httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Richards <p.richa...@elsevier.co.uk>
Subject Re: Zeus (YAIATFWS)
Date Tue, 19 Mar 1996 16:52:35 GMT
In reply to Robert S. Thau who said
> 
> I looked at the Zeus pages... and was left with a comment and a question.
> The comment is that they should probably get their technical story straight
> regarding threading --- in the high-level overview sections, they brag about
> having an "efficient single-process model"; when getting into the nitty-
> gritty, however, they indicate that you can actually have multiple processes
> running, and recommend two per processor to best exploit multiprocessor
> machines.

You can have multiple-threads in a single process. I couldn't find the
second reference and I don't know if it's a multi-threaded server but their
claims may not be in conflict if that is the case.

> The question is what would happen to the benchmarks if they put this
> in their reference config files:
> 
>   <Directory />
>   AllowOverride None
>   </Directory>

Are you saying the benchmark relies on access checks at each level and that
they do that very well while Apache doesn't?

I remember plans to improve that in Apache, did they ever come to anything?
The good thing about a published benchmark suite is we can see where Apache
has problems and fix them, nothing like a challenge :-)

Elsevier seems to be only 4 or 5 hops from www.obsolete.com, probably a 
256K line through but the server-push animation is *VERY* impressive. Several
frames a second would be my guess. (See the Zeus sites list or the
features link). I don't think Apache can do that across our ethernet :-(

-- 
  Paul Richards. Originative Solutions Ltd.  (Netcraft Ltd. contractor)
  Elsevier Science TIS online journal project.
  Email: p.richards@elsevier.co.uk
  Phone: 0370 462071 (Mobile), +44 (0)1865 843155

Mime
View raw message