RE: [technical] Transparent proxying with delay pools based on IP precedence bit

Hi Henrik, Dennis & all.

Thanks for getting back to me.

>Please note that due to the way delay pools operates, they are quite
>likely more efficient than any packet based bandwidth limiting. Delay
>pools operates at the application level, with no TCP overhead due to
>extra retransmissions or whatever.

Yes true ! I believe that Squid really works well with delay pools but how
do we discriminate and limit our clients international bandwidth separately
from their Local bandwidth, which is sold to them at different prices, and
at the same time allow them to reap the benefits of using a transparent
Squid cache on our network ?

-----Original Message-----
From: technical@pmburg.co.za [mailto:technical@pmburg.co.za]On Behalf Of
Henrik Nordstrom
Sent: 13 February 2001 01:21
To: technical@pmburg.co.za
Cc: Dennis; FreeBSD Mailing List; squid-users@ircache.net;
technical@sai.co.za
Subject: [technical] Transparent proxying with delay pools based on IP
precedence bit

David Wilson wrote:

> >we could do that in our etbwmgr product for freebsd.
> That would be great ! Especially if it could somehow interface with Squid
so
> that any Cache hits would not be bandwidth limited, but TCP_MISS's would
be
> pulled down at the bandwidth specified for that particular client, 16K if
it
> was and international site or 64K if it was a local site... all based on
> precedence bit.

Please note that due to the way delay pools operates, they are quite
likely more efficient than any packet based bandwidth limiting. Delay
pools operates at the application level, with no TCP overhead due to
extra retransmissions or whatever.

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