keeping RHEL users in RedHat.
Plus, by fighting back RHEL initiative is supporting leeches like AlmaLinux and RockyLinux: which compete with RHEL in support while not having to pay for engineering or development. At least RHEL develops things and upstreams a lot os stuff, something that Alma and Rocky don’t do at all.
The only thing I can think of is future business. Since RHEL is now even more tied to a support contract which stipulates you need as many subscriptions as you have RHEL installations to be able to claim support. A RHEL compatible stopgap gives you the option of not switching out distributions immediately. It’s clear Red Hat is tightening the screws to get more subscription revenue.
The effort to have a compatible non-Red Hat Enterprise Distribution is to forge a path forward, Where RHEL might not be the gold standard to certify against. It would be ironic if Red Hat tightening the screws would bring about a unified Enterprise Linux base (not necessarily pure Red Hat), shared between Oracle, SUSE and Rocky Linux and AlmaLinux.
Basically. Sell Rocky, RHEL support, and progressively sell more Suse tools and eventually convert them to SLES.
The enterprise Linux game is far more commercial than you’ve been led to believe, and many things like administrative tools are behind paywalls. Sure, you could download the sample and open source code, but they’re effectively useless in a real enterprise environment that needs the other tools. This is essentially SUSE declaring it’s going after Red Hat market share.
dark2,
I’ve noticed that redhat has moved some documentation, forums, and knowledge base content behind subscription paywalls too. Of course it is their prerogative to make these private for their customers only, though I find it frustrating when paywall content actually ranks higher than open sites elsewhere. I strongly prefer search engines to list results with publicly accessible content first.
Suse is putting $10M into Rocky Linux (from what I’ve heard) so there will continue to be a bit-for-bit RHEL clone they can support via their Suse Liberty Linux (https://www.suse.com/products/suse-liberty-linux/) service.
Suse already supports RHEL, and the clones, and Suse isn’t that different from RHEL. OBS has been able to build packages for distros besides Suse for years, and OBS can build other images with KIWI templates. They just need a project dedicated to tracking RHEL, which would be Rocky.
Yeah, draining RH is the idea. Two businesses competing for marketshare. There can be only one Enterprise Linux distro! LOL
Which is why the whole thing is dumb. They should rebase onto CentOS as their upstream and participate in the community forging their own way. (Don’t “at” me. We’ve already hashed this out.)
Forking isn’t that hard. And they have the cash, it gives them good PR and mindshare, and helps them be relevant outside of Europe (Which SuSe has always had trouble branching out of). Dunno if that’s worth $10 million though.
Also, it kind of takes a little bit of wind off their competitor’s sails.
I suspect that RH containers may be important for some big corporate SuSe customers, Specially since SuSe is big in the whole Kubernetes infra space.
RedHat making things weird for linux since the 90s ha ha
Suse isn’t even doing the forking, so this is even lower effort. They’re giving $10M to CIQ to keep Rocky Linux alive and being a thorn in RH’s side.
Only possible due to the inventor of GPL.
Richard Stallman deserves his place in history. Let’s not ignore all the code licensed under MIT, BSD, Apache, and other licenses though.
The BSD license is older than the GPL. The MIT license is even older and more software is released MIT than GPL ( probably in RHEL as well ).
Cool, do Canonical next. LOL
Also, spend some time to officially port RH Directory Server/FreeIPA to Suse/OpenSuse.
I did not see this coming but it is so brilliant.
This is a masterstroke of PR for starters. But it is really just an expansion of their Liberty Linux program that they started last year.
Red Hat is six times the size of SUSE and, as an educated guess from rough numbers, SLES has maybe 5 – 10 percent of the market share that RHEL has in North America.
Growing enterprise Linux share in a market where RHEL already dominates must be hard. Buying SUSE support has meant switching to SLES which is just too hard. Liberty Linux allowed SUSE to grow their enterprise subscriber base without asking those customers to switch. But why buy from an also-ran when you can get the real thing from Red Hat?
Well, now the SUSE Red Hat clone will essentially be the real thing with enterprise support available from the core maintainer to boot.
The cost of maintaining a RHEL clone is nothing if it allows them to close some of that 83% market share gap.
Amazing move.
Suse have been brilliant and shifting the narrative here. Everywhere I was reading about people moving from Redhat to other distros, mainly Debian based ones with Suse barely getting a mention. If they become the fork all the other distros base themselves off suddenly the old Centos/Redhat for prod could become SuseCent/SuseRHEL (they’ll need some sensible branding work doing here). Longer term it will be interesting is they decide to converge SEL and SuseRHEL or if they instead target them at different customers.