Is Red Hat the new Oracle?

27 Comments

Credit: Red Hat

Everyone knows that Red Hat(s rhat), the king of enterprise Linux, is banking on OpenStack as its next big opportunity. And most figured it would be aggressive in competing with rival OpenStack distributions — from Canonical, HP(s hpq), Suse, and others.

What we didn’t necessarily know until the Wall Street Journal  (registration required) reported it Tuesday night is that Red Hat — which makes its money selling support and maintenance for its open-source products — would refuse to support of users of Red Hat Enterprise Linux who also run non-Red Hat versions of OpenStack.

Since Red Hat accounts for more than 60 percent of the paid enterprise Linux market, that policy could stem adoption of rival OpenStack distributions. It could also irritate customers — many of whom don’t like the specter of vendor lock-in.

In this policy — which the company confirmed to the Journal — Red Hat seems to have ripped a page out of Oracle’s playbook. The database giant, as it expanded into other software areas, decided that it would not support customers running  non-Oracle virtualization, non-Oracle Linux etc., unless the customer could prove that its issue originated in the Oracle part of the stack. That went over like a lead balloon with users. Since Oracle announced its own OpenStack distribution this week — and it also fields its own Linux distribution — the stage is set for dueling single-source OpenStack implementations going forward. Needless to say, that could ding OpenStack’s promise of no-vendor-lock-in.

The Journal also reported that Red Hat employees were told to stop working with Mirantis, an OpenStack systems integrator that late last year started offering its own OpenStack distribution. Red Hat’s president of products and technologies (pictured above) Paul Cormier told the paper that Red Hat would not bring a competitor into its accounts.

Reached late Tuesday for comment, a Red Hat spokeswoman noted that OpenStack “is not simply a layered product on top of Linux — [RHEL] is tightly integrated into and part of OpenStack. It is much more complex and intertwined than, say, Microsoft choosing to run PowerPoint on iOS.” I will update this story with additional Red Hat comment when it becomes available. Mirantis could not be reached for comment.

This news, which came out of the OpenStack Summit in Atlanta, may unsettle corporate customers wary of committing too much of their IT budget to any one vendor. But it’s hardly surprising. All of these vendors, while pledging open-source goodness of OpenStack, also want to expand their own reach in customers’ shops. OpenStack competitors have been wary of Red Hat for quite some time, expecting it to try to replicate its dominance in the enterprise Linux realm with cloud with OpenStack.

To be sure this problem is sort of theoretical now, as IDC Analyst Al Gillen pointed out. “From a practical standpoint, this is a non-issue today since there are precious few OpenStack clouds in production use,” he said via email. He agreed that this smacks of Oracle’s support policies but attributed Red Hat’s action more to the “immaturity and fast release of OpenStack more than anything else.”

Mark Shuttleworth, founder of Canonical, which competes with Red Hat both in Linux and OpenStack, recently acknowledged that the battle front has moved from the single-node Linux server realm — where Red Hat won — to multi-node cloud deployments where Red Hat’s enterprise software licensing mentality could put off customers. This new policy will test how compliant big customers will be to such enterprise sales tactics in the age of cloud.

 

This story was updated with analyst comment and Oracle’s OpenStack news.

27 Comments

Bill Bickel

Some things I think people should consider.
In my view, when looking at a new segment (e.g. IaaS, PaaS, Hadoop/BigData, No/NewSQL Data), both existing large companies and smaller for-profit companies need to carefully study and look for the best places and best chance to participate, add value and make money. Also looking to see if there are existing companies that might have a very good shot at winning parts of that new market, and look to avoid or complement those.

In the case of OpenStack, which is an open-source centric market area, and where the key technologies relies alot on Linux, and to some extent KVM, I would think that people would give Red Hat a pretty good chance at winning the major role in this market. Looking at Red Hat and seeing that they have become the #1 contributor to the project, and given Red Hat’s track record in linux and jboss, and in being able to balance community with commercial open source to profitable success, I would stay away from the key market area that Red Hat is targeting.

If I add in the fact that openstack, when combined with Linux and kvm is kind of like “Super-sized-Linux” in many ways, and it will require a strong third party ecosystem of hardware and software participants and developers. I would then look and see if players in this market have experience in this area, have existing relationships with ISV’s and IHV’s, and are trusted by the third parties to build a sustainable platform, price it fair and provide updates and support. On that front Red Hat would be in the best position, and maybe IBM has some chops there, and VMWare some, but they are not fully in with OpenStack from what I see. Given that you have a company in Red Hat, that most all ISV and IHV companies would trust to be a solid, stable company, and have support and technical skills to back up their solution, and open source skills to balance community and commercial interests, I am not sure why people are trying to take the core Openstack part of the business. That is a very tough battle.

I think it is better to get some concentration of force around one company that can be trusted to drive it forward, and build applications and hardware and software solutions around it. Having 5-10+ OpenStack versions confuses third party ISVs, IHVs and end customers.

I can see why someone like HP fancies going after this area. They currently have no strategic infrastructure software (i.e. virtualization, operating system, database, middleware), and with IaaS emerging as a strategic new infrastructure technology, I can see why they would want to try it. The challenge is that their ability to attract an ecosystem that includes their head-head system vendors like Cisco, Dell, IBM and others, is very unlikely to happen. I just don’t see it happening.

Barb Darrow

Nothing to amend here. Red Hat, with its open source roots, is trying to slow down adoption of non Red hat parts of the stack via its support policies and its own FUD.

Sinclair Schuller

Paul’s statement is carefully crafted (http://www.redhat.com/about/news/archive/2014/5/on-openstack-and-open-source). It’s a statement that seems focused on providing a truthful statement that acts as a distraction without answering the question: Will RedHat provide support for RHEL running on other OpenStack distros? Also, given the strategic importance of OpenShift, what does this mean for their PaaS?

“To be clear, users are free to deploy Red Hat Enterprise Linux with any OpenStack offering, and there is no requirement to use our OpenStack technologies to get a Red Hat Enterprise Linux subscription.”

Yes, this is true. That doesn’t state that it will be supported. It says that (a) you can deploy RHEL and (b) there is no requirement to RHAT OpenStack to get RHEL.

The next statement is “As a matter of fact, in traditional virtualization environments we certify performance of Red Hat Enterprise Linux on offerings from our largest competitors, Microsoft and VMware” which is also true.

But statement 1 and statement 2 do not, together, constitute an answer of ‘yes’ to the core question proposed at the onset of this comment.

Steven Citron-Pousty

I am curious that now that the facts have been made clear why you don’t actually put an update at the end of the article. I think it is pretty clear that the WSJ piece and others were FUD about Red Hat and leaving this up without the update seems to not be accurate.

bjs

Gotta love the FUD, which is always proliferated. Paul sets it all straight, but it will be missed.

http://www.redhat.com/about/news/archive/2014/5/on-openstack-and-open-source

“To be clear, users are free to deploy Red Hat Enterprise Linux with any OpenStack offering, and there is no requirement to use our OpenStack technologies to get a Red Hat Enterprise Linux subscription. That’s what open source enables, and it’s not a new way of business for us. As a matter of fact, in traditional virtualization environments we certify performance of Red Hat Enterprise Linux on offerings from our largest competitors, Microsoft and VMware.”

sstave

When I was attempting to implement RHOS FOR Redhat as an employee of Mirantis (no longer with that company) — it was completely unworkable and not ready for prime time.

Sad to see RedHat fall so far to become that which it was founded on despising; the monolithic enterprise that can only defend its territory through idiocy and shenanigans rather through true innovation.

Jan-Willem Lammers

Pivotal CloudFoundry runs on Openstack, AWS, vCHS, and vSphere as I understand it.

Sinclair Schuller

I’m not so sure that’s true. From the looks of it, it’s VMware only. The opensource project might not be dependent on vSphere, but the commercial offering is. Here is some evidence:

Prerequisites are full of VMware: http://docs.gopivotal.com/pivotalcf/customizing/requirements.html

Their launch announcement 7 months ago stated: “Pivotal CF, the leading enterprise distribution of the Cloud Foundry â„¢ platform, is now available on vSphere…” – See more at: http://blog.gopivotal.com/pivotal/news-2/pivotal-one-launches-now-every-enterprise-can-develop-like-an-internet-giant#sthash.B7DYGPHZ.dpuf

The commercial offering listed as a VMware product: https://www.vmware.com/products/pivotal-cf

Last, let’s not forget that Pivotal bluntly states and views Amazon as their competition. The goal of CF long term will be to be an onboarding vehicle to *their* public cloud. AWS compatibility might be true (I have no evidence that it is or isn’t), but long term will likely not be the case since Pivotal would want to direct migration flow from on-premises to their public cloud.

Sinclair Schuller

This isn’t surprising. I recently wrote about lateral vs. vertical risk in vendor selection. This is a vertical lock-in play, which is lock-in in it’s most dangerous form.

http://apprenda.com/blog/lies-come-flavors-paas/

It’s a problem in the PaaS space as well. Technologies outside of those supplied by an owning, biased vendor will always receive exclusive support or artificially better treatment. For example, RedHat OpenShift exists for driving more RHEL and JBoss EAP. It does not exist for helping the customer achieve an independent cloud vision. Pivotal and their Cloud Foundry commercial offering? Requires VMware to run. So much for “open.” Customers need to focus on selecting technologies that afford them with vertical stack freedom. If not, they’re buying into LIES – Locked In Enterprise Stacks

Steven Citron-Pousty

As the CEO and co-founder you must realize that people should not really trust your opinions as fact. While it is an interesting take on OpenShift (which I am an evangelist for) it is not true.

Sinclair Schuller

Yet the same principle doesn’t apply to you, as an employee of the company in question? Got it. Going forward I’ll remember that it’s unlikely that a company with a $9B market cap isn’t behaving in a fashion that’s defending it’s commercial interests, and those that it employees never have a bias.

Interestingly, I don’t even fault RedHat for this. It’s expected behavior. Revenue streams and existing product lines always have a tremendous amount of pull in any company’s future strategy when the two may be at odds with each other. This is expected, but it’s also bad for the customer. Lock-in is lock-in.

Lloyd Dewolf

How quickly things change, less than a year ago:
“Red Hat is taking a leadership role in bringing OpenStack to enterprise users, and Mirantis has been an early leader in developing services and tools that are helping enterprise customers achieve success with OpenStack,” declared Brian Stevens, Red Hat CTO and vice president of Worldwide Engineering. “Our investment in Mirantis reflects our commitment to customer success and customer choice, and the on-going collaboration between Red Hat and Mirantis will help accelerate the adoption of OpenStack in the enterprise.”
http://www.mirantis.com/company/press-center/company-news/ericsson-red-hat-and-sap-ventures-back-mirantis-with-investment/

Lloyd Dewolf

Wow, this is a sad day for open source and Red Hat. Red Hat has lost their way.

Historically, Red Hat was committed to supporting RHEL as a guest OS with the caveat that a problem believed to be caused by the host OS (hypervisor) would need to be reproduced in isolation or on Red Hat’s own stack.

Barb Darrow

that is my understanding of how Oracle support works as well @lloyd. Problem in a mixed stack has to be replicated to show that it comes from the Oracle part of the stack before oracle will support it. In theory i “get” that, but customers hate the finger pointing.

Cloud Insider

I can understand competitors getting upset. But if you are RedHat, do you really care about competitors? Wouldn’t you want to crush them anyway? Much as this causes angst amongst competitors, I cant help but think that it is actually a good business decision by RedHat. They are not in the charity business of helping out competitors.

Abi

Red Hat accounts for more than 60 percent of the enterprise Linux market !

Avi Alkalay

It takes time and a lot of efforts to integrate all the components of a complex software as OpenStack along with the underlying operating system. Not to mention the way all the stack was compiled, with performance switches etc.

That being said, it is natural that RH says they’ll only support their stuff. It is difficult, almost impossible, to debug problems if you don’t control the whole stack. It is too risky otherwise. I agree with Red Hat’s position.

cl

That’s not what it says on the article. It explicitly said that RedHat will “refuse to support of users of Red Hat Enterprise Linux who also run non-Red Hat versions of OpenStack”. Meaning that if my company has 2 datacentres, one running RH’s OpenStack and another one running another instance of OpenStack, RH will refuse to support my RH OpenStack installation.

Yuri Grinshteyn

I think that’s the most severe interpretation of the wording in the article. I am sure that RH will still support your RH-only DC and most likely even the RH servers running the non-RH OpenStack distro. They just won’t offer any support for anything that even smells like an OpenStack issue on them.

Disclaimer – I work for Oracle, and this would be very similar to our support policies (as I understand them).

Barb Darrow

my guess is that what Red Hat’s policy says explicitly will not necessarily happen in the field. If it’s a big customer, RH, like Oracle, will probably be more flexible. Why tick off a big customer?
On the one hand, i get why one vendor does not want to support another competitor’s technology. On the other hand, big companies do not want to be overly reliant on one vendor — cloud migration was seen as one way to break old vendor dependencies.

thanks for all the comments.

Barb Darrow

this blog does not clear it up. Q: if you run RHEL and choose to run Canonical or hp or some other OpenStack, will Red Hat support that overall stack? The answer still looks seems to be no. I understand the rationale, but i don’t think this blog addressed the question. RHEL is supported by RH. but an outside OS, i don’t think so.

RU Serious

In what way exactly is the blog not clear?

“To be clear, users are free to deploy Red Hat Enterprise Linux with any OpenStack offering, and there is no requirement to use our OpenStack technologies to get a Red Hat Enterprise Linux subscription.” Original allegation from original article refuted. Clearly.

Is it because the blog doesn’t address your farce of a follow up question? Why would RHT commit to supporting a “stack” of software that they didn’t produce and qa? Why would HP want RHT trying to “support” their OpenStack distro?

What mythical software vendors are you dealing with that will offer such support?

RHT supports RHEL and they support RHELOSP. Anything else you care to run, you get support from someone else.

You jumped on the WSJ article, you were wrong. Be big enough to admit it.

Comments are closed.