[thelist] Questions for web traffic gurus

[Glenn Hunt]

> I think they are referring to special caching hardware (like 
> http://www.cacheflow.com/), and software caching proxy 
> servers.  IIS may do some caching of its own, but it will 
> still log the request.
> 
> In the cases mentioned I think you have something like this...
> 
> browser sends request to site.com
> 
> a) the site has a hardware cache setup, and your request gets 
> routed to the cache, if the cache has a copy, it responds, if 
> not it forwards the request to the web server and the web 
> server responds.
> 
> b) you have a proxy server setup and the proxy server saves a 
> copy of the site for you, or goes to the site and gets the new copy.

+1

I ran into this with a site design where the logout page, although
processed by CF, had no URL variables, and therefore appeared to be a
"static" page. Problem was, when someone [a demo actually ... OUCH!]
tried to logout, they never actually hit the server, just the cache.
Because of this, they never got logged out.

There is some meta-tags that you can try, but I think most caches ignore
them. The best solution we found to not have pages cached is to add a
'?' to the end of the URL.

www.yourdomain.com/somepage.htm becomes www.yourdomain.com/somepage.htm?

The cache thinks it is a dynamically generated page (even though it
isn't) and doesn't cache it.

Glenn Hunt
ghunt at hds.ca