Nginx Overtakes Microsoft As No. 2 Web Server

tsamsoniw writes "With financial backing from the likes of Michael Dell and other venture capitalists, open source upstart Nginx has edged out Microsoft IIS (Internet Information Server) to hold the title of second-most widely used Web server among all active websites. What's more, according to Netcraft's January 2012 Web Server Survey, Nginx over the past month has gained market share among all websites, whereas competitors Apache, Microsoft, and Google each lost share."

Comments

Finally!

[Voyager529]

something that's actually legitimate for Netcraft to confirm!

Re:Finally!

[cwj123]

Wait.. BSD isn't dead?

Re:Finally!

[Anonymous Coward]

Wait.. BSD isn't dead?

They changed its name, they call it "Mac OS X" nowadays.

Re:Finally!

[wvmarle]

Everybody knows it is dead already. We don't even NEED NetCraft to confirm it!

Re:Finally!

[Stalks]

Woosh.

Re:Finally!

[TheRaven64]

Well, it died, but then it came back with sharper teeth and better dress sense. Just keep it out of the sunlight...

Re:Finally!

[williamyf]

No, MacOS X is a heavily modified MACH 3 os Microkernel, running a BSD server in Userland**.

The Server in userland may as well be BeOS, ReactOS, Haiku, or, for that matter, even WinNT.

So no, BSD is BSD, MACH is MACH, and MacOS X is Mach running a BSD Server in userland.

** A similar arrangement, and probably inspired by Mach is the "personalities" of WinNT (for those who rememeber there were the DOS personality, OS/2 Personality, Posix personality and Win 32 Personality).

Quality

[Anonymous Coward]

Nginx is a great product. Not surprised.

Re:Quality

[theshowmecanuck]

Time for some legitimate competition for Apache, it's been a long time. A bit of competition, which they haven't had much of could help them too. I am curious if it is easier to configure than Apache, and how well does it integrate with a JEE containers for serving the static content?

But what a fucking name though. :) I found out it's supposed to be pronounced engine-x. Until I found out I called it enjinx. Reminds me of that movie 'That Thing You Do' where the band called themselves the Oneders at the beginning and everyone called them the 'oh-nigh-ders' or 'oh-need-ders'. Then they got a manager and he forced them to change it to the Wonders because it didn't look garbled. So I say, why not just call it EngineX. It still sounds cool and doesn't have that annoying 'I'm trying to look cool' thing going at the same time. Regardless, sounds like a good product.

The only thing that makes me dubious is that they're based in Russia, I hope Putin and his boys don't have a back door into it. But America is starting to look no better than Russia these days in terms of a government that actually cares about the people. Have you checked out the NDAA that Obama signed this week? It lets the American military arrest civilians inside America (heck Fox news AND democrat supporters are all screaming bloody murder about this one). So on second thought I think I'll give this jinx thing a try.

Re:Quality

[stanlyb]

you have the source code, so what is the problem!!!

Re:Quality

[telekon]

1) Hell yes, it's easier to configure than Apache. Has most of the plugins you could want from Apache, whilst being much more lightweight.

2) I'll echo the other comment here, YOU HAVE THE SOURCE CODE. Worry about backdoors in IIS from the U.S. Gov't., nginx has way more eyes on it.

3) You eventually figured out the pronunciation. Most of the people I know that use GNU/Linux and LaTeX ca't pronounce GNU or LaTeX, but they work great so they get used. What's the problem?

Re:Quality

[icebraining]

http://wiki.nginx.org/NginxJavaServers

Re:Quality

[exomondo]

So I say, why not just call it EngineX.

Just be thankful it's not named Libre Nginx.

Re:Quality

[Anonymous Coward]

Uh whatever... You act like it's some sort of new thing. Some of us have been using it for years. It's a solid, well known piece of software that has been around for a long time.

Re:Quality

[Ethanol-fueled]

I've been reading it as "nig-nix" the whole time, but I never really parsed it. The only time i saw that word until now was 404's and other website error pages.

Re:Quality

[msauve]

"Nginx is a great product."

Strange that 99% of the server errors I see report themselves as being from nginx. I guess there's a conspiracy against it.

(although anything with a config less obtuse than Apache can't be all bad)

Re:Quality

[Anonymous Coward]

The only thing that makes me dubious is that they're based in Russia, I hope Putin and his boys don't have a back door into it.

Right, because the evil Russians can easily hide a backdoor in the source code. You're much safer using software made by a company that obeys the NSA.

Re:Quality

[harperska]

Having never heard of the server in question until this /. article, when I read the summary my mind parsed the name as alternately 'neh-GIN-ex' or 'en-GIN-ex' with 'GIN' pronounced like in 'begin', not like the drink.

Using a solitary 'N' as a syllable at the beginning of a name is ambiguous as to whether the implied vowel should be at the beginning or the end. For example, when I first saw its name printed, I thought the graphics card manufacturer was pronounced 'neh-vid-ee-ah'.

Re:Quality

[man_of_mr_e]

I am quite surprised. nginx may be a good product, but it's also lacking a lot of functionality that a web server used as a load balancer or cache should support. For example, it doesn't support HTTP 1.1 to the backend, thus it can't do name based virtual hosts on the servers it caches.

I *WANTED* to use nginx for a large multi-tennant website we were building, but it didn't support it.

Re:Quality

[MurukeshM]

A consequence of NCurses?

Re:Quality

[GP1911]

They're usually backend errors, nginx is often used as a reverse proxy.

Re:Quality

[Anonymous Coward]

Two words, signing statement. Sure they are highly suspect too, but hey whos counting. Just once I'd like to see a president use a signing statement to say he doesn't feel the need to obey an unconstitutional law. Of course a veto would have been better, its not like the military would have starved overnight while they made the bill comply with the constitution.

Re:Quality

[amorsen]

I don't use nginx, but:

By default, the Host header from the request is not forwarded, but is set based on the proxy_pass statement. To forward the requested Host header, it is necessary to use:

proxy_set_header Host $host;

I haven't heard of a web server which refuses to accept Host headers just because the client sends HTTP/1.0 instead of/1.1.

Re:Quality

[garaged]

I actually did the test, passed all my local config from apache to nginx, the config approach is quite different, it took me like 3 day of getting it 30-60 minutes to config ngix all the things that apache had working but the final config was about 20 lines long, oppossed to a much bigger apache config that has taken year to "master".

Needlss to say, been happy using nginx for a couple of months now on local testing env., still neeed to port production sites :)

Re:Quality

[HopefulIntern]

Oh come on, Nvidia make it easy for you by capitalising the N. That's more than you can say for nginx.

Re:Quality

[mjwx]

3) You eventually figured out the pronunciation. Most of the people I know that use GNU/Linux and LaTeX ca't pronounce GNU or LaTeX, but they work great so they get used. What's the problem?

So it's not pronounced "lay tex" like people expect it to be.

Hell, we're still having the "Lynn-ix" vs "Line-ix" argument where I am.

Anyway, Vietnamese may be having trouble with it, Ng is sometimes pronounced Ngyen.

Should have been written as N-ginX so people understood it.

Re:Quality

[mjwx]

So I say, why not just call it EngineX.

Just be thankful it's not named Libre Nginx.

Yes,

People would think it's a new pad.

Re:Quality

[MarkRose]

I'd recommend HAProxy. It's twice as fast as Nginx is for load balancing, it's just as stable, and it supports more advanced balancing than simple round-robin.

Re:Quality

[tbird81]

Ng is a consonant pronounced the same as it is when elsewhere in a word. Think of the word "singer", and don't say the "si".

Re:Quality

[Bronster]

US Govt? Surely you mean those crazy russians.

I've written a couple of patches for nginx. It's a giant state machine. The code is amazingly different from anything I've worked on before. I can't believe someone manually types that shit - but it all makes sense.

Re:Quality

[dotancohen]

But what a name though. :) I found out it's supposed to be pronounced engine-x. Until I found out I called it enjinx. So I say, why not just call it EngineX.

Also, LEMP is a bit more pronounceable than LNMP.

I do suppose that they could call the web server IngnX and then package their LIMP stack with the GIMP.

Re:Quality

[dotancohen]

Just be thankful it's not named Libre Nginx.

Or IngnX. Because then the stack with Linux, MySQL and PHP will be a nice compliment to the GNU Image Manipulation Program.

Re:Quality

[mrchaotica]

Hey, I mispronounced it so bad I thought Velma was their spokesperson!

Re:Quality

[micheas]

In my somewhat limited experience the place that nginx really shines is ssl.

I am prone to putting varnish on 80 and nginx on 443 with something heavy on the back end to spit out the dynamic content to the reverse proxies,

Re:Quality

[florin]

Ng is a consonant pronounced the same as it is when elsewhere in a word. Think of the word "singer", and don't say the "si".

AH so it's pronounced nig-ger then.

Thanks for clearing that up.

Re:Quality

[Cyberax]

In Russian it's usually pronounced with unambibous 'n' followed by 'g' as in 'begin'.

However, the author's intention was to make it sound like 'engine-x'.

Re:Quality

[jbolden]

nginx uses a different strategy for how it utilizes hardware. It is not so much a configuration problem as a density problem. For things like webhosting companies, which are a huge percentage of websites, nginx cuts costs.

Re:Quality

[dingen]

Sadly, except for some native English speakers, nobody will notice the difference between LAMP and LEMP.

Re:Quality

[TheRaven64]

Most of the time you see a server error, it means that the web server is working but something else (often the database, more often the fastcgi or whatever) has failed. It's when you don't see any error that the web server itself is broken.

Since Apache 1.x is now officially unsupported, OpenBSD has imported nginx into the base system and it will be the included web server in a near future release. I've been using Lighttpd for a few years because it was lighter-weight and easier to configure. Development seems to have stalled in the past couple of years though, and nginx looks like a promising alternative.

Re:Quality

[asdf7890]

You could equally say you see a lot of Apache errors because of broken PHP code, or people sending you mistyped links (hint: that 404 error may not be the web server's fault).

nginx is very often used as a front-end to code writting in other systems like node.js, rails, and so forth: nginx serves static files directly and the other systems serve dynamic content. If the back-end is too busy (or actually broken in some way) nginx won't be able to drag dynamic content from it and will have to report an error instead.

Most sites are configured to hand out a generic 500 "server error" instead of anything more specific like "some fool missed a semi-colon near line 328", as giving out meaningful internal error messages on public facing interfaces is considered a potential security problem (it can make injection flaws easier to find), but don't have custom message HTML so said messages have "nginx" plainly visible on them, so I can understand some confusion on this point (though the same thing affects other web servers too, obviously).

Re:Quality

[fotoguzzi]

Can Slashdot inaugurate a +6 level for comments that are even better than the parent comment? I really think the end is near. I think Slashdot is now kuro5hin--completely pointless.

Re:Quality

[marcosdumay]

If you want to make a reverse proxy, Nginx is the way to go.

But unfortunately, most software that people want to speed-up with a reverse proxy are of bad quality. If it wasn't, it would also probably be fast and wouldn't need that proxy.

Re:Quality

[wisty]

Actually, native English speakers are some of the worst people in the world at noticing the difference between different vowels. You can replace every vowel in many English sentences with "eh", and English speakers will still understand you. In Texas, you can use "ah" instead - "Ah wahnt tah thah stahre, ahnd bahght sahm ahpples ahnd ahrahngahs".

Re:Quality

[Jaro]

This is indeed a real pain. We are using it to load balance our LMS at the university I work at and also offer WebDAV access to files in the system. Files are delivered via HTTP 1.1 chunked transfer encoding which is not supported in the backend. Apparently the nginx guys are working on a proxy plus module with limited HTTP 1.1 support (like chunked transfer encoding). Right now we are using a dedicated machine which is not behind the load balancer for the WebDAV traffic, which of course isn't an ideal solution.

Othewise nginx is really a fantastic little EngineX. ;-)

Re:Quality

[theshowmecanuck]

"Ah wahnt tah thah stahre, ahnd bahght sahm ahpples ahnd ahrahngahs".

lol Translation: I went to the store and bought some apples and oranges.

In Boston it's consonants. Specifically the R. Take the R's out and put them where they don't belong. Try the famous: I parked my car in Harvard Garden without pronouncing the R's. Then you can sound like a Bostonian. Throw in the name Melissa at the beginning. Hey Melissa, I parked... Except it would be Hey Melisser, I packed my cah....

Now I will duck as Bostonians throw rotten fruit at me.

Re:Quality

[theshowmecanuck]

It is not so much a configuration problem as a density problem.

I have to admit, I'm not sure what you mean. Can you elaborate?

Re:Quality

[jbolden]

Sure imagine if I have to maintain 100m sessions, which are 95% inactive on 1m websites my company hosts simultaneously. nginx uses 2.5Mb per 10k inactive sessions. So for the 95m inactive sessions I'm at 23.75 gigs. Say on the order of a dozen servers to handle those inactives.

On the other hand apache can't get to 10k session at all. So I might be at one server per 5k inactive sessions or around 19k extra servers.

That doesn't change anything about the 5m active sessions, and that could be another 10,000 servers easy.

But the difference between 10k servers for apache and 30k severs for nginx is what you are buying.

Re:Quality

[olau]

Also nginx doesn't have the weird bug that made lighttpd crash from time to time.

Re:Quality

[olau]

It's twice as fast as Nginx is for load balancing

Do you some numbers somewhere to back that up? Sounds suspicious to me.

market share v. reality

[girlintraining]

I'm firmly convinced the main reason IIS is even in the top 10 is because so many large corporations sign secret agreements with Microsoft to get discounted software in exchange for not using "free" or "open source" software. No joke -- I am working at a company right now where it is banned, and the only reason given is either that "info security" said so, or "legal" did. But when pressed, nobody can quite identify why. It's just policy, and nobody questions it. IIS' market share is vastly inflated; If it weren't for these clandestine agreements, I sincerely doubt it would be deployed very often, even WITH all the MS tech tie-ins, there's too many compelling reasons not to use it. Even Microsoft doesn't use it on it's major websites because it doesn't scale and it is prone to failure.

Re:market share v. reality

[Kalriath]

Bullshit. Microsoft uses it on every site they have. The only reason that "web server identification" surveys like Netcraft say they run Linux is because, like all large websites, they utilise the services of a CDN such as Akamai.

And there are no "secret agreements". Most of the time the company forbids such things is because there is no support, or because there is no ability of the in-house technical support to provide assistance with it. We're a very large IT company here and we have maybe 3 RHEL servers (because Linux was the best option for the task) and a couple of thousand (including virtual) Windows servers. (There's also about 2 Solaris servers, 4 or 5 Oracle Linux servers, a SCO Unix server and 2 or 3 HP-UX servers). None of this is due to any "secret agreements". It's all because there's one person trained to work with Unix based systems, and about 8 to deal with Windows. We utilise quite a large number of open-source packages across our infrastructure if it's the best tool for the job.

Re:market share v. reality

[Anonymous Coward]

In my last job, we had a client with an all Windows environment. We're talking 2 DCs, a file server, an exchange server and a dedicated IIS server on the other side of the firewall and off the domain.

One day, they decided to revamp their static HTML website (this was a government department trying to justify their existence, IT wasn't exactly at the top of their list). We talked to the outfit doing it, who told us they were using PHP. Great, I though. We can get rid of an old and outdated Windows server and replace it with a nice, lean little Linux box. Nope, I was told to install the PHP ISAPI module on IIS, because "we were a Microsoft shop", even though this server was quite literally doing nothing but serving up HTML and chewing up an unnecessary Server 2k3 license. So after much fighting, and arguing to explain that we may as well NOT go through the trouble trying to set up and debug PHP as FastCGI, another guy went behind my back and stuffed up the install, leading to me wasting 3 or 4 hours rolling it back and installing it properly. Anyway, it's all smoothed over, until I get the zip file I've been promised by the "website makers". It was indeed a website, 10 or so DreamWeaver files with the extension renamed to PHP. No Drupal theme, no Joomla install, nothing. -.- God I hate the people in this industry that like to sell themselves as professionals

Re:market share v. reality

[poetmatt]

You kidding?

Why do you think live365 exists?

It's that exact agreement. They don't say "don't use free or open source software" they just say "don't use any of our competition". It throws in the whole microsoft suite (office, sharepoint usually in the face of wikis or better solutions, live365, etc), always with the argument of "we have a MS specialist to help you migrate" (even if that won't fix problems).

agreement cost = substantial.
cost of all the microsoft stuff non-agreement = easily 10x as much.

So you see corporations sign up for this agreement as fast as possible if it's a good MS sales rep, even if it's a horrible long term outlook as any company smart enough to look down the road would know you never, ever want to rely on a single vendor for everything as that's a single point of failure.

Re:market share v. reality

[icebraining]

Even Microsoft doesn't use it on it's major websites

curl -v www.microsoft.com
Server: Microsoft-IIS/7.5
X-Powered-By: ASP.NET
 
curl -v www.hotmail.com
Server: Microsoft-IIS/7.5
X-Powered-By: ASP.NET

What sites are those, exactly?

Re:market share v. reality

[epiphani]

There's some truth to this.

Several years ago, GoDaddy switch all of their domain parking to IIS, explicitly to get microsoft's numbers up. Throw 10,000 cnames pointed at a single machine serving up parking pages, and boom - 10,000 websites running IIS.

Re:market share v. reality

[ATMAvatar]

You mean like all the press Microsoft got for the very same behavior with Windows and Internet Explorer prior to US v. Microsoft?

Putting your trust in the media to inform you of anything beyond celebrity news or Republican primaries imitating reality TV is foolhardy.

That said, it would be nice if the GP provided more than flimsy personal anecdotes.

Re:market share v. reality

[Zalbik]

Please stay firmly convinced...it will make it easier for me to recognize you as a complete idiot if I ever have to interview you.

More and more I'm finding the best way to recognize talent is to find people who understand how to apply the right technologies to a solution rather than the mindless ".NET is da best!" or "Windoze sucks!" I keep hearing from the typical zealots.

Re:market share v. reality

[Daniel Phillips]

We're a very large IT company here and we have maybe 3 RHEL servers (because Linux was the best option for the task) and a couple of thousand (including virtual) Windows servers.

Wow, it must suck to work where you work. Could you please tell me the name so that I may have nothing to do with it?

Re:market share v. reality

[Daniel Phillips]

If Microsoft were making such clandestine deals it would be all over the press...

Microsoft was found in court to have made many clandestine deals and none of them hit the press at the time, that I can recall. I find it entirely believable that Microsoft is still today using the same tactics, as the only punishment they ever got for their wrongdoing was a pat on the wrist. Plus paying out $billions in private suits, but Microsoft just regards that as a business license.

Re:market share v. reality

[Zalbik]

What the heck?

The office live 365 agreement says nothing about not using competitors software.

I think you need a new tinfoil hat...

Re:market share v. reality

[bloodhawk]

It's not about faith in the media, It's about faith in the inability of any company that is screwing people over to permanently silence millions of companies from saying anything anywhere. Hell they couldn't even manage to silence everyone for there patent agreements and we KNOW they were trying to do that, yet somehow we are supposed to believe they can successfully silence an exponentially larger group of people (many of whom dislike them) while at the same time screwing them over?

Re:market share v. reality

[hcs_$reboot]

$ curl -v www.skype.com
HTTP/1.1 302 Found
Server: Apache

Too new, maybe...

Re:market share v. reality

[Anonymous Coward]

The deals were public knowledge. The OEM deals of the time that resulted in the court cases were actually legal until MS was declared a monopoly. Companies have always made such deals and continue to make them, they were not anything special and only become a problem under monopoly rulings.

Re:market share v. reality

[flimflammer]

If they already had the Server 2k3 license paid for, what reason would they realistically really have to drop it for Linux when they can do what they want with what they have? The deed is already done shelling out for the license, putting Linux on it is just like pouring salt in the wound. Anyway, as you said, the site is just from a government department trying to justify their existence, I doubt it really matters what OS is serving up the content. It's not like they're Google or something.

Re:market share v. reality

[parlancex]

...go through the trouble trying to set up and debug PHP as FastCGI...

I think it's funny how common it is for anyone who mentions working with Windows professionally on Slashdot to be called out for being inexperienced or some kind of unauthentic system administrator with no real skills, but no doubt there are just as many who consider themselves experienced *nix system administrators who I could make fun of for being inept at basic Windows administration tasks.

Anyway, there are plenty of good reasons that web server should have been a Windows box. Even if it wasn't joined to the domain by switching that box to Linux they would lose the ability to leverage their existing update (SUS) and backup infrastructure. Also, the cost of a Windows license for a small shop like that would pay for itself probably 3 times over if they had to even try to get some kind of professional support for the Linux box even once.

Re:market share v. reality

[SadButTrue]

This has probably always been the case. However, using .NET means buying the entire Microsoft stack.

At my last job I wrote an entire back office in Java. When my company merged the decision was made, over my vehement protests, that we would recode in c# just to support a thick client that was the bread and butter of the traders at the other company. Literally everything had to be moved just because it had been marginally easier to code a desktop app in c# initially.

Microsoft makes some good stuff, they really do. But since MS stuff only works well, or at all, with MS stuff you may end up taking a heavy does of shit along with the good.

Re:market share v. reality

[Anonymous Coward]

Actually www is fronted by reverse proxies running on linux (you can check with nmap). IIS apparently can't handle the load by itself.

And, hotmail took a long time to convert to M$. They tried once, then had a two week outage since M$ didn't scale worth a shit. They rolled back to Solaris, and kept it that way for a _long_ time.

Fun fact. Microsoft also used sendmail on sun boxes for internal mail for a long time after exchange was introduced. M$ couldn't get their own software to scale to an enterprise as large as their own.

Re:market share v. reality

[timeOday]

Are you telling me this survey is by number of domains hosted, irrespective of number of pages served? That would be absurd.

Re:market share v. reality

[gstrickler]

But it's all MS, so it all integrates and works together seamlessly, right? And it's all managed centrally using AD, right? I mean, that's got to result in significant savings, right? Oh, and the agreement covers all our computers and all this extra software that would cost us so much separately, so we're really getting a great deal on it. Please, MS Salesman, tell us something that's easy to believe and we'll sign.

Re:market share v. reality

[hcs_$reboot]

Yet another GoDaddy interesting story :-) I'd mod you up... if I had some mod points and if I wouldn't have already two (now three) posts in this thread...

Re:market share v. reality

[Anonymous Coward]

Wow, it must suck to work where you work. Could you please tell me the name so that I may have nothing to do with it?

Microsoft Corporation, Redmond WA

Re:market share v. reality

[Dynedain]

Also, the cost of a Windows license for a small shop like that would pay for itself probably 3 times over if they had to even try to get some kind of professional support for the Linux box even once.

A Windows license doesn't magically come with professional support. And honestly, if you need professional support for a server *NIX is going to cost you the same as an equally competant Windows admin.

If you can't handle management of a web server in-house with qualified staff, you should move to a hosted solution. It will cost less regardless of OS choice.

Re:market share v. reality

[parlancex]

It sounds like it probably should've been hosted by a 3rd party based on the GP's post, but the point I was making about support is that if I was a small shop without in-house IT and I need support for a specific issue and I open up the yellow pages and look at my options there are going to be way more options for me that will be less expensive if I need to hire someone to take a look at a Windows server.

Re:market share v. reality

[LordLucless]

I need support for a specific issue and I open up the yellow pages and look at my options there are going to be way more options for me that will be less expensive if I need to hire someone to take a look at a Windows server.

And most of those will be sixteen year old kids who think that the fact that they know how to find the control panel qualifies them to administer a windows server.

If you're looking for qualified, certified, experienced administrators, there are plenty in both worlds. The BSDs and Linux have dominated the server market for a long time; there's a very large pool of talent to draw from.

Re:market share v. reality

[Daniel Phillips]

If Microsoft were making such clandestine deals it would be all over the press...

Microsoft was found in court to have made many clandestine deals and none of them hit the press at the time, that I can recall. I find it entirely believable that Microsoft is still today using the same tactics, as the only punishment they ever got for their wrongdoing was a pat on the wrist. Plus paying out $billions in private suits, but Microsoft just regards that as a business license.

Nothing but the facts there, mod points won't change that.

Re:market share v. reality

[clarkn0va]

leverage their existing update (SUS) and backup infrastructure

Yawn. I work in a mixed environment and I know exactly how precious (read: expensive and touchy) that SUS and backup infrastructure can be. You're right on one point, that's not something you want to have to duplicate unnecessarily.

Fortunately, Linux systems generally include excellent backup tools that need only a few minutes and frankly comparatively pocket change to acquire and configure. Once you grok these you will look back at the whole Windows backup circus with mixed horror and confusion.

Re:market share v. reality

[CAIMLAS]

There are tens of MSPs out there with dipshit MCSEs and people who are better than that who are able to fix your average Windows problem. (If they can't fix it then they probably know a Linux person who can figure it out. :P)

Re:market share v. reality

[gl4ss]

why would it be absurd? maybe it's per ip.

if it's per pages served, then whatever is google.com + facebook.com run gets top spot. but that would be guesswork estimating.

Re:market share v. reality

[gstrickler]

Yes. RTFA.

Re:market share v. reality

[Serious Callers Only]

And there are no "secret agreements".

MS has widely used secret agreements in the past to exclude competitors from a market, what makes you think they are not doing this now?

Re:market share v. reality

[Bert64]

The "issue" here is one of confidence...
You have thousands of 16yr old kids and people who "have used a computer at home to play games" who think they're experts and start offering windows support.
It's very rare for such people to have even heard of unix, so you don't get useless offerings of unix support.

When it comes to competent support, this is generally available for both platforms, there are usually similar numbers of people available and they generally cost about the same.

In short, useless windows "support" is widely available and cheap while cheap useless unix support is rare... While competent support is available for both, it generally isn't cheap.

Re:market share v. reality

[Bert64]

Support companies, unless you contract out your entire infrastructure to them, won't have a good overview of your setup...
Specific support companies are even worse, because they generally only understand their own product and have very limited understanding of anything else it may interact with, in many cases not even for the underlying os their product runs on.

Re:market share v. reality

[dbIII]

who I could make fun of for being inept at basic Windows administration tasks

At various points everyone was uniformly inept at basic Windows administration tasks until useful tools were developed and documented. It's at those transition points where everyone looks useless and start scrabbling for something that isn't there anymore or even worse is still there but doesn't work properly anymore. With wide beta releases of MS software those days seem to have gone but it has left MS Windows admins with a bad reputation that they do not deserve (I'm a *nix admin so all the idiots that like to blame the messenger can go and look for a different target).

Re:market share v. reality

[wvmarle]

And there are no "secret agreements". Most of the time the company forbids such things is because there is no support, or because there is no ability of the in-house technical support to provide assistance with it.

There is a difference between white-listing certain applications because that's what a company decides to support, and adding to this list whenever there is a need, and a blanket ban on free or open source software where it's banned for being free and/or open source.

The first I can totally understand. You can't support just everything.

The second totally not. There is no reason why a software packed becomes easier to support by an IT department because the source code is not available.

Re:market share v. reality

[Anonymous Coward]

Working at a company that let me sysadmin (with a team) a few thousand linux servers, please let me just LOL at that.

Our PCI env has like 2 windows virtualization servers, agaist some 150 linux virtualization servers (more than 600 VMS), what dpes that tell you about windows viability?

Re:market share v. reality

[RubberMallet]

You haven't been paying attention.... have you?

Microsoft was caught out blatantly bribing the ISO guys (and it was widely reported in the press) to pass the OOXML standard.

Microsoft has been caught out over and over signing deals with companies to be the exclusive OS provider... and it's been reported in the press.

I've worked in 2 companies who had these "clandestine" deals with Microsoft where the company would lock out ANY competing products in return for deep discounts or even free software and support.

My mother has run a SOHO IT company since 1990... and guess what, even she was offered and accepted a deal with Microsoft (due to sales volumes) to be 100% exclusive to MS... the deal included recycling software on updates to the local school system, ensuring that they also got MS only, and Linux was banned in the schools as part of the agreement to get free MSO and MSWindows licenses.

In other words... it happens... it's not FUD.

Re:market share v. reality

[JasterBobaMereel]

You have 8 people that deal with windows servers because you have a lot of windows servers

You have a lot of windows servers because you have 8 people that deal with windows servers ....This is why windows is still popular

Re:market share v. reality

[tiedemann]

Most server-side .NET code works just fine on Linux using Mono. It does however require not wearing a tinfoil hat.

Re:market share v. reality

[tiedemann]

...and it works well with nginx http://www.mono-project.com/FastCGI_Nginx

Re:market share v. reality

[St.Creed]

Could be any big multinational, basically. I don't see Linux used much outside universities and city councils, and even there the employees are bitching about it. Notable exception: a lot of Oracle databases run on RHEL(*). Apart from that, most of the internal servers I see are all Windows. With VM/ESX they're easy to manage and roll out. But MS SQL Server sure doesn't run on Linux and that is becoming the database of choice for a lot of smaller installations due to the braindead Oracle policies. Although if you ask me, these outfits would probably be better off with PostgresQL.

(*) But even Oracle doesn't support that very much. Try installing Oracle Warehouse Builder on a new version of RHEL, even one minor patchlevel above the certification - you'll spend weeks on the phone with Oracle support. Last time round it took them 6 weeks to get most of the errors fixed but we still have random errors.

Re:market share v. reality

[am 2k]

Unlike smelly hippies, they're not going to trash a perfectly good service just because of rah rah candy ass confusion of software and ideology.

I guess you weren't around yet when they switched Hotmail from BSD to Windows?

Re:market share v. reality

[b0bby]

I'm not in a large environment, but I'd bet that they have 8 people who deal with Windows servers because they have a lot of things like Exchange and Sharepoint running on Windows servers. Even in my small environment, I have Windows servers running Exchange and the voicemail etc; the few Linux boxes I have are doing what they do best, but the Windows boxes are needed and are doing the most important stuff from the point of view of the company..

Re:market share v. reality

[TheRaven64]

Seriously, anyone who thinks Microsoft would turn around and immediately swap out OS/webserver because herp derp, needs to be fired

So, someone who pays attention to history should not have a job, according to you?

When Microsoft bought Hotmail, it was running on FreeBSD. This was slightly embarrassing for Microsoft and they made a big show of switching it over to Windows NT. It then died horribly, and they quietly switched it back. The second time they did the transition, it worked. Neither time was done for technical reasons - hotmail had already demonstrated that it scaled very well with the old infrastructure - it was done purely so Microsoft could point at Hotmail and say that they were using a 100% Microsoft back end.

Re:market share v. reality

[b0bby]

If they really are a "Microsoft shop", why would throwing in some random Linux box which is unlikely to ever be updated improve security? Sometimes it's better to standardize rather than go for the otherwise optimum solution. I run my intranet on Linux, but I enjoy doing that; if I were to be replaced there's no guarantee that the next guy will be able to deal with it.

Re:market share v. reality

[asdf7890]

Actually www is fronted by reverse proxies running on linux (you can check with nmap). IIS apparently can't handle the load by itself.

Probably true, but disengenuously presented. The front-end is run by a 3rd party (Akami?) CDN so MS don't control what sofftware they chose to use for the task. It MS decided it was worth running their own CDN instead of outsourcing the task, their's would probably run un Windows+IIS rather than anything else.

Re:market share v. reality

[asdf7890]

It wasn't just GoDaddy IIRC, but they were by far the biggest registrar to be offered a little bribe to that end.

Re:market share v. reality

[poetmatt]

Considering it's not a live365 agreement, I'd agree that you're correct. Did you not realize this?

It's not about live365. I'm saying that live365 is bundled in as part of the agreement instead of paying a monthly cost per user and is instead thrown in for free. It's the only reason you'd see live 365 in enterprise in a microsoft setup, because otherwise google applications is both a: cheaper, b: more flexible and c: more capable/cross platform friendly.

Re:market share v. reality

[trippyd]

Throw 10,000 cnames pointed at a single machine serving up parking pages, and boom - 10,000 websites running IIS.

I hate to say RTFA, and maybe click on some of the links inside, but you should. One of the links is this one describing how they determine what is an active site.

TL/DR, it says that your 10,000 websites running IIS would be completely ignored by the Netcraft survey of "active" sites, they have an algorithm in there to completely ignore templated "parking" pages and the like.

Re:market share v. reality

[SuhlScroll]

They *are* professionals ... professionals at keeping themselves employed. Companies will learn sooner or later that trying to pay IT people as little as possible will always bite them in the end. He who controls the computer controls the content.

Re:market share v. reality

[CBravo]

it is called job security combined with cognitive dissonance.

Re:market share v. reality

[140Mandak262Jamuna]

A: Have you seen elephants hiding behind trees?

B: No.

A: That shows how well they hide

If you knew about these agreements, they would not be secret agreements. Right?

Re:market share v. reality

[TemporalBeing]

Windows is easy to admin if you're any good.

So long as what you do fits the model that Microsoft has set out to provide. Otherwise it a PITA.

Re:market share v. reality

[Sique]

I don't see Windows Servers much in use exept for some additional application servers. The company I work for has all their main products running on Linux for the server, most of the larger clients, and sometimes VXworks for small clients. Windows Server is mostly used for products the customer is supposed to use directly as a local administrator, or for reporting and statistics. (And it's not a small company, we have about 1 mio enterprise customers in 160 countries).

Re:market share v. reality

[darkpixel2k]

Blind hatred aside, what makes you think they are? Really makes you wonder how in the world MS is posting record profits, when they're apparently paying everyone and their mother to use their products.

Wasn't there something during 2011 about Nokia and a burning oil platform--and just when everyone got excited and thought they were going to say 'Android', they said 'Microsoft'. Then their stock tanked. Then it came out that Microsoft paid them billions to continue standing on that burning platform?

Re:market share v. reality

[Kalriath]

Virtualisation servers? Come on, even we use VMWare ESX for that. It doesn't really prove anything.

Re:market share v. reality

[Kalriath]

A: Have you seen elephants using mac computers?
B: No.

A: Clearly the elephants are being paid by Microsoft not to use macs.

Really, that's what half the irrational anti-Microsoft crap on this site sounds like.

Re:market share v. reality

[Kalriath]

I considered responding, then noticed where you rambled on about "garbage" and shit, so I decided not to bother because there's no way I'll ever get you to think rationally where the word "Microsoft" is concerned.

Re:market share v. reality

[Kalriath]

You do realise there are entire industries where there are no commercial Linux based offerings and all the open source offerings are festering piles of shit, right? Examples include healthcare, water processing, pretty much anything government related...

Re:market share v. reality

[aztracker1]

There's always the Mono stack, but that takes some more care than Java development.. though I'd rather have P/Invoke over dealing with JNI bridging.

Re:market share v. reality

[am 2k]

And you weren't around to realize Hotmail was a separate entity until Microsoft bought it. If anything it proves they eat their own dog food.

Uh, that's totally beside the point. Microsoft bought a well-working BSD-based webmail service, and moved it to the Windows platform for no technical reason. The result was a lot of issues and downtime.

Re:market share v. reality

[ekgringo]

That tells me you probably have more Linux admins than Windows admins or that the primarily Linux-oriented admins are also able to manage Windows. It tells me nothing at all about Windows viability except perhaps at your company.

Re:market share v. reality

[Zalbik]

Then I hate to say it, but your architecture sucked.

At the very most (assuming whoever wrote the thick client was a complete idiot and coded it to ONLY worked against .NET), you should have only needed to write an adapter layer in C# to adapt whatever interface the thick client worked against to whatever interface your java system worked with.

If you had to rewrite the entire stack, something was either very wrong with your system, or very wrong with your ability to explain choices and costs to your manager.

Re:market share v. reality

[Bert64]

If IIS is so great, then why do all the large CDNs use linux then?

Re:market share v. reality

[Kalriath]

Because CDNs aren't ordinary web servers? They sure as hell aren't running Apache either.

Re:market share v. reality

[Kalriath]

No, we have 300 IT people. The dozen are merely the server admins.

Wow, never even heard of that one

[Nursie]

I've used a load of web servers in the last few years - an early verion of IIS when I had only windows many years back, apache, lighttpd, thttpd, netscape web server (showing my age) and various others... but I didn't even know this was out there.

Suppose it just shows how out of the loop I am these days. Computer stuff covers a vast field these days.

Re:Wow, never even heard of that one

[bradgoodman]

I was going to say the same. So I was pretty surprised. From what I am reading, it is more of a "front-end" system for web servers, that does things like caching and load-balancing. So I guess it sort of depends on ones definition of "web server".

I was also going to speculate/wonder if it was one of those "rigged" deals, like a few years back when IIS was declared as "overtaking firefox" and becoming #1 because "most web sites on the web used it". The actual reason was that GoDaddy (which hosts a vast majority of "parked" domains) was paid-off (or "otherwise incented" by Microsoft to switch to IIS. So when you considered a "www." to be a "unique site", and 99% of "unique sites" to be garbage parked-domains, IIS was not the leader.

So, I wonder if some other bizarre statistical work is at-play. For example, does someone like Akamai, who hosts a lot of other people's sites, use Nginx to skew these numbers??

Re:Wow, never even heard of that one

[asserted]

no, this is genuine. it has been steadily gaining popularity over the past several years.
nginx is being developed by a russian guy who up until recently was working (as a sysadmin, apparently) for one of the major russian web portals where nginx originated as an in-house project first but was open-sourced. the guy has now left the company (which has been slowly dying anyway) and incorporated an llc or something, focused on nginx. it was already quite popular in russia 5-6 years ago (when i was still living there).

nginx is an efficient event-driven front-end server, quite often used for loadbalancing in front of traditional apache or tomcat or whatever other backends, but in a simple case of a LAMP server it can be hooked up directly to PHP via FPM or FCGI.
config syntax is quite expressive, with quite advanced uri / header - based rewriting capabilities. there is even a built-in Perl interpreter for more advanced use (which tends to be abused by people who forget what being an event-driven server means by sticking logic in there... oh well, people use things like node.js too *shudder*).

Re:Wow, never even heard of that one

[Anonymous Coward]

Nginx + PHP-FPM is the way to go for a number of high performance setups.

Re:Wow, never even heard of that one

[gr8dude]

> oh well, people use things like node.js too *shudder*
Can you elaborate on that? What is the matter with node.js?

Google Icon

[Anonymous Coward]

Why is the Google icon on the post when its MS that got overtaken. Is Nginx run by Google?

Re:Google Icon

[Hermanas]

The Google icon is probably there because the poster mentioned that Google's webserver, among others, lost share the past month.

Nginx is not a website, but rather a webserver

[lethe1001]

"Nginx over the past month has gained market share among all websites, whereas competitors Apache, Microsoft, and Google each lost share." Should read "webservers", not "websites"

Re:Nginx is not a website, but rather a webserver

[Anonymous Coward]

The usage of websites is correct. Try this similarly-constructed sentence on for size:

"Guinness beer has gained in popularity amongst college students, while competitors Molson's, Budweiser and Pabst each lost ground."

Re:Nginx is not a website, but rather a webserver

[viperidaenz]

Perhaps they mean websites, because its impossible to tell how many different webservers host a specific website. If I have 1,000,000 IIS servers behind a load balancer hosting a single website, it would be counted as 1, not 1,000,000.

Re:Nginx is not a website, but rather a webserver

[LordThyGod]

Most of the big names in that league (Facebook, Google, Twitter, etc), are running Linux. So, yes that's quite true.

Nginx is primarily a cache engine

[Anonymous Coward]

The article and summary are misleading, typical slashdot. Typically nginx is used as a forward cache engine, often on the same box as apache. People typically put apache on port 81, and nginx on 80, and configure nginx to cache from port 81...

Doesn't make it the number 2 web server. Yes perhaps the number 1 cache engine, but its generally not used as a web server.

Re:Nginx is primarily a cache engine

[telekon]

Not always... for our Rails (and Sinatra) projects, we use nginx as the frontend/static asset server to a (pool of) Ruby-based application servers (mostly Unicorn). There's no Apache anywhere in the mix, and that has greatly reduced my migraines. Perhaps in some situation it makes sense to have nginx as a cache engine or load balancer for Apache, but in my world, nginx usually replaces Apache, rather than supplementing it.

Re:Nginx is primarily a cache engine

[Scaba]

Can you link us to your data?

Re:Nginx is primarily a cache engine

[Dr.Hair]

I've dropped apache completely. I've got nginx in front of php-fpm (wigh fcgiwrapper/spawn-fcgi for things w/ legacy cgi like nagios/collectd) and uwsgi.

You can easily find appropriate nginx rewrite rules for the major php apps like wordpress and menalto's gallery2. And for performance's sake it's all in the server config, so there isn't a disk access to read the .htaccess file to figure out if there are rewrite rules that need to be considered.

It's even better for apps, like django, which keep in mind delivering static files out of the hands of the app server. And configuring nginx to serve static files instead of hitting app servers is a piece of cake.

And for those really wanting max performance, there are plugins for direct access to memcache/postgres or even writing your web app directly in the config file with something like the lua plugin.

Re:Nginx is primarily a cache engine

[xiando]

The article and summary are misleading, typical slashdot. Typically nginx is used as a forward cache engine, often on the same box as apache. People typically put apache on port 81, and nginx on 80, and configure nginx to cache from port 81...

You do know there is something called mod_proxy for apache? Apache can be configured as a proxy or a web server. Nginx can be configured as a proxy or a web server. Your point is .. what, exactly? I use nginx and I use it as a pure web server. I do not know what everyone else uses it for, but you can't just go about assuming whatever.

Re:Nginx is primarily a cache engine

[aintnostranger]

how is such non information/flamebait modded "3"?!?

Re:Nginx is primarily a cache engine

[Terrasque]

A gun can be used as a hammer, and a rock can be used as a hammer.

But we still typically use a hammer as a hammer.

Re:Nginx is primarily a cache engine

[Thing 1]

A gun can be used as a hammer, and a rock can be used as a hammer.

But we still typically use a hammer as a hammer.

Just don't use Mt. Dew as a hammer! (Protip: the edge of the bottom works great; the side of the can, you only get one strike before it's sticky everywhere.)

Great loadbalancer

[Anonymous Coward]

Nginx is a great loadbalancer for http which makes it quite suited as a frontend and thus getting counted by netcraft . There could be hundreds of apache servers behind it . E.g. on my boxes Nginx runs as a reverse proxy in front of about 20 different apache, tomcat, more Nginx, other servers that generate some kind of html. But these 20 will all be counted as Nginx while they actually run something different. So I beleive it is quite hard to actually say what Server actually is the most popular.

Re:Great loadbalancer

[butalearner]

Good info, but not entirely complete. What I was told while researching webservers was that nginx also excels at serving static files. Personally I have a PogoPlug v2 serving three static HTML websites (static HTML is generated when I change something to make it look dynamic) plus a few binary files. I've never run Apache or any others, but the resource usage is extremely low, even under some load.

Re:Great loadbalancer

[inKubus]

Yep, it's mostly used for front-end duties like connection pooling, load balancing, SSL offloading, gzip, that type of thing. If you're running PHP stuff, it's still debatable whether you want to go FCGI or PDM instead of Apache's built-in module. There are ups and downs in both cases and you'll have to see what works best for your site. At my company we use Nginx up front (with server type obfuscated) for SSL offloading and gzip and connection pooling. From there it goes into a varnishd cache on the same server (stored in 100% RAM) which handles the static stuff. Varnishd then forwards remaining requests to an L7 load balancer appliance type thing which then drops requests to each of 10 web "application" servers which are a combination of Apache with mod_php, Tomcat and Jetty Java servers. We've also used Nginx as an IMAP proxy and cache and it works quite well for that.

Apache has a good architecture but it's horrible at handling a lot of simultaneous connections and recycling them (that will change in 2.4 but it's not out yet). Also, if you're using mod_php, over time each Apache process will take the total maximum amount of RAM your php process uses, and many of our PHP applications use 128-256M of RAM or more (data management type stuff). So you can run a server out of RAM if you're trying to maximize connections.

Nginx can handle 10K connections on a little box with very little RAM due to the way it threads stuff. It's basically a copy engine and it's very fast. Varnishd can also handle a lot of connections and can serve up content straight from RAM in less time than apache takes to build a connection. That being said, Apache is reliable, and has I feel better logging at the moment and just more of everything. It's a reference implementation. It's actually fine for most purposes but if you're handling 1000 users simultaneously and they are making 10-20 connections each with various service calls and static downloads, you gotta have something that can pool the conenctions on the front end and handle static content or you're going to spend a lot of money on RAM. And if you're serving up static content with Tomcat, Tomcat is absolutely garbage. I think it has to boot the whole JVM to serve up your one file. If not that bad, it's still awfully slow, and it REALLY benefits from caching up front. BTW, Nginx does caching as well but varnishd seemed more mature and elegant.

Now lastly, you can just go out and buy an F5 BigIP and it does all this stuff on specialized hardware (Ok, special board, intel chip) and it's out of the box. But even the little ones are $20K which is a lot of software dev hours and/or web server/database/storage hardware. Would be nice and fun to have but if you can't spend the money on hardware (and training!) the nginx/varnishd frontend is pretty much the best setup in my book at the moment. A little complex but once it's set up you just let it run. I made an internal nginx cache for all our internal sites, including some Java apps (e.g. Jira) and with requests going through the cache everything just flies. If you use sharepoint on IIS, you would be prettty stupid to not try a cache server up front, it's amazing. If nginx fixed mod_rewrite stuff to be the same as apache, it would probably be possible to make it into an application server, and we're going to get a test environment set up with php-fpm and see how it fares. We'll see how managable it is though.

Re:Great loadbalancer

[Atzanteol]

So it's used more like squid than Apache?

Re:Great loadbalancer

[styrotech]

Tomcat is absolutely garbage. I think it has to boot the whole JVM to serve up your one file.

No. When Tomcat starts up, that's your JVM boot right there. It stays running until you stop Tomcat. All the apps and files served by that Tomcat instance are served using the same JVM process. It will spawn extra threads for extra connections when required though.

In fact most other web stuff uses a long running process like that. It's mostly only CGI or 'CGI-like' configurations and PHP that work by starting up from scratch for each request.

But yeah, nginx is great :)

Re:Great loadbalancer

[inKubus]

I don't know, maybe there's some session management and auth stuff that's slowing it down but I have not noticed it being quick, or even as fast as Apache. I mean, it's been pretty common practice to put apache on the front of tomcat for quite a while and use redirection to serve static content. With a lightweight cache/reverse proxy it's even easier and faster.

dodgy

[viperidaenz]

the methodology for determining "Active Sites" only takes into account the structure of the html elements of the page. If the structure of the page stops changing its considered not active. javascript heavy sites don't require any html structure change to continue to provide changing content.

...and efficiency

[hcs_$reboot]

I'm glad such a program, well designed and programmed in good old C, is rewarded with trust and confidence from more and more engineers.
I have been using it for two years, serving several professional sites, and the transition from the initial Apache setup was surprisingly smooth.

What I like in particular, compared to Apache :
- fantastic performance gain, in terms of cpu and memory
- maintenance gain: the configuration appears (at least to me) to be more "developer like", and easier to configure/extend with many options
- load balancing is ... really a piece of cake

The only drawback I (initially) found was the lack of a PHP embedded/module. But using php-fpm happened to be a good alternative, via a local port.

Oh, lovely

[gman003]

Guess what class I had today?

CIS311 - Web Server Management

Guess what we use!

IIS 7 and Windows Server 2008!

Good thing I've run both Apache and lighttpd for personal experience. And taught myself C, C++, PHP, Lisp, Perl, Python, and a little bit of Assembly. And MySQL. And how to run Linux from the command line. And... what the fuck am I paying this college for, again?

Re:Oh, lovely

[Anonymous Coward]

You're taking a 300 level course and it consists of running an IIS server on 2008? What the hell school do you go to? "Rick's College"?

Re:Oh, lovely

[l0ungeb0y]

And this is why I thank god I just up AND DID Web Development and didn't shit my money down the toilet on a CS degree.
So yes -- Why are you paying for college?

For most Software Development, the "University of Google" is a far more cost and time effective education.

Re:Oh, lovely

[Anonymous Coward]

Actual CS degrees don't have classes teaching you how to use IIS.

Re:Oh, lovely

[Confusador]

And... what the fuck am I paying this college for, again?

The ability to get your resume past HR and in front of someone who knows what any of that means. Sorry.

Re:Oh, lovely

[kamaaina]

And this is why I thank god I just up AND DID Web Development and didn't shit my money down the toilet on a CS degree.
So yes -- Why are you paying for college?

For most Software Development, the "University of Google" is a far more cost and time effective education.

Specializing in Web development is OK, not knocking that, but doesn't school also teach you about stuff you don't want to learn.

Back in the day I didn't want to do Unix, but I was forced to write C programs on Unix box. Learned vi (and a little bit of edlin), how to grep and pipe, processes and spawning.

Saves the day cause I understand how some of the stuff works.

On ESXi someone tried to delete a folder that contained a running VM, I don't think there is anything on google on bringing that back, but knew it was still running, knew about file handles, VMDK was still there and was able to recover the VM. Would have been gone if they rebooted the host.

Hardware class tought me about boot sectors and was able to recover a physical hard disk.

Being forced to telnet to port 80 and 25 taught me about protocols.

Yeah it was 4 years, but sure makes it easier to figure out stuff, pretty sure I am at a higher pay grade then other Sr Level Engineers in my company.

Maybe you can spin up a web page faster then me, I would higher you, but they would trust me to go and talk to the different departments, come up with the requirements and schedule. No I am not a PM.

Re:Oh, lovely

[HopefulIntern]

Yep, this. A CS degree these days is nothing more than a piece of paper that you often need to be considered for a job. It does not (necessarily) teach you the tools for the job. I am in the inverse situation to GP, my university taught only non-MS stuff (Java, Perl, LAMP) and my job now requires me to use Windows Server 2008 and IIS7. I had never even seen IIS before I came here..

Re:Oh, lovely

[Pecisk]

Guess what pays for this course to be implemented? Hint: not you, but Microsoft.

I know, I know - why the fuck college need so much money (from you and from Microsoft)? Because it is run by greedy bastards, that's why. What's more sad, that they are educated bastards and they think they can require so big bucks. They justify that.

Re:Oh, lovely

[gman003]

Oh, the teacher openly admitted that Microsoft set the requirements for the class. If we used anything else, even IIS6 on Server 2003, we'd lose our MSAA license.

Re:Oh, lovely

[gman003]

CIS311, not CS311. It's a "Computer Information Services" class, not "Computer Science".

Re:Oh, lovely

[petteyg359]

I wouldn't hire you.

For a good reason: nginx really performs!

[Anonymous Coward]

I am the main server admin for a very large website that has been running Apache for 10 years. Then, last year, after a period of tremendous growth, we began to encounter serious memory/CPU issues with Apache. I had been researching alternative, light webservers for a while, so after thorough research and testing, we made the transition to nginx overnight with resounding success. We've never looked back! It is very easy to configure, ridiculously scalable and highly extensible. There are plenty of how-to guides and recipes for those moving from Apache. Nginx seemed like a no-brainer. Apache is a great reference server; it has every bell and whistle imaginable, but at a cost. Our site uses PHP, so for those wondering about PHP integration, we use PHP-FPM. I'm generally pretty conservative and slow to change our architecture, but looking back, we made the right choice.

Re:For a good reason: nginx really performs!

[Zedrick]

> we began to encounter serious memory/CPU issues with Apache

Can you elaborate, please? I used to work at a webhost that used apache for around 500.000 websites, and memory/CPU was never a problem. (Not for apache, only for PHP and MySQL.) I often see people claim that Apache is bloated, but don't understand in what way (except possibly for the config files, that might might extensive but not really bloated and they don't affect performance)

Re:For a good reason: nginx really performs!

[FreakyGreenLeaky]

Exactly. Similarly, I also have to laugh at folks who blame sendmail for x, y or z problem they are experiencing with their mail.

Systems like Apache, Sendmail, GNU C, glibc, et al, are rock solid and reliable for a reason. Maturity.

Re:For a good reason: nginx really performs!

[Barsteward]

He/she sounds just like a troll either for nginx or just to upset apache users. Never trust a post from the " Anonymous Coward"

Re:For a good reason: nginx really performs!

[Terrasque]

I'm not OP, but have a bit of experience with this..

There are two things "against" apache:

1. It runs either a thread or a process for each connection
2. With all the fun stuff on, each thread / process take a lot of memory (especially with php enabled)

The new breed of popular webservers (which nginx seems to be the most popular of) are event-driven, which means it can handle a silly amount of connections in one thread/process.

This gives them advantages in a lot of situations. Especially for serving static files, doing comet stuff, and load balancing incoming connections to several backends.

It's also worth noticing that the part where apache is good, generating dynamic pages, are increasingly being moved to specialized software for the language (ruby on rails, django, php-fpm and so on), which the webserver then talks to via fastcgi or similar protocol (or even normal http).

Other event-based servers (from memory) : Cherokee, node.js, Tornado, lighttpd

IIS still wins

[recoiledsnake]

IIS still wins, based on the same metric that the iPhone wins over Android(or so the Apple afficionados keep claiming) because it corners the most profit in the smartphone market. IIS makes Microsoft billions every quarter, whereas Apache and nginx barely make their makers anything.

Re:IIS still wins

[Anthony Mouse]

You do understand that the corollary to that is the open source competitors save their users that same amount of money.

Re:IIS still wins

[Overly Critical Guy]

The metric "Apple aficionados" use is the one where the iPhone is the top-selling handset. For some reason, you're comparing a phone to an operating system. If you actually compare mobile operating systems, iOS has more share due to iPads and iPods.

Re:IIS still wins

[Anonymous Coward]

sore loser

Re:IIS still wins

[f0rk]

In what way is financial profit of the company/organization of any interest to the consumer?
I hear a lot from non-OSS people that the financial profit of the company is in direct correlation to the support and maintenance motivation. In a fully proprietary environment, this might hold true to a larger extent. Less so in an OSS project, for several reasons.

To argue that I'll use IIS over Nginx cuz IIS is racking in billions in profits to Microsoft feels like the complete opposite to what you REALLY should be comparing; feature set, security, development transparency, community, and so on.

My choice is of software is, arguably biased, tilted towards the OSS way. I believe in the fundamental ideology in OSS where people is motivated by self interest, mastery, and contributions to the larger picture.

Re:IIS still wins

[findoutmoretoday]

<quote><p> save their users that same amount of money.</p></quote>
Save some lose some, there is more than the license price to a product.

Re:IIS still wins

[rev0lt]

Hum like Godaddy did with parked domains? Oh wait...

Re:IIS still wins

[jbolden]

You meant that sarcastically, but it isn't entirely unreasonable to drop free webservers and look at profits as a different way of accounting. It does tell us something that for a long time with so many good free ones that: IIS, Oracle Weblogic (BEA), IBM's , LiteSpeed, riverbed ... are still sold. And the reason interestingly enough is generally the same. Integration.

In both the case of iPhone over Android and commercial over free webservers it appears that customers consider integration / ease of integration to be a service worth paying substantial money for.

Re:IIS still wins

[dingen]

But when you look at the top 20, top 100, top whatever you find the same pattern: Apache is by far the most used web server, IIS doesn't even come close.

Re:IIS still wins

[funfail]

You do understand that the corollary to that is the open source competitors save their users that same amount of money.

Then Android saves its users that same amount of money. So what?

Re:IIS still wins

[marcosdumay]

Yes, by not using IIS you save way more than the licesing.

Re:IIS still wins

[Overly Critical Guy]

It is a very unusual measure of marketshare to compare sale of individual phone models.

I agree, and that is what people are doing when they compare an iPhone, an individual phone model, to an operating system.

Also the notion that iPods should be included when looking at phone OS market share, is more than a bit strange, just to avoid talking about phone OS market share which has been the normal measure in the industry long before the iPhone and is a very relevant category to look at, even if Apple is not on top.

I specifically said mobile operating system, not smartphone operating system. As Android is attempting to expand outside of smartphones, and iPods remain a popular mobile device (they're simply iPhones without cellular radios), I see no reason why mobile operating system marketshare should be compared strictly on the basis of smartphones running them, anymore than you'd compare desktop operating system marketshare based strictly on laptops.

Re:IIS still wins

[marcosdumay]

Yes, and saving time for not having to deal with the Microsoft's support channel available... That kind of things.

We use it as a file server

[trawg]

A lot of people here are talking about how nginx is "only" useful as some sort of reverse proxy or cache engine or something. We haven't used it for that, although it's on our list of things to try at some point as a lot of people seem to have success with it.

We do use it for serving files over HTTP - primarily video gaming-related files, so they range in size from a few meg up to several gig. It generally performs flawlessly, although sometimes struggles under significant load.

We did a little formal testing ages back; I can't find the numbers now but generally we seem to squeeze more data out of the same hardware using nginx than we did with Apache.

(We used to use it streaming h264 video; it was great for that as well. We switched back to Apache to use some pseudostreaming mod but I can't remember why because now that I look I see the same module is also available for nginx as well.)

Nginx?

[Yuioup]

How come I've never heard of it till now? I don't live under a rock. I read tech sites daily and listen to poscasts. How did this slip under my radar?

Re:Nginx?

[HopefulIntern]

You tell me, it is pretty common these days.

Re:Nginx?

[migla]

You don't read or listen to quite the right stuff, apparently.

For a tech podcast, where you would have heard it mentioned in a few episodes, go to for example "techSNAP" (SNAP stands for Server, Network, Administration, Podcast).

http://www.jupiterbroadcasting.com/show/techsnap/

The host, Chris, can be a bit silly as if he was trying to fill too big shoes, pretending to be a professional TV host (which he technically is, as in getting paid and even having quit his day job, though), but some could certainly find that kinda nerdy cute. And he is still improving, as far as I can see. The "style" of his presenting is a tad bit too "american" to me. I still watch every show, though.

I'm sure they're way more knowledgeable than anything on the TWIT network with Leo Laporte (which is just too annoying for me to watch).

The co-host Alan is very knowledgeable about all sorts of tech and internet related things and especially server and network stuff. I believe he co-runs a serverhosting business. He prefers and runs nginx in his shop.

I don't like their commercials (and the very blurry line between selling and informing when it comes to talking about the products of their advertizers (such as godaddy(!))).

I also usually catch their "Linux Action Show".

Re:Nginx?

[migla]

... on second thought, the S in SNAP probably stands for security. I could just check their site in about a second an a half, but instead I'm preferring to be where I allready am and do what I was allready doing. The inetria in me sometimes astonishes even myself.

IIS will become legacy software

[RoLi]

Nonsense. IIS will become a legacy product

The share in Japan, Germany, Russia and many other countries already lies below 4% for many years. But also traditionally Microsoft-friendly countries can turn away from IIS, for example in the last 10 years, the share in France fell from 35% to 5%, in Brazil and Taiwan from over 45% to 15% and in India even from 65% to 18%.

IIS will probably be able to hold out another 10 years, but in the long term it's future is far from rosy.

Re:IIS will become legacy software

[RoLi]

IIS loses everywhere, also in the "top 1 million" - category on Netcraft.

I guess the one million websites with the most traffic are not "farting around", right?

Re:IIS will become legacy software

[Bert64]

Apache and nginx are both used for plenty of high traffic websites...

Not so long ago MS were also paying large domain registrars to host their parked sites (ie empty sites, but lots of them) on IIS to artificially inflate the netcraft stats.

Re:IIS will become legacy software

[Ice Tiger]

IIS is used for industrial strength websites for corporations, online shopping sites, etc... People use IIS for _real work_, .

That is correct, but a lot more corporations use non IIS solutions for industrial strength web sites.

Re:IIS will become legacy software

[LordThyGod]

Bollocks. Look at Netcraft's top 20. Only 2 run IIS, and those are MS owned domains. Everybody else is Linux + something else. The web is not so friendly to MS.

Re:IIS will become legacy software

[rvw]

Apache and nginx are both used for plenty of high traffic websites...

Not so long ago MS were also paying large domain registrars to host their parked sites (ie empty sites, but lots of them) on IIS to artificially inflate the netcraft stats.

And apparently MS has stopped paying them, so now they use Apache or maybe even Nginx and the inflation has backfired.

Re:IIS will become legacy software

[jbolden]

That article is looking at a type of webserving where the webserver using a substantial percentage of hardware. There are a huge number of applications in which the number of clients is vanishingly small but there is a strong desire to offer integration with other services, primarily intranets. In other words IIS is likely to thrive is where it always has, in the intranet where you want a non standard value add webserver solution.

As a percentage of websites though, intranets are small.

Re:IIS will become legacy software

[fuzzyfuzzyfungus]

Unless they fuck it up pretty horribly, IIS has an assured future as the embedded webserver for default installs of the various Microsoft products that have a web-facing component(ie. Foocorp installs exchange, exchange.foocorp.com/owa is going to be providing 'Outlook Web Access' via IIS... Ditto with Sharepoint and similar).

If Microsoft has plans for people to actively shell out a nontrivial amount of money just to run a commodity HTTP server, though, they'd better have something good in mind, especially now that so much commodity HTTP serving is done on an outsourced basis by 'cloud' types or CDNs, both of which generally care a great deal about per-node costs that scale linearly with volume(like software licenses, or low end admin flunkies); but care substantially less about fixed costs(like having a serious guru on staff to deal with the esoteric intricacies of whatever they are using).

Re:IIS will become legacy software

[AmiMoJo]

IIS is less focussed on high traffic web sites now. The main uses are for things linked to other MS services like Exchange and for serving vertical apps developed rapidly in ASP.NET. If you look at those markets it still does well because it is the only option.

Re:IIS will become legacy software

[thetoadwarrior]

IIS is only really used by idiots bodging together a shitty company intranet and they can't operate anything without a gui.

Re:IIS will become legacy software

[thetoadwarrior]

you say it does well and it's the only option. that's not a good thing.

Re:IIS will become legacy software

[aztracker1]

MS does have web server versions of Windows Server... around $500USD or less iirc... not a bad value. I actually like IIS myself, it's one of the few MS products I can say that about. Though the mid-late 90's with script exploits etc was a bear, I had always removed all unused extensions first. IIS today has a lot of functionality that is more of a pain to configure/automate in say Apache for instance. I would say that it isn't the best for some roles though.

I think VS development (.Net) is probably the single biggest driver for IIS usage.

Re:IIS will become legacy software

[aztracker1]

ClassicCars.com Photo Gallery is an example of a modern web application using .Net/IIS behind it. It's a self-plug (I wrote it) but the effect is nice, if you're using a current version of Chrome or Firefox it's nicer than IE.

Re:IIS will become legacy software

[fuzzyfuzzyfungus]

I don't have anything in particular against today's IIS, my point was just that commodification, outsourcing, and scale are arguably a danger to Microsoft(and anybody else who is trying to sell licenses for software that isn't leagues better than the OSS competition). If you are running a server, or a handful of servers, even fairly exorbitant license fees are small change compared to the value of being able to use whatever it is your available people have expertise in. However, as the number of nodes you are running increases, the cost per unit of having expensive experts keeps on falling, while the cost of licenses doesn't.

Microsoft's real threat is not so much that the customers currently using IIS will directly abandon them for apache or nginx; but that they will abandon running IIS internally for buying web services from somebody else, who is much less likely to be running IIS.

Public internet sites..

[RightSaidFred99]

Not all sites. IIS is used massively on the corporate interanet.

Re:Public internet sites..

[peawormsworth]

Given the choice to between IIS or Apache and the technical knowledge to use either... I have never seen anyone choose IIS. The only reason I have ever seen web services specifically hosted on IIS is because someone told them that this is how they had to do it. The only other reason to use IIS is because that is what you are familiar with or because IIS is a requirement for the application you want to serve from it... which are valid reasons. I never see anyone choose to waste money on a product that provides absolutely no benefit over the free one running on a free operating system. Sorry to state the obvious.

Re:Public internet sites..

[Dan Dankleton]

I'm a Linux guy, but if I needed to add an intranet server in a small shop I would probably do it on whatever existing machine was easiest. In a lot of cases this will mean the Windows file server which has been in place for years, and installing IIS would be the path of least resistance.

Re:Public internet sites..

[RightSaidFred99]

[Shrug]. That's because you don't don't know what you're talking about, that's all.

I have more than that technical knowledge and I choose IIS. I do so because I'm deploying enterprise applications including WCF web services, WF backend services, and MVC web applications and use IIS/AppFabric to manage them. It would be...silly to run such things on Apache.

It amuses me that Linux guys (of which I am one, though not the dweeby type) think they have a deep technical knowledge about computing and "know" why any Linux solution is better, but if you asked them any in-depth questions about why or about different scenarios they'd just stare at you blankly.

Re:Public internet sites..

[TemporalBeing]

I'm a Linux guy, but if I needed to add an intranet server in a small shop I would probably do it on whatever existing machine was easiest. In a lot of cases this will mean the Windows file server which has been in place for years, and installing IIS would be the path of least resistance.

Apache HTTPd is very easy to install and run on Windows systems too. So that's no excuse.

Re:Public internet sites..

[TemporalBeing]

[Shrug]. That's because you don't don't know what you're talking about, that's all.

I have more than that technical knowledge and I choose IIS. I do so because I'm deploying enterprise applications including WCF web services, WF backend services, and MVC web applications and use IIS/AppFabric to manage them. It would be...silly to run such things on Apache.

It amuses me that Linux guys (of which I am one, though not the dweeby type) think they have a deep technical knowledge about computing and "know" why any Linux solution is better, but if you asked them any in-depth questions about why or about different scenarios they'd just stare at you blankly.

So we all know now to avoid your websites....let the phishing begin...

Re:Public internet sites..

[RightSaidFred99]

Are you just throwing buzzwords around? What in god's name does any of this have to do with phishing?

Re:Public internet sites..

[TemporalBeing]

Are you just throwing buzzwords around? What in god's name does any of this have to do with phishing?

Well, aside from IIS being easily crackable, and all that...yes, I know with II6 and later they did a good job of cleaning things up and making it more secure than it was before. But that was also a pretty low bar to meet. So don't be surprised when your clients end up getting phished by their own websites.

Pronunciation

[dutchwhizzman]

So how do you pronounce Nagios? Most people say nah-gee-oz. But as far as I'm concerned, the proper pronunciation is is nag-yo-ass. Because, that's what it's for, to nag you when something is wrong.

Re:Reverse Proxy

[micheas]

Your comments about varnish and nginx don't match my experiences.

I have that the recent versions of varnish are faster on linux, with nginx being too close to call.

When varnish was just getting noticed you were probably right, but a lot of people with lots of motivation to get varnish working well on linux have added patches to varnish that have sort of reversed that in my experience, not that most people are going to bump heads with what varnish can handle.

The nice/evil thing about varnish's config file is that it is in C and compiled before being loaded. It is a fun system to play with/setup.

Doubt Microsoft cares

[rikkards]

They found a better way of getting IIS out there, essentially you need it for any of their server functionality especially as a Distribution Point in SMS/SCCM to use BITS.

Re:Reverse Proxy

[rev0lt]

Many big shared-host implementations actually use fcgi instead of mod_whatever. Also, while nginx's rewrite module doesn't seem as powerful as mod_rewrite, it's not that complicated to convert the rewrite rules between them. I use Nginx with Magento and our custom developed products (and internally the development is done with Apache), and all the issues were easily fixed.
I also never had any issues with php applications like the one you mention - current popular applications run in almost every php-supported environment, regardless of the webserver.
AFAIK, Nginx development started in FreeBSD, and not Linux. And you are assuming sysadmins aren't using FreeBSD, but Linux, which is, in many cases, a fallacy (our servers are all BSD). RH/CentOS users will be more inclined to use the extensively documented, readily available, supported by wizards and easy to configure Apache (and eventually throw more servers to the pool), than a handtuned Nginx installation or other "fringe" solution.

Varnish performance in FreeBSD might have something to do to the fact that was created by PHK, a FreeBSD guru responsible for the creation of many core system features, including the malloc library.

SPDY

[Lennie]

Firefox will soon, probably Firefox 11, have support for SPDY (Speedy HTTP) like Chrome, but server support is limited currently.

If nginx would support it, I'm sure it will replace or serve as a proxy in front of many, many more servers of performance oriented websites.

Re:Ummm, ok

[Lennie]

If you look a bit further than just their news 'front page'. You would see that nginx overtook Microsoft in the 'active sites' catagory (the more important catagory ?):

http://news.netcraft.com/archives/2012/01/03/january-2012-web-server-survey.html

What about Cherokee?

[simula67]

Any comments on Cherokee?
I was trying to set up a foss mirror in our college and we finally settled with lighttpd because it was smaller/easy to configure.

Enterprise backends

[HideyoshiJP]

This may be, but I suspect enterprise software will be using IIS as a backend for some time. I've seen a number of highly customized apache installs, but more often than not, software vendors are willing to ride the Microsoft train for canned solutions.

Re:U need a dose of reality

[thetoadwarrior]

Copy and paste that all you want. It doesn't change that IIS is a dying pile of pig slop.

Re:Money WASTED clearing up security breaches!

[Anthony Mouse]

I was wondering, how does Microsoft track your posts so that you can get paid when you post anonymously like that? Can't any of your fellow shills claim that post as their own and take your money? Or do they give you some kind of monitoring software so they can track who posts what or something like that?

WTF? Math skillz lacking?

[s.petry]

I don't get it, at all.

Taking into account servers for all sites covering all domains, Microsoft holds a healthy lead over Nginx, accounting for 14.46 percent of sites for a total of around 84.2 million. Nginx runs on 56.1 million of all sites, representing 9.63 percent. Apache dwarfs them both: 64.9 percent for a total of 378 million sites. Google rests in the fourth spot with a 7.9 percent market share, covering 14.4 million sites.

How is this web ranged as the 2nd highest used and IIS rated number 3?

64.3(apache) > 14.46(IIS) > 9.63(Ngingx)

Ok, maybe the author missed their own statistics. But the last paragraph:

Finally, among the world's 1 million busiest sites, Apache holds a market share of 64.4 percent (640,547 sites), down 0.36 percent since December; Microsoft's share is 14.99 percent (149,209 sites), down 0.01 percent; Nginx represents 8.49 percent (84,541 sites), up 0.28 percent month over month; and Google handles 2.4 percent (23,894 sites), an increase of 0.09 percent.

Once again, 64.4(Apache) > 14.99(IIS) > 8.49(Ngingx)

The top statement in the article, the 11.1% number and the 12.1% usage numbers do not match anything else in the article. They sure don't match any statistic I can find on web server use. Sounds like someone is spinning some yarn trying to make a hype sweater for something they bought some stock in.

You call yourself nerds.

[degeneratemonkey]

I think it's sad that so many people on SLASHDOT (of all places) have never heard of nginx .

This is me, shaking my head.

Seriously. Apache has been dead to me for two years now.

Re:Freebies = fine 4 "smallfries" (try BIG compani

[CBravo]

Why do you post this long message so many times in this thread and doing it anonymous as well?

Re:Bigger "woosh" (reality on Linux vs. MS)

[CBravo]

Why do you post this too long message so many times in this thread and doing it anonymous as well?

Re:U realize THIS is reality (not Linux 'phantasy'

[CBravo]

Why do you post this message so many times in this thread and doing it anonymous as well?

Re:Money WASTED clearing up security breaches!

[glodime]

This is interesting but a bit off topic in this thread. Also FUD doesn't mean what you think it means.

Re:Bigger "woosh" (reality on Linux vs. MS)

[darkpixel2k]

XEROX...Windows Server 2003
The U.S.' LARGEST STOCK EXCHANGE...Windows Server 2003
FUJIFILM GROUP...Windows Server 2003
HILTON HOTELS...Windows Server 2003
MEDITERRANEAN SHIPPING COMPANY...Windows Server 2003
SWISS INTERNATIONAL AIRLINES...Windows Server 2003
UNILEVER...Windows Server 2003
MOTOROLA...Windows Server 2003 & SQLServer 2005
NISSAN...Uses Windows Server 2003...and Exchange Server 2003
TOYOTA MOTOR SALES...using Windows Server 2003
DELL COMPUTER...Windows Server 2003
HSBC...Windows Server 2003
RAYOVAC Chose Windows Server 2003
CONTINENTAL AIRLINES...Windows Server 2008
7 ELEVEN STORES...Windows Server 2003
STATE OF ILLINOIS GOVERNMENT...Windows Server 2003
REGAL ENTERTAINMENT GROUP...Switching to Windows Server 2003 + Windows "Embedded"
CARNIVAL CRUISE LINES...Manages 1,000 shipboard &Windows Server 2003 & Server (every boat needs an anchor, right?
STARBUCKS...Windows Server 2003 Active Directory
RADIOSHACK...Windows Server 2003
TOMMY HILFIGER...Windows Server 2003
NcSOFT...Windows Server 2003
TDC...Windows Server 2003
GAMEWORKS...using Windows Server 2003M
SHOP DIRECT...intends to migrate no less than 350 servers from Windows Server 2003 to Windows Server 2008 R2

It gives me such a warm feeling to see that 'SHOP DIRECT' has finally saved enough to upgrade 350 of their servers. Give them another decade on their remaining Windows 2003 boxes and they might be able to save enough to finish the project...

And 'bravo' to the two other companies that could finally afford to upgrade to Windows 2008 by firing a bunch of IT staff to pay for licenses to upgrade to Windows 2008 (which is already 4 years out of date).

For everyone else who is stuck on Windows 2003 because of costs or the fact that 15 year old corporate applications won't run under anything but IE6 and Windows XP--well, you're going to have to pay to re-write the apps anyways, you might as well consider re-writing them as a web-based app that can run from a much cheaper Linux-based server. If you still feel the need to get screwed by licensing fees, you can always purchase support from Red Hat, Canonical, etc... Or just send me a dump truck filled with cash...

Re:Ur bs FAILS vs. facts

[Anthony Mouse]

R U twelve yrs old?

I had assumed you understood that nobody would actually care about your post because a list of vulnerabilities (or users of a piece of software) is totally meaningless without a comprehensive list of vulnerabilities in and users of the alternative software so as to allow a basis for comparison, which you so blatantly failed to provide.

Re:Smallfries use Open "SORES" (big ones buy MS)

[Bert64]

Ridiculous, sure plenty of companies use windows and notice how these are pretty much all non-tech companies, companies where tech is not a core competency is often outsourced and purchasing decisions are typically made by people without a tech background and are often poor choices...

When it comes to applications on a large scale exposed to the hostile internet open source is king, think google, yahoo, facebook, twitter, akamai, amazon, and let not forget the worlds top supercomputers...
The fact is when it comes to organisations who understand technology open source is the more common choice...
Windows gets used in organisations where the decisions are made by people with little or no technical knowledge, who are much easier to fool with fancy marketing.

You talk about migrations, but don't mention what was migrated FROM... Many of these stories of saving money or time with a particular windows install are based on migrating from earlier versions of windows which were even worse... It's easy to make improvements when you start with the worst technology on the market.

Amusing you mention nasdaq, it's well known their core trading system is linux based.. want a citation? http://www.computerworlduk.com/news/open-source/3268210/nasdaq-in-open-source-tech-battle-with-7bn-nyse-bid/ or try google.
Also amusing you mention the london stock exchange, they did have a windows based system and it was often cited in propaganda pieces like yours, infact ms used to advertise it on slashdot. That advertising abruptly stopped about the time the london stock exchange suffered the first of several major outages. Their windows based system was derided by those who used it, it was slow and unreliable and got ripped out.

Have you ever worked at any of these companies? I have first hand experience with several on your list, and i would not hold them up as any kind of positive example... And let's just say that in any security audit, windows is pretty much ALWAYS the weakest link.

Given up on the CIS benchmarks yet? Or have you given up pushing flawed benchmarking tools without ever making any effort to actually understand how they work?

Do you even read the links you post?

London Stock Exchange serving malware:

http://slashdot.org/submission/1484548/London-Stock-Exchange-Web-Site-Serving-Malware [slashdot.org]

Says: The site has NOT been hacked, and there is a big difference. The alert is likely a result of "Malvertising",, a growing method for attackers to distribute malware via advertising tags"

Hacked CAs? You forgot diginotar, who were compromised due to keeping important servers on an active directory domain making them easy pickings... Also the link you posted has no useful information on it, all the reports i've read have stated the compromised systems at multiple CAs were windows, or compromised as a result of compromised windows boxes (ie keylogged admin workstations).

Sites built on Linux, Apache, MySQL and PHP are the favoured targets of phishing attackers"

Do you even understand what phishing is? Here's a hint, it has nothing to do with the platform the site runs on.

Various linux related sites hacked, but have you bothered reading *how* they got hacked? Most of them appear to be due to admins of those systems being compromised, and the hackers then using their access to compromise the servers... It doesn't matter how secure a system is if you compromise the people who maintain it, now the question is what exactly was compromised initially and how was it done?

Now here's a real example from a security audit i did a few months ago...
Client had RedHat, Solaris, various cisco devices (routers, firewalls, switches) and an active directory domain with all the wor

Alexa Top 1 Million Survey

[hackertarget]

Interestingly in February 2011 I did a survey of the Alexa Top 1 Million sites. The idea here was as opposed to the large sample that Netcraft uses I focused on the more established web sites.

Alexa Top 1 Million (February 2011)
IIS :: 17.9%
Nginx :: 7.9%
Apache :: 69%

Web Server Survey Summary February 2011
Web Technology Report February 2011

A new report will be prepared in the next few weeks, it will be good to see if Nginx has indeed gathered a larger chunk of the established sites.