Does Squid beat the rest?

Aloha all,

I'm under contract to plan a medium-sized ISP buildout, and my client
wants to use a proxy cache. I need some hard data on the results I can
expect from Squid.

I have read the back issues of the mailing list on the web, and I've read
the FAQ, etc. But what I need are real numbers from end-users, similar to
what's available at http://cache.cnrs.fr/stats/stat1997.08.31.html
(preferably in English).

The numbers I'm most interested in are:

typical cache hit rates
                        Bandwidth %:
                         Document %:

hit rate as a function of cache size:
(i.e. 40% at 2GB, 50% at 10GB, etc)

Number of users accessing your cache:

Number of operations/second the proxy is handling:

Bandwidth you are saving by using Squid:

=======================

These results would help tremendously. I'll summarize and repost as well,
so we can all share the information, and get some real-world baseline
results of using Squid.

--
Jason Forester
jasonf@foad.org
SA Consultant