SlashDot: .../mode=nested&threshold=3

IIS Vs. Apache and Netcraft Results

Apache | Posted by Hemos on Monday July 10, @11:37AM
from the what-to-make-of-this? dept.
ant banks wrote to us with an article that takes issue with the Netcraft results regarding Apache's continued domination of web servers. This server looked at Fortune 500 companies and their deployment, with some differing results.

Chris McKinstry Replies: Telecopes, AI And More | Hacking Insurance For Net Businesses  >


The following comments are owned by whoever posted them. Slashdot is not responsible for what they say.

but what about the most popular web sites?

(Score:3, Interesting)
by sethgecko on Monday July 10, @11:43AM EDT
(User Info)

Sure, the article says that fortune 500 companies use NT/IIS. But that's to be expected--these are the most conservative, least net-savvy cases. How about a survey of the hot100 web sites? Yahoo certainly uses Apache + FreeBSD. Microsoft/MSN uses IIS, except for Hotmail which uses FreeBSD + Apache.

It's just not valid saying that measuring fortune 500 companies is a useful metric--these are not the companies with successful or busy web sites (for the most part--obviously Dell gets a lot of traffic. So does IBM. But what about GM? I'm sure lots of people go to GM's web site to buy a car. Not.).

Be ot or bot ne ot, taht is the nestquoi. [ Reply to This | Parent ]

Fail to see the point

(Score:3, Interesting)
by cvd6262 on Monday July 10, @11:47AM EDT
(User Info)

While I understand the idea that those who can afford Win solutions would flock to them, I feel that this article itself misses the point.

The idea that John Doe can be running a server himself is a triumph for free software and a free speech-centered internet. Netcraft doesn't care if the sites they survey are making money - but it doesn't surprise me that a pro-M$ site would care.

The article also fails to mention that very few big name server manufacturers offered linux solutions until recently. One can say that would've changed in time, but the anti-trust case did nothing but help.

In closing, is running Apache/1.3.6 (Unix) mod_ssl/2.2.8 SSLeay/0.9.0b on FreeBSD/b.
[ Reply to This | Parent ]

I love comments like this:

(Score:3, Insightful)
by AugstWest (fnord@themanagement.egg) on Monday July 10, @11:55AM EDT
(User Info)

But ENT's survey, arguably, is more meaningful, as it omits home-based systems, enthusiasts, experimenters, and smaller commercial sites that likely won't be around for any appreciable amount of time anyway.

Welcome to America, where the Fortune 500 is all that matters.

But what are these sites? Do they make any money, or are they just personal sites? And will they be still be around a year from now?

This is the important criteria? Oh, wait, that's right, making money is all that matters. It sounds more like spending money is all that matters.

Companies such as Ford Motor Company, Philip Morris, Compaq, and Albertsons are all using IIS.

What kinda PR piece is this? Listing a few cutomers? How about the Royal Family, and the Financial Times? How about the w3 Consortium themselves? Oh, wait, they don't make money, so they're not important.

--[The history of the music industry is a history of exploitation and theft - Robert Fripp]-- [ Reply to This | Parent ]

Yes, well done

(Score:4, Flamebait)
by Vanders ( on Monday July 10, @11:55AM EDT
(User Info)

First, a quote from the site:

with only 10 companies in the Fortune 500 using the upstart open source OS to deploy their production sites.

O.K, stop right there. Who but a Windows site would still call Linux an "upstart" OS these days? Sigh.

And onto the main point. What the hell are they on about? If i survey the top 500 Open Source Software sites, i wonder which server would come out on top of the results?

Limiting your survey to 500 Sites is just stupid. And like many others have pointed out, so what? Does it matter what the site is actually for, from "Look at my pet dog", all the way upto Slashdot size? Whats to stop these bedroom servers running IIS anyway?

Lastly, the most obvious thing to point out to our friends at the rabid Windows magazine, is that Netcraft are neutral, and have no interest in swaying the figures. Want to bet Win Informant is unbiased and had no interest in proving something with the survey?

This .sig here until i think of something funny. [ Reply to This | Parent ]

A general criticism of statistics (or their use)

(Score:3, Insightful)
by dsplat on Monday July 10, @12:00PM EDT
(User Info)

All statistics are simplifying the data from a collection of individual data points, each with its own individual character, to an aggregate measurement. Let's take the first paragraph of the article and reword it slightly and see what we get, shall we?

According to a survey by Fear, Uncertainty and Doubt, Inc., the Free Software Foundation's editor, Emacs, continues to be dominated by Microsoft's Notepad. Or does it? Notepad's perceived dominance might be less sure than you think, since FU&D doesn't provide any sort of meaningful data about the editors it surveys. For example, Notepad sailed past the 500 million mark, which sounds impressive. But who are these users? Do they make any money, or are they just personal computers running games, word processors, spreadsheets and web browsers? And will this release still be installed on them a year from now? According to, a neutral third party survey and benchmarking web site, the FU&D data is almost meaningless as-is. So Completely Biased took a look at the largest U.S. corporations to see what editors their developers actually used. The results, of course, were surprising.

Netcraft's survey is a valid count. It is not the only way of looking at the data. There are a huge number of ways of slicing it:

There are lies, statistics, and benchmarks. And then there are the conclusions based on them. The conclusions are often the shakiest part.

Slight disorientation after prolonged system uptime is normal for new Linux users. Please do not adjust your browser. [ Reply to This | Parent ]
  • 83 replies beneath your current threshold.