Apache Jumps In Market Share

mshiltonj writes "In case no one has noticed, the lastest Netcraft web server survey showed a marked shift in market share in just one month. Apache gained 2.63% and IIS fell 2.06%. However, the previous month showed an even larger change in Microsoft's favor, so Apache is (quickly) making up for lost ground, as discussed before. Was this turnaround due to the release of Apache 2.0? Sadly, in the last 12 months, Apache's market share has noticeably eroded, while IIS has gradually gained ground."

Comments

Sun also gained market share at the expense of IBM

[Anonymous Coward]

Sun went from 51% to 54% of the Unix server market, largely at the expense of Big Blue: IBM fell from 21% to 17% (and HP passed them, to take second place):

Sun Makes Gains In Unix Market As IBM Slips
http://www.informationweek.com/shared/printableArt icle?doc_id=IWK20020514S0004

In other news...

[thefatz]

Slashdot Junps ahead getting Apache article posted.

Film at 11

That's Nimda for you...

[canthusus]

...which is why I use TinyWeb Does very little - but does it securely.

Re:That's Nimda for you...

[driehuis]

Or thttpd...

Even though, arguably, even thttpd has too many features :-)

Re:Apache is not dying - apologies

[tps12]

It's official - Netcraft confirms: Apache is not dying.

Yet another crippling bombshell hit the beleaguered non-Apache community when recently IDC confirmed that Apache accounts for more than a fraction of 1 percent of all servers. Coming on the heels of the latest Netcraft survey which plainly states that Apache has lost less market share, this news serves to reinforce what we've known all along. Apache isn't collapsing in complete disarray, as further exemplified by not failing dead last in the recent Sys Admin comprehensive networking test.

You don't need to be a Kreskin to predict Apache's future. The hand writing is on the wall: Apache does not face a bleak future. In fact there will be a future for Apache because Apache isn't dying . Things are not looking very bad for Apache. As many of us are already aware, Apache continues to not lose market share. Red ink doesn't flow like a river of blood. RedHat Apache isn't the most endangered of them all, having not lost 93% of its core developers.

Let's keep to the facts and look at the numbers.

OpenBSD leader Theo states that there are 7000 users of Apache on OpenBSD. How many users of Linux Apache are there? Let's see. The number of OpenBSD versus Linux posts on Usenet is roughly in ratio of 5 to 1. Therefore there are about 7000/5 = 1400 Linux Apache users. NetBSD posts on Usenet are about half of the volume of Linux posts. Therefore there are about 700 users of NetBSD Apache. A recent article put MS-IIS at about 80 percent of the web server market. Therefore there are (7000+1400+700)*4 = 36400 IIS users. This is consistent with the number of IIS Usenet posts.

Due to the troubles of OpenBSD, abysmal sales and so on, Apache did not go out of business and wasn't taken over by Linux who sell another troubled OS. Now NetBSD is also not dead, its corpse not turned over to yet another charnel house.

All major surveys show that Apache has not steadily declined in market share. Apache isn't very sick and its long term survival prospects aren't very dim. If Apache is to survive at all it won't be among OS hobbyist dabblers. Apache continues to not decay. Anything short of a miracle could save it at this point in time. For all practical purposes, Apache is not dying.

Fact: Apache is alive

/. mentioned

[mnordstr]

Apparently Slashdot's use of Apache 2 is important according to Netcraft as it is mentioned: several high profile sites, including News.com and slashdot.org have started using it

Re:/. mentioned

[J'raxis]

What?

HTTP/1.1 200 OK
Date: Wed, 15 May 2002 21:48:54 GMT
Server: Apache/1.3.20 (Unix) mod_perl/1.25 mod_gzip/1.3.19.1a
SLASH_LOG_DATA: shtml
X-Powered-By: Slash 2.003000
X-Fry: And then when I feel so stuffed I can't eat any more, I just use the restroom, and then I *can* eat more!
Cache-Control: private
Pragma: private
Connection: close
Content-Type: text/html

Turnaround not due to Apache 2.0

[Cycon]

Apache gained 2.63% and IIS fell 2.06%. However, the previous month showed an even larger change in Microsoft's favor, so Apache is (quickly) making up for lost ground, as discussed before. Was this turnaround due to the release of Apache 2.0?

According to the Netcraft Web Server Survey page, the drop in IIS over the past month was due to a change in Homestead.com's policies:

Microsoft drop back a couple of percentage points this month, primarily attributable to the reaping of around a million expired sites at homestead.com. Homestead used to offer limited free hosting supported by advertising, but has revised its business model. As of April 15th has announced that it is migrating to a traditional paid-for services model.

--Cycon

Apache 2.0 lack of modules

[gruntvald]

I would think the release of Apache 2.0, which lacks php or mod_perl, the two most prominent web scripting environments, renders it to be the least useful or relevant apache release ever. Why would an IIS site ever consider switching - security has always been low on the totem pole for IIS admins, they are looking for domain integration, and other windows interfaces. Which, of course, Apache lacks. Other than the fact it now runs better on OS/2 than before, and has a BEOS port, there's little incentive to "upgrade" or switch. And those examples are only partially facetious.

Gartner recommendation

[booch]

It's probably due to Gartner's recommendation to drop IIS due to all the security holes IIS has had recently. Or more to the point, the holes themselves have caused people to lose faith in IIS and move to Apache.

Re:Gartner recommendation

[realdpk]

I've seen no sign whatsoever that people running IIS * are now suddenly concerned about security. :)

* I mean, people who weren't concerned before, of course.

Maybe Google and Netcraft should team up

[driehuis]

Hmmm... If a single web hosting company can influence the stats by that much, there is something seriously wrong with the stats.

One approach would be to count unique IP addresses (i.e., vhosted sites would not be counted twice).

But even better, it would be way cool if Google's linking metrics could be brought in. That way, a rough guesstimate of the amount of information served by all the web servers could be established.

There's lies, damn lies and statistics. I remember when a sales droid walked up to me and recommended I switch to IIS because it was the dominant web server. He had brought this list of high profile IIS installations, and on the surface it looked impressive. When I confronted him with how many of those still had Apache or UNIX somewhere in the path (either as a firewall, server for static images or ads), he started mumbling incoherently.

Re:Maybe Google and Netcraft should team up

[MadCow42]

>> One approach would be to count unique IP addresses (i.e., vhosted sites would not be counted twice).

Now, wouldn't that also pose problems of its own? I don't have numbers to prove it, but I firmly believe that you could host a significantly larger number of sites on one good Linux/Apache server than you could on an IIS server with the same hardware, due to efficiencies of both Apache and the OS.

However, you'd have the same limiting factor of bandwidth.

Just something to think about,
MadCow.

Re:Maybe Google and Netcraft should team up

[driehuis]

You're probably right with the assertion of the number of vhosts per IP limiting IIS more than Apache (on a decent OS like BSD or Linux). However, I have a hard time believing that this would outweigh the numeric advantage the mass-vhosting web server would have. As you say, bandwdith limitations will kick in way before operating system overhead kicks in (and I for one do not think weighing a gazillion of never-viewed web sites on a single IP address should count as heavy as one {Amazon|Slashdot|CNN} sized web server).

Question about IIS

[MillionthMonkey]

Do these numbers include all the people running IIS who don't realize they're running IIS, or that a web server is running on their machine at all?

Just wondering.

Re:Question about IIS

[Billly Gates]

netcraft only measures DNS registered sites so the answer is no. Whats causing the rise in IIS use is microsoft paying off huge webhosting companies like homestead.com and many others to switch. Ask anyone in the webhosting bussiness how many calls from MS salesmen do they get in a week? Ms constantly bugs them and yes even offers to pay for the upgrade costs for all servers! The reason? Marketshare. Webhosting companies have like %40 of DNS registered sites out there. MS knows the suits in the fortune 1,000 only buy whats popular and not whats best. If they see a jump in IIS they will buy it in fear of apache becoming the next netscape or lotus smartsuite.

Its total bs and crazy to even think like this but this is how IT thinks. If you look at netcrafts monthly usage you will see a jump in IIS after some big web hosting company switches to it. Then MS will make one fud report after another about bussiness migration away from expensive unix systems to NT and how its changing the whole world and yada yada yada. However now I believe alot of bussinesses are thinking about switching back to apache due to the huge security holes in IIS. If they use active server pages then I feel real sorry for them. Infact my windows2000 at home was hacked using just a small 56k modem. I believe the hacker used IIS in which I had installed to learn vb active server pages. Absolutely astounding! I checked the logs myself and no this was not a registered website but for my own personal use.

Re:Question about IIS

[MillionthMonkey]

netcraft only measures DNS registered sites so the answer is no.

What, you mean they don't just count unique IP origins of Code Red packets? :)

Does anyone trust these numbers anymore?

[Havokmon]

It's pretty crappy that NetCraft includes all those domain names that are just parked on webservers. One provider moves web servers one way or another, and it appears that there's a huge migration somewhere.

Migrating empty virtual hosts isn't what 95% of the internet will be doing when evaluating a 'new' web server.

Re:Apache

[Anonymous Coward]

Actually, 2.0 is the first version they've said should actually be run on Windows in a production environment, so, if anything, it'd be a better choice for your Windows Me web serving needs.

Apple?

[ruiner13]

Could more people be figuring out how to simply flick the switch in OS X to make their computer an Apache server? As Apple's Xserve catches on, as well as more people migrate to OS X, Apache stands to gain ground too. I think that is a good relationship.