Discussion on the future of Linux in the Enterprise

I would like to start a discussion on how we see the future of Linux in the enterprise. One realization I had:

  • A transition from paid OS support to just hiring Linux Administrators.

All throughout school, I learned on CentOS and RedHat. My professors preached “pay for the support because it like insurance”. But at my workplace we use the latest Ubuntu LTS release on everything. At first I was shocked, but the biggest thing I found that Ubuntu beats the other distros on is support. Both self-support, like though forums, and software support. So far we, and I personally, have not come across software that is not supported on Ubuntu.
The second part of my realization is loss of paid support. The reason I see for this is that we are seeing fewer errors caused by kernel bugs, but rather the software is bugged. Mainly I see companies are dropping paid OS support and paying for support of the software used. At my place of work, we have only seen about 1 kernel panic per year and recently, they have all been caused by a bug in the software, not the kernel or OS.
The Linux job market is healthy. But when you look at the RHEL market you see that full-time postions have declined while contract work has increased. Where the Ubuntu job market is increasing.

Ubuntu haters read this:
I am not here to praise Ubuntu, I am using it as an example. I use Fedora and CentOS at home.

This discussion is important because the future of Linux relies on the enterprise. Their need (and willingness to pay) for some features is what allows some development to happen.

I would love to hear anyone’s thoughts, ideas, questions, realizations, or concerns on how we see Linux growth in the enterprise.

I’m not really sure where you’re going with this. At first blush it looks like you’re concerned that companies aren’t paying for OS support, then you’re concerned about Linux jobs disappearing, but then show where Linux jobs are increasing.

Linux in the enterprise is fine, if my own neck of the woods is even somewhat normal. For the last roughly 7 years I have had a choice of Linux jobs when I was looking for new work. A lot of that was direct hire. At this point if someone at Robert Half even thinks I might be looking for another job, they have Linux positions lined up for me.

From what I can see, Linux is doing fine.

Keep going in this line of work, you’ll find Ubuntu beats the pants off of lots of traditional enterprise distros. In the enthusiast market, like around this forum, you’ll see people rocking an Arch install with a 4.12 kernel. In the enterprise, you’ll see brand new RHEL 7 machines rocking a 3.10 kernel. I feel like Ubuntu strikes a nice balance between the two.

1 Like

Kernel panic due to bad software is a kernel bug, and shouldn’t be happening.


I think it’s also a healthy practice to look at the competition. Enterprise has typically had 2 major use cases to consider, server and desktop/workstation.

It saddens me to say, but there’s a lot of shitty / niche windows desktop software out there that small and large enterprises have come to depend on.

Once that software is rewritten into SSO/oauth webapps, things will be better on the desktop side of things.

On the enterprise server side, things are better than they ever were IMHO, you can run Linux/Samba for your core user management functions and file sharing, and you can run windows server in VMs to run the other half of shitty niche enterprise software.

I think that the main drawback to Linux in enterprise is still in the lack of high quality documentation and in how to do things, when compared to Windows. For example, making a windows server a domain controller, and joining a machine to it, fits in a couple of screenshots, it’s trivial for a very junior admin to do. — not so much on Linux where there’s a lot more variety of software and configuration.

I was wondering where our weekly new thread questioning the status of Linux had got to :smiley:

My tuppence worth;

  1. Linux desktop in the enterprise is not really going to happen, not in force anyhow. Developer yes, partly depending on fashion, average business users who need more than a smartphone/tablet no. Windows laptops are easy to manage in large numbers and all the software is there today.

  2. More and more servers will deploy Linux, whilst parts of Microsoft still earn their wages/bonuses from peddling Windows licences they are not where MS is investing. MS are at the point where the Server OS is just a commodity item; they want your business to deploy its servers & applications into Azure regardless of OS or language. MS contributes a lot to ensuring Linux is a 1st class citizen in Azure/Hyper-V as a server OS (but not for desktop use, even as a VM)

  3. Linux with Kubernetes/Docker has really caught on. There will be loads of jobs in this area for sysadmins, DevOps etc.

  4. RedHat support agreements can look expensive compared to Oracle and Microsoft; it really depends on how much they want your business etc. One of my clients won’t be renewing with RedHat, they are big Oracle customers and Oracle have just thrown in the Linux support with the database/middleware agreement. The linux admins don’t really care, they hardly ever need to raise a support ticket.

  5. It’s usually only the CFO who wants a gold plated support agreement, this helps act as an insurance policy, again most of the Sysadmins/engineers release how little value is had from said agreements for day-to-day troubleshooting.

From my view of the world RedHat can’t rest on their laurels and need to make sure they are a force to be reckoned with in the Cloud. Oracle are now playing catch up with AWS/Azure and the same truth has dawned on IBM and VMware. Just selling licences/support agreements is not how you become a big player any more.

1 Like

I’d say it’s looking better and better… even MSSQL is now available for Linux! Being able to run things like SSAS on a Linux box is a big + for many companies.
(Disclaimer: BI consultant, mainly on MS software)

I want to discuss any changes we are seeing or forecast with Linux in the enterprise.

Does your work pay for enterprise support? https://buy.ubuntu.com/

You should be clear about what you mean by support as you seem to use it interchangeably between meaning that the OS has a lot of software and the ability to get some kind of technical knowledge for issues.

In most cases, enterprise support would include support for software, at minimum it will include support for all software distributed by the supplier. Remember that the “OS” in this sense isn’t just the Linux kernel and grub to boot it.

Your software bugs can in most cases also be helped by paid support if you wanted it to. But even then, your kernel panic you had. Do you have a kernel developer on staff?

If so, I assume they fixed the kernel and send it upstream? If not, you could have used that paid support to get the technical help to fix that kernel bug for you. (just as an example)

Linux has always had a mix. In reality a lot of companies will run a mix of professionally vendor supported systems and company only supported systems.

For example, many pieces of work would run test and low risk systems with CentOS and run production systems with RHEL as while your admins will fix 90% of problems without much issues the knowledge base and specialist help is there when needed for that other 10%.

Smaller companies might get away with just running things without any support, usually its to save money rather than them not needing it. Hopefully you’ve got an admin good enough to not need the specialist help, and to have the right contacts when an fucked up issues shows up that’s not on ask.ubuntu with 10 different wrong solutions.

It really depends on your needs, I think bother are going no where and both are required for Linux to survive.

edit: Good use of paid support for Ubuntu for example is if your company requires further support for Ubuntu 12.04 which has 2 years extended security maintenance under their “Ubuntu Advantage”

For companies like mine, not a chance any time soon.

All vendors support windows, and even then their garbageware is absurd in terms quality. So much would be lost in convincing manufacturers to build, support and test multiple Os’s. Then there’s the other side of the coin. Training the technically illiterate employees to use a new environment. I work with 9 other engineers and if it were me I’d rather quit then have to provide IT support for that crew.

That being said m$ is really putting the squeeze on b2b costs. Everything is licensed per user and used over the web. You can’t purchase or own hardly anything. Our company is becomming very conscious of licensing costs and maybe that will drive alternatives.

No, we don’t pay for Ubuntu enterprise support and we don’t plan to. We do pay for support for the software for our storage solution. The kernel panics we had with that were solved by talking with the developers and releasing a patch to us and fixing the issue upstream.

But what I where I see companies going is just hiring people that can do kernel and OS development to support their systems, rather than another company.

Some companies will be happy to do this, but it means they are accepting some risk, the main drivers to pay for support often comes from the Finance department rather than IT. It’s as much about de-risking as it about fixing problems. If you pay through the nose for third party support and you have a serious problem that loses revenue the board want someone to sue…