httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <>
Subject Re: proxying
Date Wed, 24 Jan 1996 03:39:42 GMT
On Tue, 23 Jan 1996, Paul Richards wrote:
> Hmm, those are reasonable heuristics to use in the absense of a proper
> protocol. It doesn't change the fact that http is missing this functionality
> in a properly formulated manner.

I don't understand why Expires: doesn't give you what you need.  The fact 
that it isn't being widely used, or correctly implemented, doesn't mean 
it doesn't exist.  

> If an enforced timeout was available you could do things like, set a long
> timeout on static pages, then when you want to actually change them you can
> first, drop the timeout to something quite small, after doing this wait for
> a time equal to the previous long timeout. 

Yes, you can do this with Expires - the real work required is in the user 
interface to that.

> It really concerns me 
> that there's no way to prevent incorrect (i.e. stale) data from being 
> served from caches. 

Not true.  If you are paranoid about stale copies, use Proxy: no-cache.  
If you are less than paranoid and willing to trust proxies to adhere to 
the HTTP 1.0 specification, then setting Expires: headers apppropriately 
will get you what you want.  But since most servers don't have a user 
interface for setting those headers, proxies are usually only left with a 
Last-Modified header to deal with - the "recommended" action is to always 
send an If-Modified-Since request, but some proxies allow the behavior to 
be configurable.  The AOL caches, for example, always send IMS requests 
for HTML documents, but cache GIF images for up to 6 hours (unless they 
have a Proxy: no-cache).  

Another note - many sites who rely upon being able to maximize traffic 
reports are also discouraged from setting Expires: headers because then 
they can't count how many hits are served out of caches.  There will be 
discussion on proxy server reporting of locally-served non-IMS hits at 
the Demographics and Data Collection workshop at the W3C I'll be 
attending next week - obviously, if servers have at least a hope 
of being able to get those hit counts they'll be more willing to allow 
caches to serve files without IMS requests.  I mean, no one cares how 
many hits your DNS server gets.  :)

I would definitely like to see some interface to setting Expires: headers 
in Apache.  A heuristic like Alexei's is interesting, I'd want to cap it 
at 48 hours or something.  An interface to compare which of the cache 
sites you are serving to which ones are reporting back numbers would be 
interesting too - i.e., "I see the AOL caches are reporting hits 
in a timely manner - I'll be liberal with the expiration dates on the 
documents I give them" versus "I see Hensa isn't giving me any data, 
fuck'm, the get no Expires header".  Etc....

Still, let me know if I'm missing something.


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--  http://www.[hyperreal,organic].com/

View raw message