From: Petr Lapukhov (petrsoft@gmail.com)
Date: Sun Jun 04 2006 - 06:03:07 ART
But then, speaking a little theoretical (putting aside all that L2 overhead
and stuff:)
if we have sum of shapers rate _equal_ to interface physical rate,
congestion will never occur at interface level. So there is no
difference (theoretically) if we have FIFO of CBWFQ at interface level -
they will never be engaged.
So maybe setting "bandwith" with "shape" still does matter only in case
of oversubscription...
Petr
This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:31 ART