Re: [LARTC] SFQ buckets/extensions

["Michael T. Babcock" <mbabcock@fibrespeed.net>]

> I disagree that the goal is to make the subqueues the same length.
> The goal is to serve them with the same bandwidth (as long as they
> don't become empty.)

Not that you need backing-up, but I agree with you.  SFQ is there to provide
near-fair queuing on a per-session basis.  As modified, it could also be
used to provide near-fair queuing on a per-IP basis instead, but having
different-depthed sub-queues simply indicates why fairness is needed (one
sub-queue would otherwise have dominated the available bandwidth).

A very long sub-queue however, indicates that perhaps fairness is not being
acheived (although, that's why its refered to as being stochastic).  This is
why I suggested at least being able to tune the size of the hash / number of
hash buckets so as to redistribute the streams through more sub-queues.
Dealing with bursts properly at Linux timer resolutions is another issue
:-).
--
Michael T. Babcock
CTO, FibreSpeed Ltd.

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/