Ubuntu vs RHEL in enterprise computing

Wednesday, March 14th, 2012

A remarkable thing happened this year: companies started adopting Ubuntu over RHEL for large-scale enterprise workloads, in droves:

w3tech.com historical analysis of web server operating systems

The trend is even starker if you look at what we know of new-style services, like clouds and big data, but since most of that happens behind the firewall its all anecdata, while web services are a public affair.

The key driver of this has been that we added quality as a top-level goal across the teams that build Ubuntu – both Canonical’s and the community’s. We also have retained the focus on keeping the up-to-date tools available on Ubuntu for developers, and on delivering a great experience in the cloud, where computing is headed.

The headlines for Ubuntu have all been about the desktop and consumer-focused design efforts, with the introduction of Unity and the expansion of our goals to span the phone, the tablet, the TV as well as the PC. But underpinning those goals has been a raising of the quality game: OEMs and consumers demand a very high level of quality, and so we now have large-scale automated testing, improved upload processes, faster responses to issues that crop up inevitably during the development cycle, a broader base of users and contributors in the development release, and better engagements with the vendors who pre-install Ubuntu. So 12.04 LTS is a coming of age release for Ubuntu in the data centre as much as its the first LTS to sport the interface which was designed to span the full range of personal computing needs.

We’re also seeing the wider community respond to the goal of cadence. OpenStack’s Essex release is lined up to be a perfect fit for 12.04 LTS. That is not a coincidence, it’s a value to which both projects are committed. Upstream projects that care about their user’s and care about being adopted quickly, want an effective conduit of their goodness straight to users. By adopting the 6-month / 2-year cadence of step and LTS releases, and aligning those with Ubuntu’s release cycle, OpenStack ensures that a very large audience of system administrators, developers and enterprise decision makers can plan for their OpenStack deployment, and know they will have a robust and very widely deployed LTS platform together with a very widely supported release of OpenStack. Every dependency that Essex needs is exactly provided in 12.04 LTS, the way that all of the major public clouds based on OpenStack are using it. By adopting a common message on releases, we make both OpenStack and Ubuntu stronger, and do so in a way which is entirely transparent and accessible to other distributions.

Quality. Design. Cadence. You can count on them in Ubuntu, and OpenStack.

99 Responses to “Ubuntu vs RHEL in enterprise computing”

  1. tom Says:

    Stats always lie.

    If you look at http://w3techs.com/technologies/details/os-linux/all/all
    you see that CentOS way more popular than Ubuntu and combined RHEL and CentOS have over 42% and Ubuntu has only 18%.

    All these stats say is that Red Hat has way more paying customers.

    Move along ..

  2. Linux Says:

    That’s great news! However, I wonder how many people use Kubuntu instead of Ubuntu? I heard many times KDE is more enterprise friendly.

  3. Stuart Ward Says:

    But if you add Centos in to that mix then Ubuntu still has a long way to go. But a good achievement all the same.

  4. mark Says:


    Yes, zooming out is interesting too, but then CentOS alone isn’t sufficient, you need to look at whole ecosystems, including Fedora on the RHEL side and Debian on the Ubuntu side. When you do that, the story remains the same. If you add up all the “Red Hat ecosystem” web servers: CentOS, RHEL and Fedora, you get 46.1%. And if you add Debian and Ubuntu, which is our “ecosystem”, you get 48.4%. What’s important, though, is that looking back a year, the numbers were 51.0% and 41.7% respectively. So in the past year, the RHEL ecosystem has shrunk 4.9% in absolute terms, and the Ubuntu ecosystem has grown 6.3% in absolute terms.

    Of course, if I wanted to use stats to create a better headline that would be easy: “RHEL ecosystem down by 10%, Ubuntu ecosystem up by 15%”. Or, leaving out the rest of the ecosystem, for a real punch, “RHEL down by 20%, Ubuntu up by 40%”. In one year.

    There are lots of other stories to tell – SUSE now down to 3% (“SUSE drops 30% in a year”) for a start. But my point was not to try and make nasty headlines, just to call attention to a meaningful shift in Linux adoption patterns, and I think the stats support that very well indeed.

  5. Marius Says:

    It doesn’t change the numbers but my feeling is that this is more caused by Red Hat’s weakness, not by Canonical’s or Ubuntu’s strength. Red Hat in general seems to have a lot of difficulty dealing with its own weight and its product and service landscape in a scalable way. We’ve been customers of them, and I regret to say that it was a really bad experience – RH felt like an “IBM without the firepower”, meaning there was lots of the typical inflexibility and slowness of a big corporation, but without its resources.

    I hope Canonical/Ubuntu will avoid that. I’ve always liked the Debian ecosystem a whole lot, and although I’m highly critical of the recent desktop madness in Ubuntu, the server system is my favourite for any size of deployment.

  6. YHVH Says:

    Hey Mark. I’ve noticed your recent use of the word ‘Cadence’. It’s not obvious to me what you mean, and it comes off like in house/marketing speak that you normally manage to avoid. Keep up the good work.

  7. Christoph Says:

    Mark, correct me if I am wrong, but these numbers don’t say anything about *enterprise* but about web servers. I mean, seriously, would you count every web server out there ‘enterprise’, even if it’s just a kid installing Ubuntu on their root server for hosting their private homepage? If you want enterprise, how about you publishing numbers of sold LTS subscriptions?

    And I wouldn’t dare calling Debian part of the Ubuntu ‘ecosystem’ because the term ecosystem indicates that there is a mutual benefit for *both* Debian and Ubuntu. I can hardly see how Ubuntu contributes back to Debian. And as long as Canonical developers file bugs in Red Hat’s bugzilla you could even count RHEL to “your” ecosystem because you benefit from their work. 😉

  8. mark Says:


    Cadence is about releasing on a predictable rhythm. It’s expressed in our six month release cycle, and in the two-year LTS cycle that is layered on top of that (every fourth release is an LTS release). It also gets expressed in the development process with one and two week iterations, and monthly milestones.

    Having that rhythm allows all sorts of creativity to flourish. The fact that one part of the complex community can know what is going on in another part without having to talk to them specifically, helps tremendously. All sorts of teams like testing and documentation and advocacy and artwork can make better plans because they can predict what’s needed and when.

    The agile development process recognises the value of tight iterations, what we’ve done is raise that to the level of the whole operating system. So people planning IT deployments, and other software companies and communities, can align with that to get the most out of the operating system.

  9. Andy Says:

    If you want to make a real comparison I’d suggest you chart Ubuntu -vs- CentOS -vs- RHEL, it’s a pity there’s no way to chart paid Ubuntu -vs- free.
    I suspect we’d find that if we compare paid Ubuntu to RHEL we’d see an enormous difference, we already know that CentOS deployments eclipse Ubuntu deployments.

    Throwing Debian into the Ubuntu bucket doesn’t make sense – there’s a huge difference between Ubunut -vs- Debian.

    What the numbers show is that people like free (as in $0) software.

    If you want to make comparisons about enterprise software then you need to use paid subscription revenue.

  10. Joe Brockmeier Says:

    “Of course, if I wanted to use stats to create a better headline that would be easy: “RHEL ecosystem down by 10%, Ubuntu ecosystem up by 15%”.”

    It would be more accurate to say *Debian* ecosystem here. It takes a lot of hubris for Ubuntu to try to claim Debian as part of *its* ecosystem.

  11. Bryn Says:

    What’s the source for the data in the chart? It’s impossible to evaluate a claim like this without knowing what the numbers are based on.

  12. Bryn Says:

    There’s a problem with your comments box. I just added the comment above and when I hit submit on the box it displayed Zonker’s personal info in the fields: http://www.errorists.org/stuff/blogform-redacted.png

  13. Bryn Says:

    There are only 18 public web sites using Ubuntu? Label your axes, please!

  14. Bryn Says:

    Umm, what? That last comment is also mis-attributed to me. I did not write that.

  15. dude Says:


    The source of this data you graphed come from…….?
    Please share this with us.

  16. Jonathan Says:

    What does the term “Enterprise” mean in this survey? Am I using a web server somewhere in my organization to serve up a website. Or is it the hosting companies increasing or “cloud” companies increasing the amount of Ubuntu servers?

    How does it compare to the support that is paid for by companies on Ubuntu vs. the paid for updates for RHEL? What % of increase in support contracts are happening?

    Any numbers about Ubuntu in the Enterprise for Desktop computing?

    Are these numbers worldwide or US?

  17. Tony Says:

    I work with 500+ node HPC clusters and would love to use Ubuntu, but we use CentOS because kickstart is the killer app. I understand there are other bare-metal solution, but nothing approaches the sanity and simplicity of kickstart.

  18. cprofitt Says:

    Some very interesting information from the link tom posted http://w3techs.com/technologies/details/os-linux/all/all

    Popular Sites using Ubuntu:

    Those are some popular sites.

    I also noticed on the main page that the “fastest growing since 1 February 2012” was Ubuntu with 37 sites and Debian with 26 sites. (daily number of additional sites in the top 1 million).

    We can all debate what the numbers ‘really’ mean, but I am impressed with the growth of Ubuntu.

  19. Hakan Says:

    Enterprise? Wake me up when Oracle starts supporting Ubuntu for running their database software.

  20. Shuttleworth: Ubuntu vs RHEL in enterprise computing | wordpress Says:

    […] Ubuntu vs RHEL in enterprise computing Posted on 2012-03-14 by Mario Scondo Mark Shuttleworth claims that Ubuntu deployments now exceed RHEL deployments for "large-scale enterprise workloads." "The key driver of this has been that we added quality as a […]

  21. nathan Says:


    “If you want to make comparisons about enterprise software then you need to use paid subscription revenue.”

    What? Then Apache and nginx can’t be compared to IIS? Wake up.

  22. Jasna Says:

    You guys misunderstood Mark totally…

    He just wanted to show as he wrote: “But my point was not to try and make nasty headlines, just to call attention to a meaningful shift in Linux adoption patterns, and I think the stats support that very well indeed.”

    If I got it right, this statistics shows change of Linux pattern in general…right? Correct me if I’m wrong.

    Yes this all distros are connected and benefit from each other from Linux Foundation to each individual distro considering kernel etc…

    But you guys have to admit that Debian is in Ubuntu’s ecosystem since Debian is, as Mark said at UDS-O, their rock… Therefore this statistics does make sense…

    That’s just my point of view… no hard feelings..



  23. Jef Spaleta Says:

    Interesting spin on the reality that Canonical as a business hasn’t figured out how to self sufficiently monetize the overall global interest in using a commodity operating system that does not have a per deployment cost.

    There is no surprise that no-cost linux solutions of all stripes are as popular as they are. Debian and Centos lead the way in those stats for a very simple reason. They are gratis deployment options. I’m a huge fan of Red Hat as a business, and yet I don’t actually purchase any of their products. I have the skills necessary to deploy and maintain the “web server” up to the level required….which is no where near enterprise grade requirements mind you. So I deploy CentOS here and call it a day.

    No the fact that Debian and Centos account for 58% of the web servers in those stats isn’t remarkable at all.

    What continues to be remarkable is that Red Hat builds a _profitable_ business around a linux operating system that has an upfront deployment cost. And even more remarkable they are able to actually persuade people to layout cash, month after month, quarter after quarter to renew that service and support subscription. Every single RHEL instance represents a payment from a customer to Red Hat as a part of a business transaction. Every single one. That is absolutely amazing. You can’t say that about those Ubuntu stats. In fact Mark, you won’t tell us anything about the size of your paying customerbase for any Canonical for-pay offerings. You continually point to the size of the Ubuntu install base as a surragate for business health. I know better. The popularity of gratis deployments say absolutely nothing about how well Canonical is running as a business. Red Hat has built and continues to reinforce brand value in their RHEL offerings that lead directly to customer dollars. And this value clearly resonates to the target customerbase for the products.

    Canonical on the other hand..I’m trying to be as charitable as I can be here….has no clue on what people are willing to pay for. You can’t look to the Ubuntu community to make the case for Canonical’s value in the enterprise. Canonical has to step up its game by a lot and be much more focused about articulating value for money to potential customers.

    And Mark you are absolutely right, computing is headed into the cloud. And I fully expected that the gains you are seeing in cloud deployments are going to be chopped off at the knees by Amazon and other cloud providers when they start rolling out and servicing their own Linux based OS brands as part of their value-add monetization of their hosted cloud infrastructure. The gratis deployment lead Ubuntu has now in the cloud does not translate into customer dollars automatically. When push comes to shove, and people need to spend money they are not choosing Canonical, and there is no evidence on the table that that trend will change. Fix that as an executive on Canonical’s management team, fix that. The Ubuntu can’t fix it for you.


  24. Jasna Says:

    Would you give it a rest with money thing 😀 (I don’t think that’s the issue here 😉

    Thoroughly looking at the Mark’s graph here, address http://w3techs.com/ got my eye :)

    Well, look at the operating system category and suffer 😀

    Just relax everyone :D… you gotta’ admit that Ubuntu team is working very hard..

    Yes I’m almost 4 year old in Ubuntu world and open source in general… buuuut a lot of things changed since Hardy Heron :)

    bye bye

  25. Ubuntu 12.04: A ‘Coming of Age’ on Servers Too, Shuttleworth Says | Got2.Me Says:

    […] interface which was designed to span the full range of personal computing needs,” he wrote in a blog post on […]

  26. Jef Spaleta Says:


    Since you have voluntarily admitted to being a previous customer of Red Hat I have some follow-up questions. Are you currently a customer for Canonical for-pay services at present? Has your organization ever contracted with Canonical for services? Are you using Ubuntu in enterprise relevant server roles?


  27. mark Says:


    Wake me when Oracle has an effective response to Hadoop.

  28. mark Says:


    Oop, looks like a theme bug, thanks for the heads-up.

  29. Shuttleworth: Ubuntu rockets up at enterprises | Laptop Service in Moscow Says:

    […] data presented in a in a blog post by Ubuntu project founder and “benevolent dictator” Mark Shuttleworth is any […]

  30. 1roxtar Says:

    Man, some of you will never let Canonical/Ubuntu have their due. One speaks positively about Ubuntu and three more have to drop their two cents in a negative fashion. Geez!

  31. mark Says:


    Your observation that every RHEL deployment pays Red Hat, while only commercially-supported deployments of Ubuntu pay Canonical, is unusually astute.

    Canonical’s proposition is much better for all Ubuntu users, Red Hat’s proposition is much better for Red Hat.

    Nevertheless your focus on revenue is entirely appropriate, because for Ubuntu as an idea to be worthwhile it has to deliver both the open access that Debian and CentOS and Fedora offer, *and* the commercial qualities of certification and sustainability that are hallmarks of RHEL. If Canonical’s proposition is too good for users, then that undermines the long term prospects of Ubuntu itself. Paying users of Ubuntu are extremely valuable for all users of Ubuntu because they decrease the risk to them of Ubuntu disappearing some day. So it’s important that we have the support of that community when we make commercial propositions around Ubuntu, and I’m glad we can count on your support for that too.

    For the paying user base of Ubuntu to grow, we need a number of things to line up:

    * Ubuntu needs to be trusted in the sorts of workloads for which there is a commercial dynamic. Being free is of no value if you are not trusted, or suitable, for the workload. This is the real focus of my post: I think there are clear signals that the commercial market trusts Ubuntu. Large public companies are building clouds on Ubuntu. And big data farms. And large desktop deployments. There is starting to be a userbase for whom paying for solutions, support and expertise is normal and expected. Rockstar startup devops have long recognised the value and quality of Ubuntu, but also been quite comfortable consuming that value unpaid, which is fine by me. I think it’s a wonderful thing that I’ve indirectly helped so many rockstar devops build things that changed the world, and don’t want to undermine that. Separately, I think it’s a wonderful thing that CIO’s are now obviously happy to see Ubuntu running on critical infrastructure for which they are accountable.

    * Ubuntu needs to be certified. People who pay for assurance do not leave gaps in the stack. People who will not pay for assurance do not put assured pieces into the stack. Crossing from a user base that is largely of one sort, to one that spans both, is difficult. It’s particularly difficult if you come along long after the market has decided it has two options and that’s enough. But perhaps the scale of adoption, combined with the fact that there is actually an institution which can be a counterparty to a certification, will open the doors to certification for Ubuntu.

    * The value proposition from Canonical needs to be good. The key metric for me is the happiness levels of our paying customers, and they are very high. We provide a very good value proposition and, when called upon, excellent service.

    So, your focus on revenue is sensible, it is one I share, and one that was never going to be a quick story to tell.


  32. Shuttleworth: Ubuntu rockets up at enterprises – Mr Pod Blog Says:

    […] information presented in a in a blog post by Ubuntu plan owner and “benevolent dictator” Mark Shuttleworth is any indication, […]

  33. Is Ubuntu becoming a big name in enterprise Linux servers? | ZDNet Says:

    […] the company behind Ubuntu wants you to know that “A remarkable thing happened this year: companies started adopting Ubuntu over RHEL for large-scale enterprise workloads, in […]

  34. Is Ubuntu becoming a big name in enterprise Linux servers? Says:

    […] the company behind Ubuntu wants you to know that “A remarkable thing happened this year: companies started adopting Ubuntu over RHEL for large-scale enterprise workloads, in […]

  35. Jef Spaleta Says:


    Can you create a space for customer testimonials for the for-pay services Canonical is offering. Right now the case-study area is a garbled mess that mixes the value of gratis Ubuntu with the value of Canonical for-pay services. Make a clean space for testimonials from paying customers who are paying for Cloud Advantage for example. I still don’t understand your Cloud Advantage pricing model and its non-utility pricing structure..pretty much the opposite of how cloud providers are pricing their managed cloud services which leverage Ubuntu as the commodity OS. There are hints in the current case studies concerning landscape but its buried pretty deep under a lot of praise for gratis deployment Ubuntu. I’m particularly looking forward to seeing some testimonials from customers who purchase the Cloud Jumpstart service to roll out a private cloud. That seems to be the most relevant promising new for-pay service you have up on deck at the moment.

    What is your plan to go after certifications? For example, lets talk about the Common Criteria certifications such as any of the EAL certs. How are you even getting into the door for service contracts in large companies and government contracts without Common Criteria certs in hand? Is Canonical actively working to get through any of the Common Criteria processes? Honestly, without best of breed selinux support I don’t see how you get there with the certs that actually matter for government contracts. I’m not aware of anyone at all trying to make the case for apparmor based security in any security oriented certification process of merit. But hey, I’m more than happy to be proven wrong on that. And the security certification question in the cloud is still looming for the entire industry. I do not think Canonical is well positioned with its lack of commercial support for selinux to really go after contracts once the need for svirt and similar technologies are understood to be imperative. I know you guys have dabbled with making apparmor work with svirt…but generally speaking it seems you are making the minimal investment in apparmor to keep it relevant instead of really pushing it as competitive advantage in the markets you have an opportunity to monetize.

    The value proposition of anything that has no cost associated with is a bit tricky to evaluate. Dividing by zero is always fun. It’s either zero, finite or infinite. You can only know by looking at what happens if you ask people to incur a small finite cost instead of zero cost. I’m prepared to rationally argue that there are just as many people who run webservers in their business on Ubuntu who actually put zero value on it as there are people who put a finite value being able to use it. There I go again, thinking like a scientist…sorry. Good look monetizing your userbase.


  36. fcoir Says:

    always when i read Jef i remember jono’s post “i am jef spaleta” the herculean sagacious freedom warrior

    Seriously, it’s tiresome to read all of the long well written post of Jef attacking on one way or another to ubuntu/canonical (directly or on a more subtle way) and i agree with the others peoples comments on the article “Ubuntu in your pocket ->http://www.markshuttleworth.com/archives/1011” of mark about jef being a troll


  37. Clive Says:

    Sometimes we see fairly heated debates about whether or not GNU/Linux distributions should embrace the corporate world or not. My employer uses Linux for servers (RHEL) and Windows for desktops. Personally I don’t mind if Linux is used in both places, for 2 reasons. One, it helps to ensure that the OS remains flexible, evolving and improving. Two, think about the alternative for a moment…

    If we want to make GNU/Linux more readily acceptable to business [and to Home/SOHO users] then the focus needs to be on Administrative Tasks. Ubuntu has been incredibly popular at driving innovation in the presentation layer, but it so far has ignored some of the deeper structural challenges with this OS. So here’s our chance.

    What GNU/Linux needs is an opt-in strategy for service management on a Linux host… a transparent, open, flexible and dynamic framework, into which any software service can register it’s own management API. Appearing, for example, as tabs in a single service manager window, each tab would contain config information, daemon startup/shutdown controls, the works. This would bring together all of the individual components necessary to manage a software component, but delivered in an open framework with a simple API that the wealth of existing server side software could be integrated with.

    Give hard-pressed sysadmins in small and medium-sized [and large, for that matter] business the ability to control all their servers from a single GUI console, and all their favourite software packages from the same toolset, and you will conquer the desktop and massively simplify the support of Linux into the bargain.

    You have proven, with Unity, that you have the capability to deliver this scale of engineering. It just remains to be seen whether you would agree that this would make a difference, and whether you’ve got the appetite to go after it. But just think about such a project as a glue that could bring together all the disparate projects that live in the Linux eco-system – a chance for all of them to work together to make the best simplified, rationalised management utility conceivable.

    Make it tailorable for distros, skinnable, customisable by consumer organisations… Integrate it with software distribution and patch management of a server farm, and you’ll leave the competition so far behind they’ll need a telescope to see you…

  38. Jasna Says:

    Very interesting point of view Jef…

    Sorry for this…

    We can all number here what Mark should do/not do, but what’s the use of it?

    Ok, Maybe the use of it is that Mark can see other points of view..

    From the other side, our comments are pretty much useless here..No offense….

    Why do I think that?

    Because we are not in his position, we don’t see Canonical and Ubuntu from his side…Therefore we can only complain or defend… What is needed in every company to be done is to measure organization performances… I’m pretty much sure that Mark is aware of that — strategy goals (in metrics, not just pure words… Words mean little if you run a company), strategy plan, mission, vision, SWOT analysis — which has to be a good one.. .. Metrics and measurements is what counts alot….

    So, only Mark can do that because he sees the best which strenghts,weaknes’,threats and opportunites Canonical has…Therefore he can take certain steps due to strategy goals he has …

    So Jef, you have to think like a business man not scientist….

    There, I just threw my college gained theory knowledge here… Mark knows how it happens in reality….

    You can all call me a kiss ass or whatever you like (sorry on expression)… But I care about Ubuntu (I stepped with that distro into open source. Therefore I’m very attached to it.) I want to see it grow..

    Never stop fighting Mark. Continue with great work. World is cruel and you know it..When individual wants to accomplish something it’s like whole world fights you back.. You know it better than the rest of us.. You hold that pressure “survival of the fittest” more than any of us… I see it by watching your interviews, reading your blog… You can do it.. You’ve done alot so far specially for users of Ubuntu..That’s why my friends and I have a big respect for you and your work…

    Keep up,


  39. Red Hat falling behind Ubuntu in enterprise sector. | LORDSITHTECH Comtips&Technic, Network tips, Linux Says:

    […] Shuttleworth, creator of the Ubuntu, said in a blog post Wednesday that his open sourceoperating system is becoming more popular than rival Red Hat Enterprise […]

  40. Shuttleworth: Ubuntu rockets up at enterprises • The Register Says:

    […] data presented in a in a blog post by Ubuntu project founder and "benevolent dictator" Mark Shuttleworth is any indication, then […]

  41. Shuttleworth: Ubuntu rockets up at enterprises | UnixProfs Says:

    […] Ubuntu rockets up at enterprises 15 maart 2012By rssposterIf data presented in a in a blog post by Ubuntu project founder and “benevolent dictator” Mark Shuttleworth is any […]

  42. Benjamim Góis Says:

    That’s a very truth that ubuntu is growing Mark. I’ve being working in IT and Telecommunications for 5 years a see a tremendous move in a lot of datacenter from RHEL to Ubuntu/Debian. I use to see companys that develop huge billing systems for telefony carriers that certified RHEL for their software, maybe like you said “certification” is the way to go. I see the point of Jef, so far i didn’t see any of my clients thinking about getting a paid subscription from Canonical, they didn’t see advantages on that, but maybe that’s the next milestone/obstacle to Ubuntu. Jasna make a very good observations, we as users can only complain and “wish” that the system is not way we personal like, it might be a very dificult position from your part Mark and thinking a little about that, makes me regret some comments i made in a feel webblogs. You are doing a pretty good job Mark, ubuntu isn’t just another irrelevant Linux distribution in the world and you know that, i really hope precise helps it jump over the Geeky/early adopters.

  43. Benjamin Kerensa Says:

    @tom: You want to know why CentOS has popularity over Ubuntu (http://w3techs.com/technologies/details/os-linux/all/all) its because a good majority of shared webhosting servers and dedicated servers run parallels or cPanel which depend on CentOS because both companies have really not expanded developer their hosting platform for other distros dispute hosting companies requesting both Parallels and cPanel, Inc to begin support Debian-based distros.

    There is no doubt that that huge market share would prefer Debian/Ubuntu over CentOS if they could use the above mentioned platforms on Debian/Ubuntu.

  44. Ubuntu 創辦人 : 從市場份額上我們比 Red Hat 優勝 Says:

    […] Shuttleworth 在 w3techs.com 的網站之內便提供了一張聲稱是 Red Hat Enterprise Linux 與 Ubuntu […]

  45. Shuttleworth: Ubuntu rockets up at enterprises Says:

    […] data presented in a in a blog post by Ubuntu project founder and “benevolent dictator” Mark Shuttleworth is any […]

  46. techtings» Shuttleworth: Ubuntu rockets up at enterprises Says:

    […] data presented in a in a blog post by Ubuntu project founder and “benevolent dictator” Mark Shuttleworth is any […]

  47. Ubuntu: la sfida a Red Hat in ambito enterprise | oneOpenSource Says:

    […] altro punto importante, continua il patron di Canonical nel suo post, è la cadenza dei rilasci coordinata tra Ubuntu e OpenStack: “non una coincidenza, ma un […]

  48. Марк Шаттлворт заявил, что Ubuntu начал теснить RHEL в области крупных промышленных систем | PROUBUNTU Says:

    […] Шаттлворт опубликовал в своем блоге сведения, что все больше […]

  49. lzap Says:

    Enterprise computing is not equal to public web servers. There’s still a long way for Ubuntu to get there. But congratulations. Good work.

  50. mehdi Says:

    Ubuntu Server LTS or not is not supported by most of IBM (Software/Hardware), HP (Hardware), Oracle/Sun (Software/Hardware) and other big OEM so *IS NOT* comparable with RHEL or SLES in Enterprise Linux market…

  51. Mark Shuttleworth: o Ubuntu está crescendo e passa Red Hat | Blog Seja Livre Says:

    […] Shuttleworth , em seu blog oficial, declarou que não só o ambiente “home” está crescendo, mas também a adoção do Ubuntu em […]

  52. mark Says:


    Yes, we’re agreed, this stat is just a lightweight taste of how companies think about their public-facing, less important servers. The picture would be different if we could peek into the guts of the data centre. Nevertheless, since we don’t have reliable data for what happens behind the firewall, this is valuable even accepting and acknowledging its limitations.


  53. mark Says:


    Yes, wouldn’t it be great if that changed?

  54. Andy Says:

    @Benjamin Kerensa
    I think you’re wrong – the reason that many companies use CentOS rather than Ubuntu isn’t to do with cPanel, it’s because customers want the stability and features of RHEL but want it for free.

  55. josh Says:

    I permanently moved all of my servers and workstations from centos to Ubuntu couple months and I made a damn good decision. Ubuntu Lts is really stable, that is, perfect for mission-critical environments.

  56. Pieter Says:

    A bit odd to call the article “Ubuntu vs RHEL in enterprise computing” and then compare Public Web Services. They are hardly the same. Anyone in the Large Enterprise business knows that there are a zillion more servers behind the firewalls, all invisible from the outside World. And the majority of those servers run Red Hat Enterprise Linux (& CentOS) and to some extent AIX, HPUX, Solaris and zOS. Given this reality this post seems wishful thinking.

    I can imagine that Mr. Shuttleworth is enthusiastic about OpenStack as indeed it presents an opportunity for Canonical. But I wonder how big that opportunity is. Red Hat participates in the development of OpenStack and OpenStack is already available on RHEL6 (CentOS6) and Fedora. This means the RHEL/OpenStack combo is on the table too as a solution. Where’s Canonical’s USP that will make the Enterprise choose Canonical over Red Hat? With RHEL (& CentOS) already deeply entrenched into the Large Enterprise, will Canonical be able to convince the Large Enterprise that they are better than Red Hat? That’s a tough sell. For one, Red Hat’s Support is unmatched and one of the most important reasons for the Enterprise to choose Red Hat year after year. OpenStack may be available for Ubuntu but does Canonical’s support match Red Hat’s? Time will tell. I look forward to see the best solution win.

  57. Andy Bailey Says:

    Without a Common Criteria certification or FIPS 140-2 validation, I can think of one really big enterprise in which Ubuntu is a non-starter.

  58. Shuttleworth: Ubuntu beats RHEL | Install Ubuntu Says:

    […] Shuttleworth happily reports that companies have started adopting Ubuntu over Red Hat Enterprise Linux for “large-scale […]

  59. Jef Spaleta Says:


    I’m sure Mark can correct me if I’m wrong about this… But Canonical’s concept of the “partner” repository as originally envisioned was meant to provide exactly the enterprise level application validation from the first LTS release onward.

    I believe the idea was Canonical charged those application developers money to be included in the partner repository so Ubuntu users could gain easy access to the software. It’s not really clear what the business arrangements were with different partners, but I would not point to the partner repository concept as a successful approach (regardless of the number of glowing press releases generated when something was being prepped to enter the repository.)

    Mark seems to be hinting that Canonical is ready to try something different. It would be nice to see him articulate how this new approach is going to be better than the previous partner concept and to address some of the lessons learned in its failure to generate interest from enterprise application developers.

  60. yman Says:

    I think we should all watch this TED Talk:


  61. Adam Williamson Says:

    Mark: so, you admit Jef’s points, but you significantly don’t produce the data he asked for – how many of those Ubuntu deployments are *Canonical-supported, revenue-generating* Ubuntu deployments? Because that’s the _accurate_ comparison to RHEL. No-one, but no-one, is running their public-facing web server on a non-supported – i.e. non-paid – RHEL. At least, I dearly dearly hope they aren’t. However, it seems reasonable to assume that a lot of the Ubuntu deployments are unpaid ones.

    There’s nothing _wrong_ with that, but it’s clearly somewhat unhelpful to compare paid with unpaid deployments.

    (For anyone playing along at home who doesn’t know – I work for Red Hat. Don’t trust me. They sign my paychecks. :>)

  62. Adam Williamson Says:

    For giggles…


    Operating Systems


    (used until recently)

    Yeah, I don’t run my site on RHEL either. =)

  63. Greg Says:

    HA! This is the best post I’ve read in a month and I barely got through the first sentence before I knew it… Every time you write a new post, I think, I should work for this person.

    Linode, the fastest growing VPS provider, is solidly Ubuntu: http://www.linode.com/about/ up from 48% Ubuntu to 57%! RHEL doesn’t even make their list.

  64. Догоним и перегоним в исполнении Шаттлворта - Linux в Беларуси Says:

    […] своем блоге Марк Шаттлворт опубликовал статью в которой он показывает как Ubuntu обходит RHEL как основа […]

  65. Martin Says:

    Dear Marc,

    please do your maths: When talking about RHEL please don’t forget to add CentOS. And even then you compare apples and oranges: We have a free operating system without support on the one side and a operating system with support on the other side. The chart would be more interesting when comparing supported platforms with unsupported ones.

  66. Ubuntu Is Spanking Red Hat, Says Ubuntu Founder | Install Ubuntu Says:

    […] a blog post on Wednesday, the founder of Ubuntu and its commercial steward, Canonical, cited numbers from web […]

  67. Ubuntu vs RHEL 服务器市场谁更符合潮流? | 极地 EA163 Says:

    […] 原文链接,OSChina.NET 原创编译 此条目是由 Wu 发表在 IT资讯 分类目录的。将固定链接加入收藏夹。 […]

  68. Frans Thamura Says:

    all meruvian using ubuntu in desktop 100%.. no windows, except my mac (1 mac).. and the server all centos .. i am also forgot, how to install RHEL :) … we try openstack, right ubuntu in server….

  69. mark Says:


    The maths are quite sound. Ubuntu is available with full commercial support. When you factor in CentOS, Fedora and RHEL vs Ubuntu and Debian, for a whole-ecosystem view, the trend is just as pronounced.

  70. Adam Williamson Says:

    mark: The ‘ecosystem’ numbers don’t seem particularly indicative of anything much. As others have pointed out, it’s somewhat hubristic to describe Debian as part of the ‘Ubuntu ecosystem’; if anything, it’s the other way around. I’m not sure what comparing Debian+Ubuntu numbers with Red Hat+CentOS+Fedora numbers tells you about Ubuntu and RHEL; you can’t assume that there’s really any particular relationship between a Debian deployment and Ubuntu / Canonical, or, honestly, between a Fedora deployment and RHEL / Red Hat.

    The comparison you claimed to be making is ‘Ubuntu vs. RHEL in enterprise computing’. You’ve acknowledged the rather large shortcomings of the numbers you cite, the most significant being that you continue to refuse to split the Ubuntu numbers into paid/Canonical-supported and unpaid/un-Canonical-supported. Yet at the same time you seem to continue to place much more weight on the problematic numbers than they seem remotely capable of supporting…

  71. Jef Spaleta Says:

    Mark is fundamentally wrong to imply that somehow Ubuntu is taking marketshare or clients away from RHEL to any extent. Certainly the data doesn’t support it. But the fact that he’s taking potshots at RHEL says a lot of who he thinks Canonical’s competitors actually are now. Hint: It’s not Microsoft. He hasn’t talked about fixing bug #1 in years now. He’s setting his sites a little closer to home (again) and trying to chip away at marketshare from another linux vendor. He need to learn to avoid drawing direct comparison to other linux vendors. He could have just noted the upward trend of Ubuntu alone without commenting on “the enterprise” or RHEL and muddling the message with overreaching propoganda. The magic art of the underpromise/overdelivery seems to be lost on him.

    Though it is useful to run a series of thought experiments concerning the actual state of the webserver marketplace.

    For each of the gratis distributions if it became for-pay-only…what would happen to the market dynamic?
    For each of the for-pay distributions if it became gratis deployment…what would happen to the market dynamic?
    For each distribution if it stopped being produced entirely…what would happen to the market dynamic.

    When I run through the full set of such thought experiments..here are the conclusion I draw.
    When choosing a webserver distribution people make 2 choices in this order.
    1) Do I need support: If yes then SUSE or RHEL
    2) Which gratis deployment option do I use: CentOS,Debian,Ubuntu

    The reality is Ubuntu is more likely than not taking webserver marketshare that would otherwise go towards Debian if Ubuntu wasn’t offering a gratis server instance. The real important insights in the trendlines are really about the Debian and Ubuntu relationship. Ubuntu and Debian and pretty much interchangeable for gratis server deployment needs are they not? The real question of merit is not why are people choosing Ubuntu. The real question is why are those people not choosing Debian. I’d wager the answers you get are pretty orthogonal to the identified needs that paid support have successfully targeted for SUSE and RHEL businesses. Canonical needs to understand why people are choosing Ubuntu over Debian and build a compelling business around those reasons.

  72. Jasna Says:


    Canonical will understand why people are choosing Ubuntu and what do they want with it… Just wait for survey results 😉

  73. Jef Spaleta Says:


    There are already published survey results. I have read the pdf announced on Feb 14 on Canonical’s blog. I am a voracious reader of pretty much everything Canonical decides to make public. I make every effort to be as informed as I can be about any topic ahead of engaging with Mark directly. If I am misinformed it is only because Canonical has declined to provide the information and then declined again to answer questions I. I am more than happy to integrate new sources of information when they become available and are adequately explained as to how they were captured.

    Sadly none of the survey data that Canonical has decided to make public speaks to commercial support uptake as far as I am aware. The latest survey summary does talk about the types of community support people are using interestingly enough. But nothing public in the survey results about commercial support interest. That omission is very very interesting to me. The omission itself screams at me in a way that a muted purple and grey bargraph can’t.

    Either Canonical is afraid to ask people to comment on commercial support in the user survey, or they are afraid to make the answers concerning commercial support public. Either way it doesn’t really look good when Canonical simultaneously lifts up Ubuntu as a commercially supported offering that is winning against RHEL but at the same time declines to publish survey results concerning that very same commercial support that the tout. They go through all the trouble of making the survey and publishing the rah-rah results..and nothing about commercial support. Message received, loud and clear. The Ubuntu server using community..whether it be traditional server usage or cloud workloads doesn’t want or need or care about Canonical’s support offerings. But even still Canonical needs to try to play this game of implication that somehow Ubuntu usage has a hope making the leap into a business model for Canonical support…just need enough of the market right….just a enough and boom..paying customers. Total fantasy. Canonical needs new executive blood to chart a sustainable course for its business.


  74. Mike Says:

    Ubuntu is becoming better than ever!!

    Thank you Mark and the Ubuntu community for this great work.

    I’m also actively learning Python, to make open-source extensions for the SchoolTool application funded by the Shuttleworth Foundation.

    This is improving IT technologies from all over the world, more effective than the work of Bill Gates Foundation, because this is IT in human centered design.

    Kind regards,


  75. Jasna Says:


    Can you give me the link of results here ?

    I do follow almost everything related to Ubuntu via facebook, google+ and twitter.. Haven’t seen the notification that results are done.. Last week was announced that filling the survey will be prolongated..

    Maybe I’ve missed them… I’d like to see them…

    Look, I think you’re a bit harsh with this commercial and money thing…. From the outside I can conclude that Canonical’s strategy goal is in the area of users… Didn’t get the picture that they have strategy goal in the area of finance yet. I didn’t see that Mark would emphasize it out very strongly as users…

    It’s up to him what he wants to achieve and in what direction he wants to take Ubuntu.. Due to that he’ll take certain steps to achieve it…

    You can’t expect something will happen over a night… You can’t twist a firm upside down over a night….

    It’s funny for me that you comment what Canonical needs and you don’t run that company shoulder to shoulder with Mark… Maybe I’m to harsh now…

    I know we can all write here…And it’s great that someone like Mark actually want’s to hear us…

    But a little objectivity wouldn’t hurt right?

    I mean we all have to be aware that this is Internet and a lot of things can be misunderstood right?

    So, I didn’t get Mark’s post like he was claiming some *enterprise* victory over RHEL.. I got it like Linux patterns are changing… by enterprise I got it like *different firms are developing different distros*

    Understanding something over Internet can be difficult… You can see that by only reading our comments here…

    One side got it like I did… the other side got it like you did – commercial stuff and money….

    At the end I started to laugh actually, because this somewhat turned into a pissing match…. 😀

    Must say in the end … You should put yourself into Mark’s shoes for one day and then you’ll see how in a difficult position he is just like anybody else who runs a company and project of Ubuntu’s magnitude…

    This subject what one company should do or not do is very case sensitive… Yes Mark can see our comments… maybe even see them as usefull maybe not… But he’ll draw the line, he has to…That’s why I was a bit harsh by saying on my previous post that our comments are pretty much useless, because we don’t run this company nor project…

    At least that’s what I think and see it….



  76. Jef Spaleta Says:



    Over night? of course I dont expect anything to happen overnight. Canonical however has been incorporated as a business now for 8 years. And all along those 8 years Canonical has been trying to break into the enterprise.
    Let me give you a partial summary of the history of merit here:

    2006, Canonical’s now defunct Ubuntu Professional Certification was marketted to “enterprise” customers…mimicking RHEL’s professional training concept.
    2006 Canonical commits to support Sun SunFire servers using the Sparc arch. Lots of blah-blah about enterprise hardware. Drops the arch entirely in 2008
    2007, Landscape was original marketted as “enterprise-ready” management…basically mimicking RHEL’s rhn concept.

    8 years… multiple attempts to fire up some sort of enterprise revenue stream…and here we are again. ssdd. Except this time its the Cloud, and in the Cloud each every one of the hosting providers who are leveraging Ubuntu deployments are going to absolutely _crush_ Canonical’s hopes of support revenue.

    Mark completely misses the core problem he is going to have to face in the Cloud. Its not Red Hat, its not RHEL. Its every single hoster he wants to partner with to make Ubuntu ubiquitous in the cloud right now is the future competition that is going to bleed Canonical dry of revenue. Those hosters are going to provide bundled services to support Ubuntu at price points that Canonical simply can not compete with. Moreover, because customers are already paying those hosters money there will be an established vendor trust relationship. Even if canonical’s support is better (it wont be) those hosters already have a direct line to the wallets of their customers for value-add.

    So yes Ubuntu will dominate the cloud in terms of deployment numbers, maybe surpassing Debian maybe not. And yes people will profit from supporting Ubuntu. But it will not be Canonical. It with be Rackspace and DreamHost and Linnode and even the 500 lb gorilla that is Amazon..they will all roll out support bundles as part of managed solutions built on Ubuntu (and Debian) which trickle back absolutely no revenue to Canonical.

    The potshots at RHEL that Mark is doing right now are just wasting ammo. Ammo he’s going to need for the real fight is ahead. And its going to be bloody horrible mean fight..a civil war of sorts…because Canonical is going to be fighting tooth and nail for revenue with companies that are ostensibly partners in driving the Ubuntu server adoption right now. Rackspace as the driver in openstack is going to pummel Canonical for support revenue to support Ubuntu instances. And I fully expect Rackspace to reach out from a position of revenue strength and start diving into the on-premises private cloud infrastructure market. There is no bright spot for Canonical’s cloud strategy I am afraid.


  77. darislav Says:

    Greeting Mark!
    I really want to know what you think about an idea: not only rip the UI from MacOS but sell similar computers too, like make Canonibook with beautiful design and Highly Optimized (for Ubuntu) hardware?

  78. Jasna Says:


    Thanks for the results…

    Wow your post… what can I say.. to much negative stuff for me :(

    Hmm …How come you don’t write one, just one positive sentence… When I’m reading your posts I have a feeling like Canonical can’t survive for a day.. I’m not attacking you.. I just get that vibe…

    I know being coutious is ok … but looking everything from the negative side is not my kind’a thing…

    And I’m still not getting the vibe *Canonical being in trouble*….nor that they’re a copy of another company… nor that their main target are enterprises…

    I still see, from the outside, that users are the main target… especially users on the proprietary side of a fence…

    As I wrote here already, I don’t see Canonical with Mark’s eyes… Therefore I can just superficially comment here….

  79. Jasna Says:

    Looks like my previous comment disappeared or Mark erased it 😀


    That’s not the survey I meant… I meant the Ubuntu user survey :)

    I guess results are still not done :)

  80. Jasna Says:

    My bad it’s back 😀 (some magic happens here 😀

  81. Ubuntu batte Red Hat in Champions League! | Luca Ferrari Says:

    […] il tutto accompagnato dall’ormai consueto post auto celebrativo in cui si elogiano le grandi caratteristiche del sistema e il grandioso lavoro […]

  82. Ubuntu 12.04 ma być systemem dojrzałym także dla serwerów | Adam Grzankowski Says:

    […] Mark Shuttleworth napisał na swoim blogu, że Ubuntu 12.04, dzięki rosnącej grupie osób uczestniczących w […]

  83. Ubuntu 12.04: A ‘coming of age’ on servers too, Shuttleworth says | Install Ubuntu Says:

    […] which was designed to span the full range of personal computing needs,” he wrote in a blog post on […]

  84. InterKnight Says:

    I have used Ubuntu for nearly seven years now, and I must say that I feel that it is probably the best Linux distribution to date…for a sighted user. Sadly enough, however, Ubuntu has had a number of releases which worked perfectly for sighted users, but were extremely broken for visually impaired users. This also seems to be the case for the upcoming Ubuntu 12.04. If the problems which exist for blind users also existed for regular users, the release would be put off until things were rectified. I still use Ubuntu, and probably always will, because it is what I know. But I must say that I do not feel at home in my system any more. Do I think Ubuntu could take a great lead in the enterprise? Most definitely. It has a lot to offer. But I think that a lot of work needs to go into the accessibility of the system. Despite all of the setbacks in accessibility, I do appreciate all of the work which is put into Ubuntu. The Accessibility Team also does a wonderful job, but it does not seem that accessibility is an important factor to Canonical. How does accessibility relate to this post? Well, as a blind graduate of an IT-related program, I feel that Linux is the best way to go when it comes to business. If, however, blind professionals cannot fully access their systems, thee can be no progress. I’m sorry for the rant, but myself as well as other blind users have been getting tossed aside since the release of Ubuntu 11.04. I want to feel at home in my system again. I think Unity is great, but as a blind user, it is not completely ready for production use. All the best.

  85. Shuttleworth: Ubuntu Mengalahkan RHEL | LinuxBox.Web.ID Says:

    […] Shuttleworth dengan bangga melaporkan bahwa perusahaan-perusahaan mulai menghadopsi Ubuntu dibandingkan Red Hat Enterprise Linux untuk […]

  86. Shuttleworth: Ubuntu Mengalahkan RHEL | LinuxBox.Web.ID Says:

    […] Shuttleworth dengan bangga melaporkan bahwa perusahaan-perusahaan mulai menghadopsi Ubuntu dibandingkan Red Hat Enterprise Linux untuk […]

  87. ¿Ubuntu mejor que RHEL en servidores? Says:

    […] este caso el señor Mark Shuttleworth quien ha escrito un artículo (que se ha sacado debajo de la manga) donde compara a Ubuntu Server con nada mas y nada menos que […]

  88. kjh Says:

    What concerns me most about Ubuntu is that I have been seeing a new class of Linux ‘Admins’ having the same unremarkable level of experience as the most recent generation of Windows Admins:

    I run Windows on my home PeeCee. I know how-but-not-why to click-click-clickety-click. I think I am a Administrator, therefore I am a Windows Server Administrator.

    now take that statement and: sed -e ‘s/Windows Server/Linux Systems/g -e ‘s/Windows/Ubuntu/g’

    — kjh

  89. Mark Shuttleworth, Ubuntu Server e le statistiche rispetto a Red Hat | Indipedia – Indipendenti nella rete Says:

    […] Via | Mark Shuttleworth […]

  90. Las empresas prefieren Ubuntu a Red Hat | TechWeekEurope España Says:

    […] en lo que sabemos de la nueva modalidad de servicios, como el cloud computing y el Big Data”, comenta […]

  91. Ubuntu, le aziende migrano in massa? | infropy - information entropy Says:

    […] brillerebbe in particolare nei confronti di Red Hat Enterprise Linux (RHEL), dice Shuttleworth, citando i dati di W3techs: tra marzo 2011 e marzo 2012, Ubuntu avrebbe superato Red […]

  92. Shuttleworth: Ubuntu rockets up at enterprises | DailyTechWire Says:

    […] data presented in a in a blog post by Ubuntu project founder and “benevolent dictator” Mark Shuttleworth is any […]

  93. Ubuntu ultrapassa o Red Hat Enterprise Linux | DPKG Says:

    […] Shuttleworth demonstrou entusiasmo em seu blog pessoal ao relatar que empresas começaram a adotar a distribuição Ubuntu, ao invés […]

  94. j03 Says:

    All you Ubunutu haters are funny. You’re just mad because your peepees are getting smacked down. Ubuntu should be hailed and respected – for bringing an attractive open source desktop OS to the masses and away from the destructive corporatocracy which has made them dumb sheep-like beings for so many years :)

  95. Lilian Says:

    If Ubuntu will take over RedHat customers that will be very bad news. 60% of the Linux Kernel is written and maintained by RedHat, a bunch of good software is written by RedHat and they contribute it to the Open Source World. While Canonical… If RedHat looses customers then it won’t support Linux so much then Canonical wouldn’t be able to say that look how good our operating system is because almost all the stuff they are using is actually RedHat’s work… Bye, bye RedHat equals Bye, bye Ubuntu.

  96. mark Says:


    Customers who want work done on the kernel, and hardware providers who are shipping new silicon and want a kernel to show it off, will always drive kernel development. Canonical is just as effective a conduit for that as any other company, so I don’t think your fears will materialise.

  97. Andrew Ampers Taylor Says:

    @mark @lillian

    Life always goes on. There are so many examples in this world where people have thought something has ended, but on it goes. Mark is right when he says your fears won’t materialise.


  98. fossfreedom Says:

    Mark – its not clear what the LTS upgrade strategy is beyond 12.04.

    i.e. – could a 12.04 business jump to 16.04 directly – i.e. plan to upgrade in 4 years time? If a 12.04 user keeps going for the full 5 years – can we assume they need to first move to 17.04 … 17.10 and then 18.04 LTS – doesnt sound ideal for business users.

    Any chance of clarification?

  99. mark Says:


    We support LTS-to-LTS upgrades, so yes, you could stay on 12.04 LTS, then move to 16.04 LTS (in our current testing plan we would test the migration from 12.04 to 14.04 and then to 16.04, not directly).