httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Dessent <>
Subject Re: [users@httpd] *.shtml or *.html
Date Thu, 12 Feb 2004 15:29:02 GMT
Ben Crane wrote:

> In terms of overhead, is it really a threat to speed
> as websites get bigger (I mean this one will never
> grow in excess of a 100 pages) or is that more of an
> urban legend...I've tried the SSI on my pages, in
> fact, my demo loads a fairly complex webpage into a
> webpage and it seems to not represent too much of an
> overhead.

The number of pages on the site isn't really relevant.  What matters is
the rate of hits that Apache must service that require parsing.  If
you're getting 50 hits per second that require parsing then it will take
its toll in terms of CPU.  However, if you're only averaging one page
every 3 seconds then it will be hardly noticeable unless you're running
this off an old 486 with 8MB ram or something.

The more significant thing to realize is that if every page is generated
with SSI then you're now running a 100% dynamic site, just as if every
page was generated through PHP or Perl.  That means that no pages will
be cached, except for images of course.  Normally, all of the static
elements of a page are cached by the browser so that the next time that
page is loaded, the web server can just respond with a "304 Not
Modified" without retransmitting anything -- assuming that the page on
disk hasn't changed.  But with SSI and every other form of dynamic page
generation, the server has no idea ahead of time whether the page has
been modified or not and so it cannot respond with a 304, it has to
serve every request in full.  This will cause a lot more outgoing
bandwidth if you have lots of hits for large HTML files that are being
generated with SSI.


The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message