From: Joe Chang (changjoe@earthlink.net)
Date: Fri Nov 26 2004 - 02:08:50 GMT-3
My understanding from the Wendell Odom text is that frames selected for priority service by LLQ or IP RTP will subsequently be assigned to the interface high queue. The dual interface queue selection is not based on whether a frame has been fragmented.
> Show queueing interface helped, but there are a few problems:
>
> If using just fragmentation (300 bytes), all packet normal Dual Fifo Queue
>
> If using fragmentation and LLQ selecting all packets, the results are inconsistent; or maybe my interpretation :).
> I was expecting that all frames, even the fragments, went to the High Dual Fifo queue. But giving it a second though maybe this test is inconsistent, because we would not want packets matched by LLQ to be fragmented. Any Coments?
>
> This is the result without LLQ, just fragmentation:
>
> r3#sh queueing int ser 0
> Interface Serial0 queueing strategy: priority
>
> Output queue utilization (queue/count)
> high/1 medium/0 normal/511 low/0
> r3#
> r3#
> r3#sh queueing int ser 0
> Interface Serial0 queueing strategy: priority
>
> Output queue utilization (queue/count)
> high/1 medium/0 normal/665 low/0
> r3#
> r3#sh frame-relay fragment
> interface dlci frag-type frag-size in-frag out-frag dropped-fr
> ag
> Serial0.34 304 end-to-end 300 0 1421 0
> r3#
> r3#sh frame-relay fragment
> interface dlci frag-type frag-size in-frag out-frag dropped-fr
> ag
> Serial0.34 304 end-to-end 300 0 1660 0
> r3#
>
> Just is the result with LLQ (selecting all frames) and Fragmentation:
>
> r3#sh queueing int ser 0
> Interface Serial0 queueing strategy: priority
>
> Output queue utilization (queue/count)
> high/83 medium/0 normal/2900 low/0
> r3#
> r3#
> r3#sh queueing int ser 0
> Interface Serial0 queueing strategy: priority
>
> Output queue utilization (queue/count)
> high/85 medium/0 normal/3031 low/0
> r3#
> r3#sh queueing int ser 0
> Interface Serial0 queueing strategy: priority
>
> Output queue utilization (queue/count)
> high/90 medium/0 normal/3156 low/0
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Thu Dec 02 2004 - 06:57:50 GMT-3