On Wed, Jul 27, 2005 at 10:18:29PM +0200, [EMAIL PROTECTED] wrote:
> Hello Jaap,
> 
> Thanks for explaining.
> It is somewhat late to include it in Bering-uClibc 2.3, but it would be a
> nice addition for 2.3.1 or so. Please keep us updated about your
> experience with it (usability and stability).
 
Well, I myself have been using it since January 2005 and without any
problems. The code is a relatively simple but useful modification of
the sfq code seems to have never made it into the standard kernel
(maybe because there's not really a maintainer?).

I will do some checks again, but in the past I performed some checks
and upload bandwidth gets nicely divided to all hosts asking for
it. Download is (as said) a bit trickier, but I think it might be
possible too if the bottleneck is laid at the internal interface and
esfq queue is short, so that fastest downloading hosts get packet
drops first. (This assumes standard configuration of router externally
connected to ISP and internally one interface to all users).

Jaap

Attachment: signature.asc
Description: Digital signature

Reply via email to