Netcraft Says IIS Gaining on Apache

benjymouse quotes this month's netcraft survey "In the August 2007 survey we received responses from 127,961,479 sites, an increase of 2.3 million sites from last month. Microsoft continues to increase its web server market share, adding 2.6 million sites this month as Apache loses 991K hostnames. As a result, Windows improves its market share by 1.4% to 34.2%, while Apache slips by 1.7% to 48.4%. Microsoft's recent gains raise the prospect that Windows may soon challenge Apache's leadership position."

Comments

Re:Google Web Server

[Leffe]

With this month's survey, Netcraft has begun tracking Google's custom web server software known as GFE (Google Front End), which is currently found on 2.7 million hostnames, or 2.3% of all sites. Google customizes its web infrastructure, with in-house solutions for software and hardware, including energy-efficient servers and power supplies. GFE is the server found on Blogger sites at blogspot.com, while Google uses GWS (Google Web Server) on some other services, although none with the volume of hostnames seen at Blogger.

http://news.netcraft.com/archives/2007/05/01/may_2 007_web_server_survey.html

Re:From the person above

[mwilliamson]

Yes, I'd say the proper choice would be to match the talent you have on your development team with the tools required to build the desired product. It wouldn't make sense to force a bunch of LAMP developers to switch to IIS/Windows/MSSQL just to become a "pure windows shop". It also would make no sense to force a team of IIS/ASP/DNET developers onto LAMP just to change OSes. At my place of work we have both, and neither is going to take over the other.

Uptime

[ciryon]

Apache has a vast majority of sites with longest uptime.

Re:GoDaddy and the like?

[morgan_greywolf]

In fairness to the GP post, his statement is accurate. People were paid to develop IIS and therefore can be called professionals and MS is a well known name. If that makes him sleep better at night and doesn't mind throwing his money around for that solution and its artificial restrictions, so be it.

People were paid to develop Apache. Open source != everybody worked for free.

IIS dying out in Germany

[SgtChaireBourne]

There is also a serious discrepancy in that other stats seem to show IIS on the last moments of extinction in hi-tech zones like Germany. NetCraft report doesn't really have any explanation of the figures it presents.

What's really problematic is that over time NetCraft has become less informative. No mention has been made lately of what the changes in market share are attributed to. In years past, even a percent or two got a few lines of explanation or analysis. Did one of the service packs or 'security' upgrades install and turn on IIS for all Windows users? Or are more domain parkers and cybersquatters using IIS in the server identification string?

This downturn started last year when MS paid GoDaddy to swap out (or claim to swap out) its domain parking. GoDaddy did get the OSS community to lay off by throwing some chump change to OpenSSH and we can see the result of these last 12+ months. The money did some good, but if it's just a one-off donation, then it's questionable whether then benefit offsets the harm. Either way it's funny to see GoDaddy decision makers thinking they can buy indulgences. Maybe it ought to become an annual fee.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:What?!

[krgallagher]

"So if the Apache team makes a GUI option for configurations then it would probably start gaining the lost market share back? "

Here are several GUI administration tools for Apache.

Re:GoDaddy and the like?

[hab136]

And, by what magic of the new HTTP 2.0 protocol are you running two different server software types on the same IP and on the same port?

Cisco, Foundry, and several other vendors make load balancing devices that allow you to have one public-facing IP distributed to dozens of back-end machines. If you connect to 1.2.3.4 on port 80, you can actually be connected to machine A, B, C, etc.. these machines not necessarily running the same web server software or operating system. In this case the actual public-facing IP is often called a VIP (Virtual IP) since it's not assigned anywhere except on the load balancer.

Counting by IP address is by far, more accurate than comparing hostnames and sites. So counting by IP is MORE VALID than the method they used.

You're right, it depends what you're counting. If you're counting the number of boxes that run a particular web server, then IPs will probably be more accurate, although load balancing will skew this. If you're counting the number of customers those chose IIS vs. Apache, whether or not they are jammed onto a large hosting server with other customers? Counting by sites will be more accurate there, although skewed by domain parking.

Despite your ignorant little point.

I love you too man!

If you want to count individual BOXES, then IP is as close as you are going to get without doing a survey or special fingerprinting of the data to find differences in machines. (It will still be too big, I run 122 IPs with about 350 sites on them. The ratio changes all the time due to customers coming/going and reconfigurations...)

I'm guessing 121 of those sites are SSL? Name-based multihosting and load-balancers mean you normally only ever need one public-facing IP for non-SSL sites. Better yet, all your back-end boxes can be configured identically, with all of the sites on every box, so you can spread the load evenly. Even if your application needs to keep the user on a particular box during a session, the load balancer can be directed to do so.

I am going to guess, that the fact that millions of "parking" domains are run for the most part on Apache causes the popularity of that particular activity of lowlife scum to weigh heavily in the Netcraft numbers.

I agree (see above about skewing).