Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

.<<<<QOS issues>>>>>

1.i define a class with CBWFQ and I am yring to mark the class with bandwidth command. If I use ''bandwidth percent 60 '' will this be 60 % of the total BW or 60% of the remaining 75% left after class-default took 25 %. What about ''bandwidth remaining percent'' too.

2. I want to classify VOIP traffic , whats the diffence between the two commands.

match protocol rtp audio

match protocol rtp payload-type 4 .

3. My single backhaul link connects me to an ISP,then the ISP connects to my three branch offices. I created a tunnel btw HQ and each branch so that all traffic from the branches goes to ISP cloud (through each tunnel)then through a single backhaul link to my HQ. I have classified and mark my traffic ,what I need do now is to apply the policy-map BRANCHES. I want the policy-map to affect the traffic going to the three branches. Can I apply it at the physical interface connecting the backhaul?

4.If I classify and mark some specifi traffic to use certain percentage of BW. Since a ping packet will default to the class-default class, can I see the effect of the CBWFQ from the smoothness of the ping reply packet?


Re: .<<<<QOS issues>>>>>

Hi there,

1. It would be 60% of the 75%. Bandwidth remaining essentially divides up the bandwidth that is left after the priority queue and the 25% default have taken their share of the bandwidth.

2. if you're using cisco voip, you could just match on the EF traffic coming through from the phones (and the CS3 for signalling). To answer your question however, 'match protocol rtp audio' matches up to 23 payload types, the table on the link following lists the types. 'match ip rtp payload-type-4' actually specifies the match to be G723 (see the table on the following link).

3. Yes, you should be able to apply that policy map to the physical interface connecting to your ISP. If your HQ router is a layer 3 switch, you might have some issues with applying bandwidth statements etc, but if it is a router you shouldn't have any issues with apply the policy to the interface.

4. Probably not in the default sense. If you choose to mark the ping packet either by the policy map if the ping is coming from within the HQ, or using the extended ping commands if you're pinging form a router (you can write a TOS byte in the extended ping commands of a IOS capable device). Other ways to check to see if your policy is working is simply to look at the hits your policy is getting. eg permutations of the show policy-map command.

Hope that helps you along!


Super Bronze

Re: .<<<<QOS issues>>>>>

#1 (different from Brad's understanding, my understanding) bandwidth percent is from interface's total. Bandwidth remaining percent, I understand, to be after implicit and explict reservations.

e.g. (assuming default max-reserve)

policy-map A

class LLQ

priority percent 5

class X

bandwidth percent 70

policy-map B

class LLQ

priority percent 5

class X

bandwidth remaining percent 100

Class X should have the same allocation in both policies.

#3 Answer depends whether you need to manage congestion at near side physical interface (egress), remote site ingress, or both.

#4 Not really. Locally, pings within one class might indicate congestion within that class, from other classes or both; and there's return congestion to consider too. Further, if you're running on any type of shared WAN bandwidth, pings could be impacted by congestion not directly related to your traffic.


Re: .<<<<QOS issues>>>>>


You are absolutely correct about the bandwidth percent command. It is the percentage of the interface's bandwidth value.

eg bandwidth percent 10 when the interface has bandwidth 256 configured would be 25k.

Justkennie, apologies for misleading you.