Re: [TCLUG:18516] Squid as an httpd-accelerator

[Timothy Wilson <wilson@visi.com>]

On Thu, 1 Jun 2000, Clay Fandre wrote:

> Squid acts just like your local Netscape cache, but in a global sense.
> (Squid is meant to be on your local network) When a page gets accessed
> via squid, it stores it in cache locally on the squid server. Anyone
> else accessing that site will get the cached (local) page, thus avoiding
> going out to the internet to get it. If you have a slow connection to
> the internet, it will help a great deal. I set it up here and have
> around 20 users using it. On average it gets about a 5-10% hit rate. The
> more users you have using it, the higher the hit rate. (usually)

Squid really performs two functions. The caching you described is probably
the most common. We use a Netware BorderManager firewall/proxy here at
school performs that function. We pull up to 70% out of the cache instead of
going out to the net.

I'm interested in the reverse. I think it's usually called reverse proxying.
Squid people apparently call it httpd acceleration. I want to set up Squid
so that it can sit in front of Zope and Apache to cache those Zope pages
that stay relatively static. Presumably you wouldn't need a huge disk cache
for that. Just load up the Web server with tons o' RAM and let Squid pull it
from there.

From the docs, it looks like you'd set up Squid on port 80 and apache on 81
(or some other port) and tell Squid to forward requests to port 81 for those
it doesn't have in cache. The docs in question are at
http://www.squid-cache.org/Doc/FAQ/FAQ-20.html

-Tim

--
Tim Wilson      | Visit Sibley online:         | Check out:
Henry Sibley HS | http://www.isd197.k12.mn.us/ | http://www.zope.org/
W. St. Paul, MN |                              | http://slashdot.org/
wilson@visi.com |   <dtml-var pithy_quote>     | http://linux.com/