I'm not familiar with RSVP, can I can help with WFQ.
WFQ is enabled by default on nearly all serial interfaces with speeds less
then 2Mbps. (Exceptions are if you are using SDLC, X.25, LAPB, compressed
PPP, and maybe a couple others).
WFQ is able to sense the various 'conversations' between end stations by
identifying packets based on a couple of things -
Source/Destination network address
Source/Destination MAC address
Source/Destination port number
Frame Relay DLCIs
If queueing is necessary on a link, then packets are put into the fair
queue. They are serviced from the queue in order based on the time of
arrival for the last bit of each packet.
So, even if a big FTP packet arrives into the router before a small E-Mail
packet it is likely that the E-mail packet will get transmitted first
because the E-mail packet finished arriving before the FTP packet.
Because they service the queue like this, smaller packets will have a
greater chance of being transmitted first which gives them a semblance of a
higher priority.
The router can also keep track of the 'conversations' and small packet
conversations can have their queued packets interleaved with the packets of
a large packet conversation like FTP. This keeps them from being held up in
queue too long if there is already a bunch of packets in queue for the large
packet conversation.
The only adjustable variable for WFQ is the congestive discard. If a
high-volume/large-packet conversation has packets in queue, it can only have
the number of packets up to the congestive-discard value. Once the number
of packets reaches that number the router no longer queues packets for that
conversation until the number of packets in queue drops to 1/4th the
congestive-discard value.
I don't know about how RSVP works, but I can explain some about how
Frame-Relay traffic shaping works in association with queues, and it may be
pretty similar.
With FRTS you can give each PVC a CIR & Burst speed. If the traffic for
that PVC goes over those amounts it needs to be queued. After the router
determines that it needs to be queued, it sends it to the queueing process
in effect on that PVC (WFQ for our example). The router then begins
queueing and de-queueing the packets as explained above.
I would guess if you reserved 50K of traffic on a link, and the link was
full, and you had more then 50K of traffic to send, the remainder of the
traffic (after the 50K) would be queued up, and then the queueing process
would determine in which order packets were serviced until the traffic load
dropped and the queue had been emptied.
Hope that helps some,
Mike
>
> I have basic questions about RSVP and WFQ
>1- How WFQ works ? I mean how are the number of queues
>are governed when you enable WFQ ?
>2- How is algorithm works ? i,e how are the packets
>classified and how are the low volume data are placed
>on the queue front ?
>3- How does RSVP interact with WFQ ? I mean if I
>reserved 50k on 128k interface how is the interface is
>going to handle the mixed traffic in terms of queues
>and weights of the queues.
>Your help will be highly appreciated.
>4- How is packet scheduling and classification work ?
>i,e How are the internal algorithm works
>Thanks
>Mohamed Shommo
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
___________________________________
UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html
FAQ, list archives, and subscription info: http://www.groupstudy.com
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]