Good to see some healthy debate on the topic. It really gripes me when benchmarks are used as propaganda

Personally, I think web-server choice depends on your needs and a simple, “lighty is faster” is just obviously not the case. Performance will depend on many things, config, usage scenario, system config, etc…

Personally I use Apache, mostly for the power, flexibility and maturity of the software but am just looking into lighttpd for it’s ‘fast’ static file serving

Although I found to get the same sort of config as apache with light it uses roughly 100M more ram.
But for the extra speed boost on small and large php apps, its worth it.

Again thanks for the wonderful coverage of this performance stats, i agree that as a web-server apache is good choice, but not as forward proxy or reverse proxy.

Do you have any performance data on Apache and lighty in proxy mode ( forward and reverse)?.

-thanks
Harish

I don’t have numbers nor graphs right now to “backup” what i’m about to say.

I admin a site that takes around 31.000 simultaneous hits at peak hours,

Using 4 long-time-tweaked apache web servers this site took those 4 (dedicated) server to their knees.

i tried my best not taking apache away, i swear, it took me 2 weeks (with work-days of 12hours, sometimes more) of tunning, balances, caches, proxies, php code and sql tuning; until i decided to test lighttpd.

Now Lighttpd is not only serving this site very well but is also serving 3 other sites on the same 4 servers and there is room for more.

don’t take me wrong, i actually think apache is a great software.
My 2 points are:

1- as you already said: Lighttpd outperforms Apache
2- changing the web server soft DO (can) solve performance problems (at least in my case, this is not an absolute statement applicable to everyone)

to let you be happy let’s say: maybe i just did not made the right conf for apache. I think i did