Re: WWW Proxy information
altis@ibeam.jf.intel.com (Kevin Altis)
Errors-To: listmaster@www0.cern.ch
Date: Mon, 2 May 1994 04:40:27 +0200
Errors-To: listmaster@www0.cern.ch
Message-id: <m0pxnqh-0003V3C@ibeam.intel.com>
Errors-To: listmaster@www0.cern.ch
Reply-To: altis@ibeam.jf.intel.com
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: altis@ibeam.jf.intel.com (Kevin Altis)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: Re: WWW Proxy information
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Content-Type: text/plain; charset="us-ascii"
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Mime-Version: 1.0
At 6:36 PM 5/1/94 -0500, burchard@geom.umn.edu wrote:
>OK, so it sounds like all CGI output really does need to have Expires
>meta-information added in order to be fully compatible with proxying.
Well, every CGI script doesn't NEED to have Expires, since the default that
most proxies will use is to never cache items that don't have Expires or
Last-modified information (CGI scripts).
>What I understand you to be saying is that the proxy server's caching
>behavior is based only on the meta-information fields and on the URL
>-- not including POST query data! -- that generated the page. And
>that furthermore, lack of meta-information results in
>server-dependent behavior. Both of these are quite reasonable, but
>imply that every CGI script needs to preemptively expire its output.
Based on Last-modified and Expires. Your POST query information going to a
remote server is never cached. The results of a POST or simply a cache
possibility. Content only has effect based on size, since a cache can use
data size to determine whether an item should be cached or for how long.
For example, I don't cache things over 16MB (Internet Talk Radio), even
though I have a very large cache on our proxy. Keep in mind that how the
cern_httpd caches today is not the only way to cache either and that other
proxies to come, say that people on this list write might cache
differently.
>Presumably it would be OK to add a fixed date which is guaranteed to
>be in the past, like:
> Expires: Sunday, 01-Jan-00 00:00:00 GMT
>Is that right?
Hmm, I suppose that might make sense. That should override any proxy settings.
In general I wouldn't worry about adding Last-modified or Expires to your
CGI scripts. What I do encourage is adding Expires information that
supports caching if applicable. For example, if you have a database that is
updated and reindexed at 2:00 GMT every day, then when you send the results
of query, you know that the results are good until the next index at 2:00
GMT, so that can go into the Expires meta-field. This applies to many of
the CGI queries on the Web today.
ka