Ubuntu, the only web server OS showing growth?

According to W3Techs, Ubuntu is the only web server OS showing a continuous growth rate for (at least) the last year. After passing Suse and Fedora last year, we passed in front of RHEL usage in July.  CentOS and Debian are still ahead though.

Usage of Linux for web sites, W3Techs

 More details at: http://w3techs.com/technologies/history_details/os-linux

Share this

Comments

Options d'affichage des commentaires

Sélectionnez la méthode d'affichage des commentaires que vous préférez, puis cliquez sur "Sauvegarder les paramètres" pour activer vos changements.

Not just growth, but about

Not just growth, but about 60% increase in just the past year!

Graphs!

each axis needs a unit and a dimension along with the numbers!

Sorry, the link to get more

Sorry, the link to get more details was not prominent, it should now be more accessible.

Debian stable

What's interesting to see is that Debian keeps it's share very stable (look at yearly statistics). All RPM based distributions are losing share. RPM still has majority, but that might change by the end of the year…

Well...don't trust statistics

Hey Nicolas,

the stats are giving me a smile on my face, but you and I know never trust a statistic which isn't forged by yourself ;)

Therefore, the stats, when I understand it correctly, are coming from the results of head requests against apache webservers running on different distros.

So, the total number of hosts running apache on different distros could be higher, because many Admins are hiding the OS from the return result. Furthermore, there are other services running on other hosts like Mailservers etc, which are not counted (sadly).

What' more interesting would be the absolute numbers of hosts of users/companies using Ubuntu, RedHat, SLES etc. I know this is not that easy, but RH does sell support for RHEL releases so they should have a good number of hosts per customers who are running RHEL. SuSE does the same.

The number of this statistic doesn't surprise me, really, because having Amazon S3 Cloud and the easy to deploy solution of Ubuntu on Amazons Cloud gives people more chances to get started easily with a host running apache, without a clue on how to secure their hosts.

Regards,

\sh

I won't go as far as to say

I won't go as far as to say don't trust these graphs. But I will say this…

Since Nicolas was kind enough to point to the data source as a reference, instead of asking him to tell me how the OS is determined I asked w3techs concerning details about how they fingerprint hosts in an email today. I was told in the reply I got back that they consider their methodology a proprietary algorithm and would not go into details about it with me. They implied it was based in part on header information, but they flatly declined to go into specifics concerning the heuristics they use. The person who responded to me was very polite and very prompt with the information.

I have a very large problem with the attitude that it is acceptable for a data analytics company to publish data publicly but refuse to make the methodology public. Paying customers who purchase data products from analytics companies get access to methodology as part of the deliverable under contract/licensing terms which restrict republication as part of the business relationship. If you are going to publish data publicly you should be prepared to explain the methodology to the public on an equal access basis to the entire readership. Methodology matters.

And while I understand why Nicolas reached for these numbers, and I do not find the trends unreasonable on their face, I'm disheartened that he feels its acceptable to republish data from sources which flat refuse to publish their methodology for public scrutiny. It frankly unacceptable to continue to republish data like this without first vetting the methodology yourself so you can explain it in detail (or being able to point to a concise and accurate methodology statement if the methodology is itself beyond your understanding)

The question I would _love_ to be able to ask, which they don't provide a trend for is the following. How does Unknown Unix trend against Linux in time? Is Unknown Unix growing? How much of Unknown Unix is really Linux they can't determine via their algorithm and thus shows up in the wrong place? Is the rate of growth of Unknown Unix outpacing the grown of Ubuntu in their trending?

  • jef

Jef, since I checked first

Jef,

since I checked first the source of data, and, like you, didn't get any clue on how they gathered the data, I think people need to put the data into the right context.

As Nicolas, I'm really happy about the growth of Ubuntu Servers, but people who are reading this on planet.ubuntu.com or on the ubuntu cloud portal, need to put the message + the data into the right context:

  1. Nic works for Canonical as Cloud Solutions Lead, so there is a commercial interest into pushing the message „Ubuntu is the better Server Platform, you can see the growth here…“
  2. The numbers _are_ questionable, because, as I wrote, they don't take all other hosts which are not able to be fingerprinted into account.

2a) http://w3techs.com/…unix/all/all ← 50.2% Linux and 47.6% Unkown , also have BSD, Solaris, Darwin and HP-UX. What hosts are the 47.6% Unkown? Could I read this number as „there could be 97.6% hosts using Linux, but we only can determine 50.2% of them as real Linux Servers“? 2b) Their explanation of „Technologies“ on http://w3techs.com/technologies is more then accurate 2c) Their disclaimer on http://w3techs.com/disclaimer is just a bla bla of what you found out, that they don't want to provide their method of data gathering

  1. When you follow several links, you will find this: http://w3techs.com/…untu/all/all

3a) Means, there is eventually a growth of Ubuntu, but only 5.3% of all hosts are running Ubuntu as OS. 3b) When you read this number, a growth of 6.5% between october 2010(10.3%) and october 2011 (16.5%) is really a good growth, but the absolute numbers of hosts are not.

And this is only my interpretation of the known numbers. Eventually you can find another interpretation of those. And all the other readers will do the same. Therefore, don't trust statistics, you didn't forge yourself.

\sh

Yes, I really really wish at

Yes, I really really wish at the very least they had a trending graph for the unknown Unix segment. If Unknown growth is also decreasing, then the conclusion Nic wants to draw is validated to some extent. If Unknown is grown as large or larger than Ubuntu, then you can't make any conclusion as to regard to relative growth as the Unknown category is hiding a lot of information.

  • jef

I think Ubuntu is not the

I think Ubuntu is not the only web server OS who is having a consistent growth rate because if you study the graph for the past six months, you can see that Mac OS aren’t lagging that far behind, although the same can’t be said about Microsoft. It is true that Ubuntu leads Mac by a million miles when it comes to the rate percentage but then again the high demands for Ubuntu rising every day has played a huge role in its 60% increase.There is growth everywhere private cloud hosting is becoming more and more popular than normal hosting..

Options d'affichage des commentaires

Sélectionnez la méthode d'affichage des commentaires que vous préférez, puis cliquez sur "Sauvegarder les paramètres" pour activer vos changements.