From: Pierre-Alex (paguanel@hotmail.com)
Date: Tue Jun 13 2006 - 11:19:55 ART
I am trying to picture myself the flow of a packet between the schedulers and
various queues in the case of
a sub-interface of dlci.
From what I have read and understood, the following happens (example with a
dlci) :
1) traffic is directed to the proper dlci
2) it is identified as belonging to a given class
3) the CBWQ scheduler pickup a packet depending on its class / bandwidth
reservation and put it the FIFO queue for the dlci
4) the same thing happens for dlci 102
5) The frame-relay scheduler looking at the speed of the dlcis , picks up the
next packet from on of the two FIFO shapping queues
and put it on the interface sofware queue (which itself can be dual FIFO when
fragmentation is enabled.)
6) The packet is put in the Tx ring
Diagram:
traffic dlci 101 | Classe1 |
| Class2 | CBWFQ |-----> FIFO
|
| Class3 | SCHEDULER| SHAPPING QUEUE1 |
|------>FRAME-RELAY-->INTERFACE
traffic dlci 102 || Classe1 |
| SCHEDULER SOFTWARE ----> Tx Ring
| Class2 | CBWFQ |-----> FIFO |
QUEUE
| Class3 | SCHEDULER| SHAPPING QUEUE2 |
Any flaw or misunderstanding ?
Thanks
Pierre-Alex
This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:32 ART