LLQ and Shaping

[Jed Renton]

I Notice in the Networkers presentation RST-2502 WAN and Branch QoS Design that when LLQ and FRTS are implemented that the priority queue bypasses the shaping process.

Has anyone seen a reference for this

Is this also true for class based shaping

thanks

[ebreniz]

Traffic shaping must be enabled in a Frame Relay environment for accurate admission control of resources (bandwidth and queues) at the congestion point; that is, the VC itself. Refer the following URL for more information

http://www.cisco.com/en/US/products/sw/iosswrel/ps1835/prod_release_note09186a0080086fb0.html#wp27575

[pkhatri]

By its very nature, traffic in the priority queue is not queued at all. All traffic classified as priority is immediately scheduled out to the line. If it happens that the priority bandwidth assigned to the priority traffic classes is less than the actual bandwidth required for that class, all excess traffic will be dropped. Therefore, there is no question at all of shaping this traffic.

Hope that helps - pls do rate the post if it does.

Paresh

[Jed Renton]

Paresh

I understand the priority queue part but if the interface is shaping the class-default how is this algoritm maintained if there is excessive priority traffic bypassing the shaping.

The extension of this appears to be that the shaped rate would need to be the carrier policed ingress rate less the priority queue bandwidth