Linux Foundation’s Latest Survey

It’s about large business’ use of GNU/Linux. While that’s a niche market for GNU/Linux, it’s large and growing:

  • 8 out of 10 “inivtation-only” respondents plan to spend more on GNU/Linux servers in the next five years while only 2 out of 10 plan to do that for that other OS,
  • twice as many plan to use GNU/Linux rather than that other OS to deal with “big data” issues,
  • those with technical problems preventing adoption of GNU/Linux dropped 20% compared with 2010. Now only 12% find technical obstacles,
  • there has been an increase in contributions of code and bug reports,
  • more than 2/3 see GNU/Linux as more secure than other operating systems.

see Enterprise Linux User Report 2012

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology. Bookmark the permalink.

25 Responses to Linux Foundation’s Latest Survey

  1. ch wrote, “Please allow me to repeat: “is not intended to be an assessment of the overall penetration of Linux in the market””.

    We know who is a member of the Linux Foundation. Are any not unrepresentative of large businesses? There’s just about every kind of business in there. Of course, the survey was not a market study. It studied users and there are a lot of them. The message is that big business can and do use GNU/Linux in spite of twits who comment here and write that that does not happen.

  2. ch says:

    “I suppose those who had above average uptime.”

    Unless you have a cluster running, a huge uptime simply means you don’t do updates – something I wouldn’t want to boast about.

    “That site is biased beyond belief.”

    Right. Worse, he is biased in the opposite direction of your bias. But can you prove him wrong ?

    “Hundreds of large businesses contributed input to the survey but somehow the anonymous authour finds it biassed.”

    Let’s not haggle about “large” and “500 employees”, but a survey that is based on invitation-only responses by the Linux Foundation does smell biased, and at least the LF says itself:

    “The Enterprise Linux User survey is not intended to be an assessment of the overall penetration of Linux in the market, nor is it designed as a cross-platform study. It’s also important to note that enterprise users who are motivated to take a survey from The Linux Foundation are not an unbiased lot, but the size of these organizations, their buying power and technical prowess – as filtered by The Linux Foundation and Yeoman – can provide important guidance both for Linux vendors and developers, as well as their competitors.”

    Please allow me to repeat: “is not intended to be an assessment of the overall penetration of Linux in the market” and “enterprise users who are motivated to take a survey from The Linux Foundation are not an unbiased lot”. In other words: “Please be careful what you read into this report.”

  3. Dr Loser wrote, “However, if you combine this with the 25% response rate, I submit that the difference between the two terms is negligible.”

    Of course no one is compelled to respond to a survey…

    Whether or not the survey has a narrow response, hundreds of largish businesses made those responses so something is happening and it’s an indication that GNU/Linux is mainstream and increasing in adoption. Is there any other viewpoint that is consistent with the results? I doubt it. Some surveys are glad to have a 10% response rate. 25% could indicate joy in the users’ hearts. The results are certainly consistent with the benefits of GNU/Linux I have seen in my small organizations.

  4. Someone wrote, “What IT professional from a large enterprise would in this day and age boast about uptime?”

    I suppose those who had above average uptime. Certainly no one boasts about downtime… oldman even demands SLA (Service Level Agreements) so uptime is often a part of contracts for service/maintenance.

    That site is biased beyond belief. Hundreds of large businesses contributed input to the survey but somehow the anonymous authour finds it biassed.

  5. Dr Loser says:

    As a mild correction, I now recall that the phrase was “invitation only,” not “self-selected.”

    However, if you combine this with the 25% response rate, I submit that the difference between the two terms is negligible.

  6. Dr Loser says:

    @oiaohm:

    “Its a way of collecting a randomize sample.”

    I don’t know precisely what mathematics you would use to collect a randomized sample, young man, but in standard statistical terms, “self-selected” does not constituted randomization.

    Sodomization, possibly. Randomization, not.

  7. Clarence Moon says:

    The proof of the pudding lies in the eating, it is said, and Microsoft’s most recent results show a significant rise in their server OS business. FRom the cite found elsewhere on this blog:

    “The Server & Tools business posted $4.77 billion in second quarter revenue, an 11% increase from the prior year period, reflecting double-digit revenue growth in Windows Server and SQL Server premium editions and more than 20% growth in System Center revenue.”

    Hear all the hopeful chants you want from the Linux preachers who lead your choir, Mr. Oiaohm, but eventually services will be over and you will be back in the real world of balance sheets and quarterly reports.

  8. oiaohm says:

    Hanson if Steve Ballmer did write a book on Linux it could be a really interesting read.

    Opps typo 12000 should have been 1200+ resonance. They were looking for parties large enough.

    “This year’s survey was conducted in partnership with Yeoman Technology Group late last year and received responses from nearly 1900 individuals”

    Basically if you are not from a large enough company for that survey you data is dropped even if you respond. Ie you must have over 1/2 a billion in income or 500 employees. Anything smaller is not for that survey. Yes they must have more than 400 respondents a year that qualify.

    Call it random testing. How many companies do we have to ask to get 400 found.

    Enterprise Linux User is the big parties. Latter on in the year there will be smaller Linux users reports. The under 1/2 billion and 500 employees is always the more interesting report.

    Its a way of collecting a randomize sample.

  9. Hanson says:

    “So from 12000 responses. 438 are pretty huge setups.”

    And you pulled the number 12000 out of thin air? Besides, I found the number of non-respondents:

    “The Linux Foundation, in partnership with Yeoman Technology Group, recently conducted an invitation-only survey of 1893 enterprise Linux users.”

    ARE YOU KIDDING ME? 428/1893 = 0.226…

    You get a response rate of less than 25%, and you claim that your report has any meaning at all?

    And let’s not forget: this is already a highly specialized, pre-selected group. We’re not talking about the public where such a low response rate could be acceptable.

    “Besides Linux Foundation is not about FOSS Hanson using closed source is perfectly permitted.”

    The same old excuse. Would you buy a book by Steve Ballmer about Linux, even if it’s very well written?

  10. oiaohm says:

    “identified by 428 respondents at organizations with $500 million or more in annual revenues or greater than 500 employees.” So from 12000 responses. 438 are pretty huge setups. Please note there is only allowed to be 1 respondent per company. Person must be confirmed as working for that company and the value of that company has to be confirmed.

    So you are not talking individuals here.

    Besides Linux Foundation is not about FOSS Hanson using closed source is perfectly permitted. Linux Foundation has very little interesting in smaller businesses at this stage.

    Its all about getting the resources for future development.

  11. Clarence Moon says:

    “Who would want that?”

    All the customers of those who sell them, Mr. Pogson. You seem to see the world as a place where it is good to avoid spending money and shake your head over the folly of paying someone for a product or service that you can provide for yourself. That is a noble image in romantic circles and has been the sort of thing that inspired the Marlboro Man and Crocodile Dundee (to give Mr. Oiaohm his thrill) images as role models.

    But it makes for a lousy business and the business of selling server hardware and server OS software as well as server support applications is the focus of a lot of much more down to earth types who are only interested in making a buck. The cheapskates who make servers out of junk box parts and Linux freebie distributions are not really of concern to those grinds who look constantly for the money flow.

  12. Dr Loser, explain how IBM or Cisco or any other member of the Linux Foundation is not representative of mainstream industry, especially when it’s about big business?

    I personally don’t feel the Linux Foundation is representing Mom and Pop small businesses, nor desktop users but they certainly are mainstream with the big guys. Whoever is making a lot of money in IT is a member, even “partners” of M$ like Yahoo! They “invited” hundreds, far more than just their membership.

    “This year’s survey was conducted in partnership with Yeoman Technology Group late last year and received responses from nearly 1900 individuals. Like the 2010 report, it filters the data to surface trends among the world’s largest enterprise companies and government organizations – identified by 428 respondents at organizations with $500 million or more in annual revenues or greater than 500 employees. The number of respondents from the largest enterprises in the world grew by more than 10% this year, providing an even better picture of enterprise Linux adoption trends. “

  13. Hanson says:

    Finally! A report you can trust. Made by the trustworthy Linux Foundation who creates its public documents with proprietary Adobe software. And only managed to get answers from 428 so-called IT professionals. The much more interesting number is deliberately omitted, the number of non-respondents. That’s how transparency looks like.

  14. Dr Loser, the point of the survey is not to quantify stuff but to see which way the wind is blowing. It’s not blowing towards M$ but away. More of a good thing is a good thing and no number is required to understand it.

    The big takeaway from this survey is that GNU/Linux, on the server at least, is mainstream and businesses from every area of industry are using it widely. There is a clear preference for GNU/Linux which is something that I advocate for all IT. Make FLOSS the default. Only in very unusual cases should it be necessary to use that other OS or apps specific to it. I find it rather interesting that my experience in small organizations in the bush should be consistent with wealthy corporations operating globally. There must be a common factor. I think it must be price/performance. It just doesn’t make sense to go to a store and demand to pay more for a commodity like bread or rice or beans but that’s what people do when they specify that other OS.

  15. Dr Loser, in a pissing contest, wrote, “Mine was an Oxford BA in Modern History”.

    Ah! That’s the problem then, innumerate. I have several degrees all requiring heavy mathematical skills and attention to detail.

  16. Dr Loser, failing utterly to understand what a server is, wrote, “Good Lord, man, why?” customize servers.

    A server needs apps. Apps need to be configured. You have to set them up with names of things. M$ doesn’t know the name of my blog or the policies I want on it. M$ isn’t going to generate content for me either. This blog has tens of megabytes of text on it none of which were present on the bare metal of the machine nor the software as installed. That’s a huge amound of work and the installation of OS and application is insignificant.

    I have thousands of choices to make when I set up some servers. Even a simple web server needs to have a fixed IP address and a name if I am going to use it humanely. For some servers that is enough but almost any server doing anything important will have security settings, services started, applications installed, configured and fed content. Then the whole thing needs a backup. No OS as installed is likely to know much about what I want the server to do. That’s true for GNU/Linux or that other OS and is quite different than for a personal computer which may be filled with generic apps and the users take care of providing the content. A typical desktop application needs to know nothing of the network on which it sits that’s all taken care of by DHCP and the OS. On a server, however, you might have Apache running a dozen web sites and each one might be serving different clients on different subnets under different domain names. That all has to be configured by someone.

  17. Dr Loser says:

    Remind me again what degree you took? Mine was an Oxford BA in Modern History, which involved no mathematics whatsoever, and even I can see the fatuity in this disinformation.

  18. Dr Loser says:

    And just to reinforce the incredible foolishness that this quote represents: you have two different contexts for “more” here. Neither one is quantified; both are ranks and nothing but ranks.

    What if the 80% are planning to spend 1% more, whereas the 20% are planning to spend 5% more? I mean, it could be the other way around. Who knows?

    But even given the notorious bear-pit of selection bias, you have for some reason chosen to fall into the even deeper bear-pit of comparing two ranking systems that are not, alas, mathematically comparable.

  19. Dr Loser says:

    @Robert:

    And, just to reinforce your obvious lack of statistical methodology, I will repeat:

    “Good grief. Even large multinational pharmaceutical companies are a little more rigorous in their methodology.”

    I don’t care what the individual or cumulative earnings of the respondents were (unless, by some mischance, they represent 95% of global GNP, in which case they are statistically a certainty).

    “8 out of 10 “inivtation-only” respondents plan to spend more on GNU/Linux servers in the next five years while only 2 out of 10 plan to do that for that other OS.”

    Are you actually trolling yourself here? Because it sure looks like it. Ever heard of selection bias?

  20. Dr Loser says:

    @Robert:

    You have to “customise your servers heavily?”

    Good Lord, man, why?

    It’s people like you who give Linux a bad name on servers.

    Buy a Windows W2K 2008 Server, and it Just Works. No customization required!

  21. Dr Loser mumbled about “totally disinterested sources“.

    Chuckle. I agree, but then usage of Linux is so mainstream that it is not unreasonable. I expect the Linux Foundation just polled their members like Fujitsu, Hitachi, IBM, Intel, NEC, Oracle, Qualcomm, AMD, China Mobile, Cisco, Huawei, Etri, Google, HP, Motorola, Netapp, Nokia, NYSE, Panasonic, Toyota and a host of others. I would say they are a representative set of big businesses which was the intended target of the survey.

  22. Clarence Moon went on about servers writing “according to sales surveys.”

    Yeah. As if I have ever bought a server running GNU/Linux… I have made servers out of PCs, paved over that other OS on servers but I have never bought a server running GNU/Linux from the factory. Who would want that? I have to customize it heavily anway to match my network and services so installing the OS is small part of the work. The same would apply to anyone replacing the hardware of a GNU/Linux server. They might just move the drives into the new server or copy software and data over the network. That’s easier with GNU/Linux than with that other OS because most drivers are a part of the Linux kernel.

    Yep. Tell us the number of units at each selling price and the OS and you might have some useful data but bundling the OS with the hardware makes the numbers almost useless. How about those CALs, eh?

  23. Clarence Moon says:

    Some 20% to 25% of the servers in use in the world are Linux based, according to sales surveys. Roughly 20% of the companies surveyed by the Linux Foundation replied to the survey request. Seems to correlate fairly well, eh?

    The Linux Foundation does warn:

    “It’s also important to note that enterprise users who are motivated to take a survey from The Linux Foundation are not an unbiased lot…”

    One can only imagine the glee that Mr. Pogson would show upon being offered the opportunity to wax about his Linux preferences. Doubtless the same affliction is present for others operating Linux installations.

  24. Dr Loser says:

    I love it when you quote totally disinterested sources, Robert.

    “8 out of 10 ‘inivtation-only’ respondents plan to …”

    … what, eat their cat? No, wait, I’m quoting from the wrong commercial.

    Good grief. Even large multinational pharmaceutical companies are a little more rigorous in their methodology.

    Keep taking the Ten Tablets.

Leave a Reply