G-WAN Web Server Claims Speed Records, Features

[droidhacker]

Server software, even if it is free, is a HUGE investment. When you spend hundreds of thousands of man-hours on developing web sites and services, the availability of that server software becomes absolutely critical.

1) As other software and as hardware changes over time, the server must be something that can be maintained in relation to what you've developed. If the source is open, you can take it on yourself for maintenance, even if those responsible for it go belly up. In other words, you have a GUARANTEED future availability.

2) If your needs grow to something slightly outside of the abilities of that server, you need to have the ability to implement your needs without having to do something radical, like replacing a 100k man hours of infrastructure development.


Apache, and other open source servers guarantee a certain level of FUTURE-PROOFEDNESS. As long as the source code for your current version is available, you can deal with it forever.

Closed source server software is absolutely out of your hands to control, and for this reason, is NOT a viable investment, even if the entry cost seems to be very low.
With closed source dependencies in your infrastructure, you could be very easily fucked in the future.

I will say this; anyone serving any more than their own person photo album with this software is an absolute and total RETARD.

[droidhacker]

From the g-wan site:
http://gwan.com/benchmark

I didn't go over this in particularly great detail, but something very interesting caught my eye;

Where they were comparing the performance of gwan+C against apache+PHP, glassfish+JSP, balmer+Cspaz....

Uh, yeah... gwan is faster running compiled C code than all the other three are at running interpreted scripts.

HOW ABOUT AN APACHE+C TEST TO MAKE THINGS FAIR???!?!?!?!

WTF??!?!?

I once had a project I had to implement through Apache, written as a PHP script, it was impossible to make it fast enough to be used. Build the EXACT SAME CODE into an apache MODULE (compiled C), and it was about 1000x faster.


I call BULLSHIT on every bit of "literature" put out by these hacks.

[bulletxt]

LOL I love these trolls. They should get hired from Microsoft!

[Danny3]

Hi guys,

I have more than 1 year experience with G-WAN. I discovered this server searching the web for web server benchmarks and I found this site:
http://nbonvin.wordpress.com/2011/03...server-to-use/ and I was intrigued by the speed of G-WAN. I downloaded it and runned some bechmarks:

On laptop - Core2Duo @ 2GHz (dualcore):
G-WAN ~ 36 000 RPS (requests per second)
Apache ~ 11 000 RPS
On desktop Core i5 @2.6GHz (quadcore):
G-WAN ~ 170 000 RPS (with weighttp 303 000 RPS)
Apache ~ 45 000 RPS (with weighttp 62 000 RPS)

All bechmarks requested a 100 kb html file (G-WAN can't use php).
Benchmark tools used were ab (ApacheBench - singlethreaded) and weighttp (developed by lighttpd team - multithreaded)

In my opinion it's extremely fast against Apache and it uses less CPU, but unfortunately it can't run PHP. I'm sorry I didn't benchmark it against Lighttpd and Nginx.
Pierre, the guy developing this webserver says that he started coding at the age of 11 in Assembly and the server is written in C and can scale very well with multiple CPU cores. Unfortunatelly Pierre closed the forum in November last year, a very bad decision in my opinion, and stopped releasing updates for G-WAN in Mach this year. I don't know why. Only way to contact him now is the contact form on his site.

Benchmarks made by Pierre:
http://forum.gwan.com/index.php?p=/d...-wan-rpscpuram

[balouba]

Yep GWAN are trolls. Lot of super flawed claims etc. And the no-source-code is of course because a part of it is ripped from nginx (scour gwan and other forums, you'll see)

that being said,an http benchmark at phoronix made with proper methodology etc would be extremely cool
we keep seeing crappy benchs for all the servers and most of them are terribly misleading.


every now and then someone discovers that apache+mod_php is way faster than nginx+fcgid+php due to mod_php not having any IPC.
likewise for apache+event+fcgid+php vs nginx+fcgid+php (and you can go on with lightttpd etc). nginx seems to be the fastest on static serving too.

finally.. if that would also including caching and non caching reverse http proxies (which are really the first in line when you're hosting a large site) would be double awesome

[uid313]

Monkey is 60kb, faster than gwan, and open source (gplv2). gwan beating apache and lighty is believable, for nginx it depends (nginx default config is absolute crap and doesn't scale).

See
http://monkey.io/benchmarks/x86_64_monkey_gwan


Disclaimer: I'm working on monkey this gsoc.

Interesting.
But does it support PHP, Python, Ruby, Mono/.NET, Java?

How does it compare against Apache, nginx, lighttpd?

A little while ago I was intrigued by GWAN's performance and decided to understand how it works. Using strace and other tools, I figured some things out and began working on an opensource alternative using just the general idea of its architecture (event-based, 1 thread per CPU, etc)...

That doesn't seem very novel these days.
Aren't there a lot of web servers out there with the same approach?

[Danny3]

Hi guys,

I have more than 1 year experience with G-WAN. I discovered this server searching the web for web server benchmarks and I found this site:
http://nbonvin.wordpress.com/2011/03...server-to-use/ and I was intrigued by the speed of G-WAN. I downloaded it and runned some bechmarks:

On laptop - Core2Duo @ 2GHz (dualcore):
G-WAN ~ 36 000 RPS (requests per second)
Apache ~ 11 000 RPS
On desktop Core i5 @2.6GHz (quadcore):
G-WAN ~ 170 000 RPS (with weighttp 303 000 RPS)
Apache ~ 45 000 RPS (with weighttp 62 000 RPS)

All bechmarks requested a 100 kb html file (G-WAN can't use php).
Benchmark tools used were ab (ApacheBench - singlethreaded) and weighttp (developed by lighttpd team - multithreaded)

In my opinion it's extremely fast against Apache and it uses less CPU, but unfortunately it can't run PHP. I'm sorry I didn't benchmark it against Lighttpd and Nginx.
Pierre, the guy developing this webserver says that he started coding at the age of 11 in Assembly and the server is written in C and can scale very well with multiple CPU cores. Unfortunatelly Pierre closed the forum in November last year, a very bad decision in my opinion, and stopped releasing updates for G-WAN in Mach this year. I don't know why. Only way to contact him now is the contact form on his site.

Benchmarks made by Pierre:
http://forum.gwan.com/index.php?p=/d...-wan-rpscpuram

[M1kkko]

a free but binary-only web server

http://en.wikipedia.org/wiki/Oxymoron

[waucka]

Man, that quote from G-WAN in the article is propagandalicious.

[allquixotic]

The downside of not being able to fix bugs that are important to your organization way overshadows the upside of higher performance. Unless they also claim their software is bug-free... hah... that'd be the day!