Slashdot Mirror
Latest Netcraft survey shows Apache increase
The latest Netscraft Survey is out. Apache enjoyed an over 1 percent increase, with Microsoft and Netscape showing some decreases. According to the survey, Apache has a 54.81 percent "market share." Also reported is the fact that Webjump actuals uses a hybrid setup with NT serving static content and the dynamic content with a Solaris/Apache/Perl system. Tucked away in the report is a small factoid that PHP is on over 1.1 million domains.
Percentages are good, but what I liked was taking the change in the number of hosts and looking at the sheer numbers.
Do the math. Approximately 4 Apache hosts went up for each IIS/PWS host that went up in the past month.
Yes, lots of those hosts are virtual hosts on the same machines, but even that says something about Apache's penetration at web hosting providers.
From the netcraft survey
.com domains with the most distinct certificates found by the September SSL survey, run 213 Netscape servers [out of a total of 341 sites], against 42 Microsoft [of which 28 are in the microsoft.com domain], and 24 Apache. Equally notable is their choice of operating systems, where both NT and Linux, strongly represented in the SSL Survey as a whole have a relatively small share. Just 57 sites run NT, with the most common Unix systems according to signatures detected in the tcp/ip characteristics, being Solaris and AIX. Only one site was detected as running Linux.
The top 10
This shows that on the really big servers netscape servers still rule. When you scrap the ms domains from the survey apache has a larger marketshare (scrapping ms is a good idea since price/performance/support probably did not play a role in choosing a webserver there).
The survey also tells us that both IIS and Apache saw a rise in marketshare for smaller SSL based stuff.
The last line is sort of interesting too since it shows that linux does not play a big role as a webserver platform for very large sites.
Jilles
- The ability to map files from other servers into its url space.
- I want to be able to control caching directives passed to the client. So like I could make files within
/forum/ passed with no-cache to the client. - I would like the proxy to have basic methods of determining if a fresh copy of something needs to be fetched. So "If
/forum/ not-cached by no-account in last 5 minutes, get new /forum/; But if client-is-user and not sent in last 3 minutes get a new forum for user" - I want to be able to manipulate cookies so that some files and cgis mapped within the server space can't get ones not intended for them (user server).
- I want to be able to protect mapped URLs by password (at the load-balancer) and by ip.
- How about running a https server and mapping the requests to unencrypted http servers
- Here's what I want to use it to handle
- A user box with some SUID user cgis and mostly static html (apache)
- A Slashdot-like forum (apache w/ php/mysql)
- A file archive server (thttpd)
- A group of small webservers for displaying stats on from developer machines (thttpd)
- A discussion forum on a mod_perl
- A web-based email server (custom running on OpenBSD)
Anyone have a solution for this?maybe the government should break up apache into....err nevermind.
Have you seen Ironstayn vs Supergovernment yet?
Others have commented on the margin of error part. As to whether it's a real increase, I'd say what matters is the trends over time. For instance, notice
One of the resident astroturfers in comp.os.linux.advocacy bragged last month about how Microsoft was killing Apache with such a large gain, while Apache showed a small loss. It was nonsense, of course, but it would be equally nonsensical to say that Apache is now on a roll due to this month's report. Clearly, Apache is "winning", but that observation is based more on its current share and the absence of any moving trends against it in the chart.
--
It's October 6th. Where's W2K? Over the horizon again, eh?
Sheesh, evil *and* a jerk. -- Jade
Since the vast majority of Linux servers ran Apache at the time you are talking something like 28% Apache + Linux. The majority of Netscape Servers were actually deployed on NT so NT + IIS probably amounted to 15%.
It's only not a lie if they count different kernel, Distribution and Apache versions separately. What about NT Service packs ?
They only count around 1/4 the net which would give them less than 1% margin of error.
--= Isn't it surprising how badly I spell ?
It's an old version, so what's the point? Do you want a code fork, where any improvements are obviated or made incompatible with the current functioning version? At what point do the /. developers decide 'enough is enough' and let people at the code? Is there a foreseeable actual date for this?
/.'s nature as an advocate of OSS, that it be held to the same standards that it (and its 'community') holds to other projects/corps/developers.. Glass house, stones, etc..
I think, given
And if noone shows interest in this, it'll never be seen... Hell, look at the interest that has been shown and the results so far.. Will the recent source (or a public CVS server) just magically appear?
Your Working Boy,
I'd have to agree with the original guy here. In some fields where precedent is being set, like PERL and X Window, open-source shows great innovation. But in area's where it's not, features seems to be looked at an added only when there's a glaring weakness. Like Linux and SMP, etc...
I don't think that any project will prosper without a form of competition. It doesn't have to be financial, but just another group of people doing something that produces the same result but by a different means. That way each group could look at each others work and pick out the best.
Maybe someone or ones should begin a new HTTP server project with a completely new source tree. Take nothing from Apache, but just build the "best" server they can.
Wrong. Go here. See in the type near the bottom. It says "The host you examine will be included in future surveys".
If you click the "add your site" link, it just brings you to the generic query page, which seems to me to mean that the way a site gets checked or scheduled to be checked is by actually querrying it. That, to me, seems like it'd be incredibly easy to tilt the results one way or the other.
Not that I'm trying to defend anyone. But to say NetCraft is unbiased, to me, seems false. The sites that get queried are the ones that users ask to have queried.
I'll email you with... oh. Never mind.
--
Xenu loves you!
I notice that the Microsoft entry includes Personal Web Server as well as IIS. Now, all Microsoft has to do is configure Windows and NT workstation so that PWS is on by default, just like Apple does with the Mac web sharing, and every Windows user with a Cable Modem or DLS line will count as a Microsoft server. Then watch them brag.
Sure, it would slow things down and open up some security holes, but that would be a small price to pay for the greater glory of Redmond.
Note for the humor impaired: just kidding, folks.
Information is not Knowledge
Others have mentioned possible problems in interpreting such data which include (but are not limited to) the following:
OK, having said that, it might be useful to pretend that none of these were concerns, and that we really did want to know whether a 1% increase in the number of domains served by Apache meant anything. Here's the short answer:
I can't tell you that.
This is especially true if the domains surveyed in some sense are the population. In that case, whether or not you care that Apache added 500,000 domain names to the population while IIS added 125,000 is basically up to you. There are many explanations for why this could have happened, and not all of them are very interesting. (Again, others have pointed out why.)
Personally, I would have been more interested in certain kinds of longitudinal breakdowns rather than the overall numbers. Some of those questions would include:
Call me a geek, but these are questions I think could be more interesting to ask. And, yes, some of the answers to these are given or hinted at on the netcraft website.
But there is one more question, which is the one the original poster asked:
But what if this really were a sampling question; is a 1% difference likely to be reliable?
If all N of the netcraft domains were independently and randomly sampled from the total population of domains, then a 95% confidence interval for a given market share, M, where M is between 0 and 1 is:
[M - 1.96*(M*(1-M)/N)^.5, M + 1.96*(M*(1-M)/N)^.5]
For Apache's market share in November, we would get the interval [.5479, .5483]. For the October share, the interval is something like [.5365, .5369]. Those are pretty tight intervals, but the sample size is over 8 million...
And this is the real point: when you have random samples this huge, error bars are pretty danged small. So it's too bad these really aren't random samples...
Babar
It's marvellous to see an Open Source solution win such a clear victory over proprietary rivals, but I hope someday to see Apache start to lose market dominance again, in favour of some of its open source rivals (like Zope). The way Apache does things isn't always the best way to manage Web content provision, and a monoculture of Web servers would certainly be a Bad Thing.
--
Xenu loves you!
It shows the usage by platform. There are a couple significant Apache derivatives that aren't grouped into the more conservative number that is used for the graph.
In fact, the bulk of the tailing off shown in the graph for Apache was actually slack picked up by Apache derivatives!
Not sure if this makes sense or if I even understand how Netcraft works ... Just curious how reliable these results are ...
--elint
"...So if you're cute, or even beautiful, remember: There's more of us ugly motherfuckers than you are." --Frank Zappa.
I just have to wonder, with an "increase of 1%" what the margin of error is - does this represent a real increase?
:)
Also, how many of the sites in question are the Apache "Congratulations!" page when Apache is installed and enabled by default on various Linux distributions?
Not to be a wet blanket - just wondering.
----
Not at all. In fact, I hope there will be a competitor of Apache which occupies roughly the same percentage of the market. Competition is always healthy. Whenever there is no competition, there is no reason not to sit back and relax. And that's when quality starts to drop. I'll probably get flamed for this, but MS products didn't suck that much when MS was still a small company. It's only when they became a dominant force that their products began to really deteriorate.
IMNSHO, shouting hooray to Apache because it's the dominant factor in the webserver market is no different from MS declaring how good the world would be if everybody switched to Windows. I'm not saying Apache sucks (I use it for a website project in fact), but that if there is no competition, eventually it will suck. (Note: this is not intended to be flamebait)
mikre he sophia he tou Mikrosophou.
Zope can run as a web server on its own, or handle requests through Apache. I think the latter is recommended though.
At least, this was so last time I looked. I know they've put quite a bit of work into making the Apache path faster since then.
But you're right, what I wrote was misleading.
--
Xenu loves you!
NT/IIS seems to be running a lot of commerce sites these days.
It happens that if you wanted to run a commerce site and wanted to use SSL and have a "plug" saying that you were a good guy, you need a certificate from one of the authorities like VeriSign or Thawte. But, until the last year or so, they wouldn't issue a certificate to any site running a "free" web server. You had to have a commercial server. And of the commercial servers, IIS is by far the best and most well known to web site developers.
-sw
this is primarily only true for closed source products where it's not profitable to continue to dump resources into it that product anymore. people are always going to tinker with software like this, if nothing else to make their lives easier. the other danger is that everyone starts "thinking inside the box" and no new features are added, which likewise isn't true because we all have the source and can tinker if we think of some cool new feature.
i can understand your concern, though. the problem is this: who gets forced to not use linux?
Sorry about the re-post forgot to format. This should be bit clearer
/Tobias
Your probably refering to Hotmail
Try
bash-2.01$ telnet www.hotmail.com 80
Trying...
Connected to www.hotmail.com.
Escape character is '^]'.
GET / HTTP/1.0
HTTP/1.1 200 OK
Date: Fri, 08 Jan 1999 01:07:20 GMT
Server: Apache/1.2.1
(Example taken from http://photo.net/wtr/dead-trees/)
The note about Webjump using IIS for static pages and Apache for "processed" pages is much in line with the Mindcraft test stating that the IIS is good at pushing out bits, but that Apache is good at CGI/PHP compared to ASP technology.
It might be too, that the static pages are not that static and the people changing them, "just wants to press at button", where the "processed" pages often involves a lot of database/monitoring/organising actions that are far better handled in a UN?X environment.
Maybe this is not war but common sense.
:-) = I am happy
:^) = I am happy with my big nose
C:\> = I am happy with my OS