Slashdot Mirror
2003: Year of Apache
John Chamberlain writes "Netcraft's numbers for the new year are in. The trend graphs tell a story: 2003 was the Year of Apache. If Time magazine had a server-of-the-year award the cover would be featuring a feather. Since October 2002 market share has grown from 53% to 64%, a 20% gain while Microsoft IIS, its nearest competitor has shrunk from 36% to 24%, a 33% decline. The change in server totals was even more dramatic. Apache HTTP Server increased from about 20 million to 32 million (+60%) while all other competitors remained flat."
Our department is moving from IIS 4.0 to Apache 1.3.29 within the next few months. The server is up & running and I'm working on porting our site over. The reasoning for the switch?
:)
While MS requires patching & monitoring, so does Apache/Linux (although it's not as time-consuming IMO). I also haven't had up-time issues with IIS although I inherently believe Apache would beat IIS in that category.
The true reason is that Apache processes SSI from the outside, while IIS processes them from the inside. I can make more modular code using apache (i.e. a single template for the whole site that the index files link to, and that template links to "content" and "data" files in a given directory). It also seems to perform better, but that's because I was using Access on the IIS machine, and MySQL on the Apache machine. Also Apache/MySQL are cheaper (putting SCO aside).
The only other good reason was to learn something new/different to make myself more marketable.
People don't generally switch web servers just for the heck of it. Obviously, there must be something seriously wrong with IIS to make people switch - I wonder what that could be...
From excellent karma to terible karma with a single +5 funny post...
We need to a series -- a long series -- of Slashdot interviews with key Apache people. I mean, look at all the stuff they're into. And the list doesn't seem to have any vaporware or bogged-down projects -- which is damned remarkable in the Open Source community, where people tend to be big on ideas and short on followthrough. Let's get these people under the microscope and find out what they're doing right!
It would be nice to see how this would look for percentage of http traffic rather than percentage of domains. I'm not sure who would be favored, but it seems like a better metric.
...so many even tiny sites - home PCs, private tiny hosts and such, run Apache.
It's big. It's slow. (okay, it can stand a big load without much slowdown, but overall latency is high) It's a system hog. These computers are often older Pentiums, sometimes 486s, sometimes used as clients/terminals, sometimes serving several other tasks.
Why people so rarely use tiny HTTP servers like Boa, Mathopd, thttpd... especially, that those tiny thingies are extremely fast under light load, light on system resources, have most of features every "amateur webmaster" wants, and because of small code base, usually completely bug-free.
Field for "Evangelism"?
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
Did anyone notice that in July 2002, Apache took a hit in numbers, and Microsoft gained for a brief period of time? (Check the graph, you will notice a spike in Microsoft's numbers, a dip in Apache in July 2002.) Does anyone know what this corresponds to?
Netcraft gives very specific rules by which it measures webserver counts here: http://www.netcraft.com/Survey/mechanics.html
:)
Always helps to actually visit the site. Their methods will favor Apache somewhat, as IIS does not generally play very well in hosting environments with virtual domains for various reasons. Of course that in itself is an indicator of server quality
These statistics make us happy, but they're not the whole story.
When we bragg about these numbers, Microsoft respond with:
"Our webserver is used by more Forbes/Fortune 500 companies and is used by more secure websites. Apaches numbers are only high because a lot of amateurs use it".
What is our argument to that? (we don't have one. We just ignore it and continue patting ourselves on the back.)
If we are to progress, it's better to look at what going *wrong*, and try to improve that.
Expert in software patents or patent law? Contribute to the ESP wiki!
Even then, how do you count them? How many machines are running any given web site? My sites currently have 8 servers behind a pair of load balancers. But it appears to the outside world as if it's a single machine. Also, do you consider all servers equal? Should my personal site be given equal weight with my company's banking sites? I'd be interested to see a weighted graph so that sites with more traffic have a greater impact. But the problem with that is, how do you measure it?
As an aside, I'm getting mildly concerned about Apache's market share. Not because I don't like it -- I do, and run both personal and corporate sites with it. But I distrust software monocultures, and I fear Apache's heading that way. So I hope that Apache gets some viable competition. I also hope, however, that it comes from somewhere that isn't intent on displacing it with proprietary, incompatible servers. So that'd be something other than IIS, then...
"The invisible and the non-existent look very much alike." -- Delos B. McKown
We run an online testing and certification engine, written in perl. It WAS hosted on a Win2K/IIS box, but about once a week the server would lock up with IIS hitting 100% CPU utilisation and the only way to 'fix' it was to reboot. The same code's been running on a Redhat 9/Apache server for about 2 months now with no downtime.
Our MD was so impressed with the port (which was very trivial), that she's asked me to consider migrating our main in-house server to Linux too - it's mainly a 'file and print' box so this should be a piece of cake.
We WERE looking at a contact management system (possibly Maximizer or Goldmine), but now we're seriously considering an open source alteratives-should save us about 7000UKP in apps and licences.
AT&ROFLMAO
Most definitely. I think these numbers will finally silence those misguided idiots who continue to say that Apache doesn't get exploited as much because "it's less popular" or "it's not used by anyone" or "it's written by a bunch of unpaid amateurs". Apache gets used because it's clean, simple, reliable, robust, and most importantly, EASY TO CONFIGURE.
If 2003 was the Year of Apache, then 2004 will be the Year of the LAMP.
SCREW THE ADS! http://adblock.mozdev.org/ Proud user of teh Fox of Fire - Registered Linux User #289618
Too much stuff does not work well under Apache 2.0.x. Specifically mod_perl has some interesting gotcha's, HTML::Mason has some issues, there are some Apache::DBI issues, .... long list.
Basically when the server went to 2.0.x, the rest of the supporting community wasn't ready. Most of it is still in testing mode. The 1.3.x branch is "good enough", and it doesn't break stuff. 2.0 is good, but it breaks stuff.
Another way to look at it is that my company ships product based upon 1.3.x. Moving to 2.0.x would require several things which don't yet exist. As we are happily operating under 1.3.x, we have no reason to move. If the Apache folks decide to completely abandon 1.3.x, thats OK as we have source and can fix it as needed.
I suspect that most folks will stay with 1.3.x for the forseeable future. The 2.0.x migration will cause more headache than it is worth, and it will cost money/time.
Y'all are forgetting that a good layer 7 proxying firewall is also going to skew things.
With the combination of URLScan header removal and a Unix-based firewall (few folks are insane enough to put up IIS webservers and Windows Firewalls on the same network...) my IIS5/6 hosts don't look anything like a Windows box as far as Netcraft is concerned.
Throw in a hardware load balancer doing SSL offloading, and the client connections are never going to see my hosts directly for Netcraft to count.
While IIS is not the leader, it is interesting to note the trend on open-source webservers:
Apache from 22M to 31M (40%)
Jetty from 1150 to 3731 (324%)
Resin from 24224 to 57113 (235%)
vs. Closed source ones:
IIS from 9.7M to 9.6M (-0.1%)
Lotus-Domino from 78k to 86k (10%)
Oracle from 6629 to 8167 (23%)
Weblogic from 5344 to 7844 (46%)
It looks like
a. The big boys have a trend that is slower than the small ones
b. Open source grows a lot faster.
That says a lot about the dynamic of open source webservers in general, and probably all open source tools to some degree.
Write boring code, not shiny code!
It's nice to see that Apache is gaining ground. Now it if could only send out WMV and RAR files with the correct MIME type, that would be great!
What a fool believes, he sees, no wise man has the power to reason away.
If Microsoft are trying to sell it on the basis of "big professional companies use IIS", it doesn't really work for me. I'm interested in what sites like Amazon, Google, the BBC, Tesco, Natwest, BT, British Airways and the IMDB run on. Stuff that either gets a lot of traffic, has to be secure or both.
I don't have a list of the Forbes 500, but I've had a look at the FTSE 100 in the UK, and a great deal of those companies don't have what I'd call major websites - their businesses don't depend on high traffic. There are companies involved in Biotech, Construction, Mining and Food production.