From: Petr Lapukhov (petrsoft@gmail.com)
Date: Sun Jun 04 2006 - 02:45:03 ART
Thanks, Pierre and Chris :)
Your posts make think a bit more on that thing :)
I was wrong: using bandwidth with shape DOES make sense.
Basically, it enables CBWFQ _after_ shaping. So we have
two-stage complex queueing scheme here.
Without putting any "bandwidth" weight in class we got sofware
FIFO queue per physical interface (just as with FRTS, btw!).
However, setting bandwidth turns sofware interface queue to
CBWFQ, and assigns "shaped" conversation's packets to a
class.
I just wonder what happens if you set "bandwidth" with one
class, and omit it with another :) That is, how is that "another"
class served with interface CBWFQ?
Petr
This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:31 ART