And to support my claims about minimal improvement ( I'm asuming you'll
trust me on the risk of putting things in the kernel :)
http://www.fenrus.demon.nl/performance.html
Which shows that khttpd is only a 26% improvement in speed at low
concurrencies, and about 12% at high concurrencies over Zeus. ( Did I
mention I love Zeus :) )
Now 20 odd percent speed up is not worth the risk of putting a server in the
kernel if you ask me. ( But then if you can afford Zeus, you can afford the
extra power, and not the extra risk. Just FYI, Zues costs about a thousand
British Pounds, yes it's a lot, but well worth it )
Of course, it's 4.5 times ( 350% greater ) faster than Apache, and there it
might be worth looking at, especially if you serve a lot of static content,
and not much dynamic content. As someone on the Linux Mindcraft challenge
team said, "And we all know how heavily the world is moving to static
content".
But seriously, it seems like a nice idea, and if you could integrate Apache
and khttpd tighter, then it might be worth it, but I hear that Apache is
about to get a whole lot faster with 2, so it might not be worth it in the
long term.
I would guess khttpd could find a lot of work in phones and fridges
though...
Adam
--
SLUG - Sydney Linux Users Group Mailing List - http://www.slug.org.au
To unsubscribe send email to [EMAIL PROTECTED] with
unsubscribe in the text
