Anytime you apply QoS to to an interface, the QoS scheduler needs to know exactly how much bandwidth is available before it can divide it up and guarantee it at Layer3.
With ethernet or serial interfaces, the router assumes they are point-to-point, and uses the default interface bandwidth to make these calculations.
In the case of frame-relay and ATM, you need to configure traffic-shaping to do 2 things:
1. Guarantee bandwidth at layer2
2. Provide the bandwidth information of the PVC to the QoS scheduler to divy up the bandwidth correctly.
#1 is especially important in ATM, in the case that you have multiple ABR PVCs off the same ATM link. Since they are all sharing the links bandwidth, if you try to guarantee 300kbps at layer3 for an application, there is no guarantee that once you get to layer2 that 300kbps is still going to be there. If you have a bunch of abr PVCs, and 1 vbr-nrt PVC, the vbr-nrt PVC will take precedence over the ABR according to IP to ATM CoS, and as long as you have enough bandwidth configured in vbr-nrt, the guarantee can be made at layer3.
Remember, also, if you're planning on implementing a CBWFQ policy with real-time traffic (ie. voice or video), you need to tweak your tx-ring-limit under your PVC in order to create the necessary back-pressure for the Class-based queues to kick in during congestion.
If you're on a pa-a3 port adapter, put your tx-ring-limit at 3, if you're using a lower platform with AIM-ATM, or NM-1A, put it at 5.
Thanx for the info.. I am deploying QoS for real time traffic ( Voice) over a large enterprise network.The customer has WAN network via 3 tier IGX/MGX architecture.. ATM Pvc's are the standard wan links used. I had selected from 75 location 2 locations to test the QOS over WAN. I had configured the vbr-nrt value using the IGX end values for specfic PVC.
llq was configured.Later network congestion happened on that segment during high traffic and vbr-nrt was removed to stabilize the tfraffic.. I monitored via MRTG and found the traffic was getting checoked at the router end.
1> Query one: What values should we use when confguring the vbr-nrt in the router pvc? I need some help here.
2> What is the recommended approach when implementing vbr-nrt ?
When configuring LLQ, best practice is to set PCR equal to SCR, with an MBS of 0 (in later IOS, 'vbr-nrt 3000 3000' can be put in, without explicitely configuring 0 mbs).
As for the value to set it to, best practice is to match whatever the link is provisioned for.
Configuring vbr-nrt enables traffic shaping on the PVC, essentially limiting the amount of traffic that can cross it. If your performance suffered after configuring it, that means the PVC's provisioned speed isn't enough for the traffic you're pushing across.
Configuring LLQ on the PVC, though, should have protected your voice, or whatever other traffic you guaranteed in your CBWFQ classes.
Were you seeing a high number of drops on the 'show atm pvc'? This is a sign there is too much traffic for the shaping mechanism, and you either need to increase your buffers (queue-limit under your CBWFQ class), or increase the speed of the PVC.
We are pleased to announce availability of Beta software for 16.6.3.
16.6.3 will be the second rebuild on the 16.6 release train targeted
towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are
looking for early feedback from customers befor...
Introduction Featured Speakers Luis Espejel is the Telecommunications
Manager of IENova, an Oil & Gas company. Currently he works with Cisco
IOS® and Cisco IOS XE platforms, and NX to some extent. He has also
worked as a Senior Engineer with the Routing P...
In this session you can learn more about Layer 3 multicast and the best
practices to identify possible threats and take security measures. It
provides an overview of basic multicast, the best security practices for
use of this technology, and recommendati...