Re: [SQU] SQUID with a M$ Parent...?

Hi Max,
    I hope you don't obejct, but I've cc'd this back to the squid-users mailing list. It's pretty standard netettiquette to stay on
the list unless it's agreed in advance to move the discussion off list.

My point was not about a single CARP parent, it was about a single parent. CARP is a tool for deciding where your upstream requests
get sent. On a single parent configuration it has nothing to do :]. If you have two parents and want to share the load between them
in a certain ratio, using CARP can be useful. I.E. you can give one parent 1/3 of the requests, and the other 2/3.

note: CARP DOES NOT QUERY. This is important. CARP decides on the client(be it a browser or a downstream proxy) what upstream proxy
will serve the request. This is the opposite to ICP and Digests which provide the downstream proxy with information to see what
parents have hits.

I think CARP may be useful in your environment...

Rob

----- Original Message -----
From: "Max Clements (Swist)" <max.clements@vodacom.co.za>
To: "Robert Collins" <robert.collins@itdomain.com.au>
Sent: Tuesday, December 12, 2000 7:00 PM
Subject: RE: [SQU] SQUID with a M$ Parent...?

> > If you only have a single parent, CARP is of no benefit. The no-query
> > option Timothy is suggesting is very useful, as well as 'default'.
> I have two parents - one a squid cache - and the other a M$ cache - but I do
> take your point on CARP and that is is useless with one CARP parent...? My
> objective was to be able to query the cache of both the M$ machine and also
> the squid machine - but careful thought brings me to the conclusion that the
> no-query option and the default to the squid would be more useful - as my
> own cache cannot do direct fetches...
>
> Thanks
>
> Max
>

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html