- Site menu:
- Latest Discussions
- Latest Newswire
- Archives
- Submit Story
- Story HOWTO
- Syndicate
- E-mail Alerts
How Can Linux Market Share Be Accurately Measured?
Quoting from the article: IDC analyst Al Gillen pointed out that the number of servers shipped does not perfectly equal the number of operating systems in the market. This is particularly the case with Linux where a substantial portion of the overall market opportunity comes from deployments aboard recycled servers, PCs and workstations deployed as servers, and Linux deployed as a guest operating system. I should note that eWeek often functions as a cheerleading section for Microsoft. Take, for example, the links to other articles with such catchy titles as:
It is not surprising that Mr. Gillen’s statement and that the article as a whole tries to minimize the fact that sales figures are not an accurate way to measure Linux market penetration. What is surprising is that the article mentioned this at all. In 2004 and 2005 I did five months of contract work for a large insurance company in Cincinnati. Their methodology for server purchasing is one I’ve seen time and time again in corporate America. They would underestimate their needs or else wait until the last minute to make purchases. They bought their servers based on what was in stock or what they could get a corporate discount on. They didn’t care if the servers were going to be used for Windows or Linux. They took them with whatever was on them, which was always Windows, and then promptly erased the hard drives and loaded a corporate standard image of whichever OS they wanted. Each and every server they loaded Linux on was counted as a Windows server when sales figures are tallied by organizations like IDC. I find the claim that Windows now owns 70% of the server market hard to believe. Granted, I don’t work in Windows-only shops but I wonder how many of those really exist. Most large organizations have a mix of Windows, UNIX, and Linux. I also suspect the 20% figure IDC assigns to Linux is accurate in terms of server sales but grossly underreports what percentage of servers actually run Linux. This leads me to a question I just don’t have a good answer for. How can Linux’ real market share and penetration be accurately measured? |
|
|
Comments (18)
Then the slant of articles like these would not be "Windows wins because customers are spending way more money on it!" but "Linux wins because customers are saving truckloads of cash!"
But.... who would pay me?
I recently worked for a firm that provided 1-5 tier tech support for various ISP's across the country. Unfortunately, every work station in every techie's cube was running XP...it would be a simple step-over assumption to think that the server room would smell suspiciously like Windozed IIS.
Nope...this corporation ran pure Debian servers from top to bottom. I had a chance to speak with the lead hardware and software engineers on several occasions and since we speak the same language the topic always ended up with Tux standing there as a mutual friend. There is not one byte of Microsoft server code in that entire operation. The server software was installed by local talent on machines, for the most part, that had once ran Windows server craep. So the question is only extended...not answered. How many thousands of small and medium businesses go about their business with Linux servers quietly doing their job? We won't ever know for sure...but one thing is certain...
For Microsoft to claim that they have captured the server market again is a sure indication that they have solid but confidential figures to the contrary.
h
Nope...this corporation ran pure Debian servers from top to bottom. I had a chance to speak with the lead hardware and software engineers on several occasions and since we speak the same language the topic always ended up with Tux standing there as a mutual friend. There is not one byte of Microsoft server code in that entire operation. The server software was installed by local talent on machines, for the most part, that had once ran Windows server craep. So the question is only extended...not answered. How many thousands of small and medium businesses go about their business with Linux servers quietly doing their job? We won't ever know for sure...but one thing is certain...
For Microsoft to claim that they have captured the server market again is a sure indication that they have solid but confidential figures to the contrary.
h
Ken -
One thing I wonder about. First, market "share" is an interesting number.
If you grow, but the market grows faster, your "share" will decrease.
I have noticed an real uptake in desktop Mac use. I see people showing off their Powerbooks and you'd be amazed at how often my nosy little ears hear things like "I gave up on Windows", etc.
My partner and I have a very fine HP Ubuntu server that came our way because it started out life as a Windows 2003 server for a small manufacturing company who got so frustrated that he pulled the plug, replaced it with a Mac server, and sent the old box to us just to get it out of his sight.
I find it impossible to believe that Windows is winning anything except numbers. They ain't winning hearts, that's for sure. That should concern them because numbers change.
One thing I wonder about. First, market "share" is an interesting number.
If you grow, but the market grows faster, your "share" will decrease.
I have noticed an real uptake in desktop Mac use. I see people showing off their Powerbooks and you'd be amazed at how often my nosy little ears hear things like "I gave up on Windows", etc.
My partner and I have a very fine HP Ubuntu server that came our way because it started out life as a Windows 2003 server for a small manufacturing company who got so frustrated that he pulled the plug, replaced it with a Mac server, and sent the old box to us just to get it out of his sight.
I find it impossible to believe that Windows is winning anything except numbers. They ain't winning hearts, that's for sure. That should concern them because numbers change.
OK, first point. As a general rule, businesses buy servers, not individuals. Individuals buy desktops, which may be used as servers, but won't show up in the figures.
Second point. Business use of Linux has come of age. Businesses have realized that they can support Linux in house, and don't really need those expensive support contracts except for mission critical applications. If they do need support, they're going to deal directly with the distributor or their existing support structure; not buy it from a vendor. They've also realized that they don't necessarily need to buy that expensive server they needed with Windows. They can use an older existing server or a higher end desktop machine and it will run Linux with no problems.
Result: Decreased sales of Linux servers from vendors. Not decreased use; merely decreased sales.
I haven't seen this first hand, as the company I work for is relatively Linux hostile, but it's my guess as to what's happening.
Second point. Business use of Linux has come of age. Businesses have realized that they can support Linux in house, and don't really need those expensive support contracts except for mission critical applications. If they do need support, they're going to deal directly with the distributor or their existing support structure; not buy it from a vendor. They've also realized that they don't necessarily need to buy that expensive server they needed with Windows. They can use an older existing server or a higher end desktop machine and it will run Linux with no problems.
Result: Decreased sales of Linux servers from vendors. Not decreased use; merely decreased sales.
I haven't seen this first hand, as the company I work for is relatively Linux hostile, but it's my guess as to what's happening.
jdixon -
I used to work (sigh) at a place that was moving to Linux. They did it both ways -- buying Linux preinstalled on servers, and moving Linux onto existing servers.
From the economics standpoint, paying for support isn't as big a deal as you think because corporations always pay for support. They can pay Red Hat or they can pay their own people, but they will pay somebody.
I used to work (sigh) at a place that was moving to Linux. They did it both ways -- buying Linux preinstalled on servers, and moving Linux onto existing servers.
From the economics standpoint, paying for support isn't as big a deal as you think because corporations always pay for support. They can pay Red Hat or they can pay their own people, but they will pay somebody.
> They can pay Red Hat or they can pay their own people, but they will pay somebody.
I agree. However, most larger companies already have an IT staff which can handle most support issues, so they're already paying their own people, whether they realize it or not.
I agree. However, most larger companies already have an IT staff which can handle most support issues, so they're already paying their own people, whether they realize it or not.
>so they're already paying their own people, whether they realize it or not.
They absolutely realize it, down to the penny, just as they do when they choose to buy their servers from a vendor rather than assemble them.
They absolutely realize it, down to the penny, just as they do when they choose to buy their servers from a vendor rather than assemble them.
Wouldn't a better way of measuring server share be some kind of application that connects to all the Web servers out there and finds out that way whether they are running Windows, Linux, BSD or what have you? Whenever I go to a bad URL on a given domain, I get that Apache screen that contains all this information. So it's right there to be found.
>Wouldn't a better way of measuring server share
No. That would actually be worse because web servers are only one kind of server. Worse, web servers are dominated by Apache and most of those apaches are running on a flavor of *x. In short, that would deliver a different bias. It's a bias we might like, but certainly no better as a measure of the truth.
No. That would actually be worse because web servers are only one kind of server. Worse, web servers are dominated by Apache and most of those apaches are running on a flavor of *x. In short, that would deliver a different bias. It's a bias we might like, but certainly no better as a measure of the truth.
> ...some kind of application that connects to all the Web servers...
That only works for servers running an http service and facing the web. That's only a small subset of servers.
That only works for servers running an http service and facing the web. That's only a small subset of servers.
> That's only a small subset of servers.
One place I worked had their one outside web server, but a dozen redundant RedHat systems ganged together doing financial transaction processing.
It also didn't count the 5 or so Linux desktop machines in the office.
So yeah, visible servers are not a reliable count. Heck, the 5 Linux systems here at home wouldn't be counted at all.
One place I worked had their one outside web server, but a dozen redundant RedHat systems ganged together doing financial transaction processing.
It also didn't count the 5 or so Linux desktop machines in the office.
So yeah, visible servers are not a reliable count. Heck, the 5 Linux systems here at home wouldn't be counted at all.
For the sake of accuracy, how about an open source "Linux call home" program!
It would be developed, supported, maintained, and signed by The Linux Foundation.
Users launch it at their discretion. Users select & decide what hardware can be used by program to generates a unique checksum from various hardware information. Users see data being collected. User decides to send or not to send information.
Problem solved.
Users launch it at their discretion. Users select & decide what hardware can be used by program to generates a unique checksum from various hardware information. Users see data being collected. User decides to send or not to send information.
Problem solved.
> ...how about an open source "Linux call home" program!
A couple of distros (I believe Debian is one of them) have such a program.
Slackware includes a link to the Linux Counter in it's welcoming email.
A couple of distros (I believe Debian is one of them) have such a program.
Slackware includes a link to the Linux Counter in it's welcoming email.
Quoting:A couple of distros (I believe Debian is one of them) have such a programDo you know if they released any of their stats?
Abe:
Some checking revealed that what I was remembering from Debian is actually a package popularity tracker, not a system tracker, but if you track the popularity of a required package, it should match the number of installed systems fairly closely. However, only people who install the package will be counted.
The stats are at: http://popcon.debian.org/
Linspire has a user tracking service, but it doesn't look like they make stats available. You can look at a map presentation of the data at: http://www.lraiser.com/index.html
Added: As a publicly traded company, Red Hat's revenues are available. If you take their subscription revenue and divide it by the per system price, you'll come out with a low end figure for the number of active Red Hat systems being supported.
Some checking revealed that what I was remembering from Debian is actually a package popularity tracker, not a system tracker, but if you track the popularity of a required package, it should match the number of installed systems fairly closely. However, only people who install the package will be counted.
The stats are at: http://popcon.debian.org/
Linspire has a user tracking service, but it doesn't look like they make stats available. You can look at a map presentation of the data at: http://www.lraiser.com/index.html
Added: As a publicly traded company, Red Hat's revenues are available. If you take their subscription revenue and divide it by the per system price, you'll come out with a low end figure for the number of active Red Hat systems being supported.
Thanks JD. Very nice of you to furnish the links.
Now, how about a single source! :) Just kidding.
Now, how about a single source! :) Just kidding.
You can see the web server stats here. http://news.netcraft.com/archives/web_server_survey.html
Post a comment
A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.
Your message:Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help
Sign up and post (it is free) or Log in and post