On Tue, 29 May 2007 18:35:35 -0700, toby wrote:

> I've heard other sceptical reactions to this benchmark. So, instead of
> (sigh) fielding accusations of trollery and ignorance, let's turn this
> thread into something constructive. Let's make a new benchmark between
> yaws and a C web server (Apache, lighttpd, or whatever you suggest). I
> have an open mind. I have hardware to test on, am experienced in
> configuring Apache and have set up Erlang servers before. Any takers to
> define more realistic benchmark conditions?

Ok, one more try...

Why would we "make a new benchmark", are you under the impression that
there is a general lack of HTTP benchmarks? The canonical monster one is:

http://www.spec.org/web2005/

...although there are a lot of usable ones that are much smaller, but only
give you limited information of varying degrees of usefulness. A small
collection that I've used being:

http://del.icio.us/james.antill/benchmark+http

...note that "ab" is intentionally missing. Here is a list of webservers:

http://del.icio.us/james.antill/webserver

...you probably want to include, at least, 4 of the above that play to
different strengths. Feel free to include more, including Apache-httpd
if you want.

If done well you'll then have something to say that will be worth
paying attention to.

TIA. HAND.

--
James Antill -- james@and.org
C String APIs use too much memory? ustr: length, ref count, size and
read-only/fixed. Ave. 55% overhead over strdup(), for 0-20B strings
http://www.and.org/ustr/

[ Post a follow-up to this message ]