Link to home
Start Free TrialLog in
Avatar of c-h-r-i-s-t-o-p-h
c-h-r-i-s-t-o-p-hFlag for United States of America

asked on

VOIP phone QoS on Cisco infrastructure

We are a 100% Cisco shop. Curently, we have no qos running on our infrastructure.

We will be deploying phones in the near future, what is the easiest / best way to implement QoS for the phones that we will be deploying?

How is QoS configured / deployed on a Cisco infrastructure?

What considerations should be made?
Avatar of c-h-r-i-s-t-o-p-h
c-h-r-i-s-t-o-p-h
Flag of United States of America image

ASKER

Also, I plan on splitting the phone traffic into seperate VLANs for QoS.
Avatar of jjmartineziii
autoqos is your best bet. its simple and almost requires no work.

on your trunk links use the command "auto qos voip trust" and for access ports use "auto qos voip trust cisco-phone"

Here are some videos:

Auto Qos on routers:
http://www.youtube.com/watch?v=mRqON3_ffJU

Auto QoS on switches:
http://www.youtube.com/watch?v=_p9X7QJ-HsY
My mistake. Our router / switch infrastructure is 100% Cisco.

However, we are going with the Avaya VOIP solution.

Will this affect the config?
Yes, im not familiar with the way avaya works with qos.

The trunk link commands are still the same but the access ports won't be dynamic. You would have to extend trust to those ports that have phones. With Cisco phones, trust is handed to the phone when it is detected through CDP.
ASKER CERTIFIED SOLUTION
Avatar of AnthonyHamon
AnthonyHamon

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Here is an example of policy maps on your routers, just to augment Anthony's comment

policy-map MPLS_QOS
 class MATCH_VOICE
    priority percent 40
  set ip precedence 7
 class MATCH_VIDEO
    priority percent 30
  set ip precedence 6
 class class-default
    fair-queue
     random-detect
I have to disagree with atrevido.

It's not a good idea to mark any traffic higher than 5. 6 and 7 are used for network and internetwork control.
This is good information. I have been involved in Cisco VOIP deployments in the past, but this will be my first run with Avaya.

I am familiar with the access vlan & vlan voice config.

Avaya uses DHCP to tell the phones which vlan is for voip traffic.

We have Cisco routers at the heart / along the backbone of our network.

We also plan to use several voice vlans. Lately, I have been wondering what sort of configuration would be required at the core of the network in order to provide qos to multiple voice vlans.


@AnthonyHamon:

I am assuming that the config that you're describing can be applied in a blanket config on all access ports. And then the switch will apply this config only to the voice vlans in question. Is this accurate?
Avatar of AnthonyHamon
AnthonyHamon

Yes, the configuration can be applied to all access ports (using the interface range command).  If, however, you have servers connected to some of the access ports, I recommend that you do not apply a voice VLAN to those: this is a security best practice.
I am not sure what you mean by 'And then the switch will apply this config only to the voice vlans in question' - please could you clarify?

At the core of your network (routers rather than multilayer switches in your case, if I understand correctly?), you need to configure class maps (to classify traffic) and policy maps to specify what proportion of the bandwidth each type of classified traffic will receive.  If you do not implement QoS at the core, QoS markings can be lost for traffic passing through the core, and, therefore, voice traffic will be delivered 'best effort' for the final leg of its journey.
@AnthonyHamon
Meaning that I can apply this config at a global level. And after that, the switch will ignore the config for standard data traffic (vlans). However, the switch will activate the config for standard voice traffic (vlans).
Yes, the global QoS configuration allocates frames to queues based on their CoS (class of service) marking.  So, a data frame will gererally have a CoS value of 0 (best effort), but your voice media frames will have a CoS value of 5 (and signalling frames will have a CoS value of 3).  These will be given preferential service in accordance with the global QoS configuration.
I understand how the qos is set at the edge, and what the function is.

I guess that i'm not understanding how the qos travels across the network or is passed from edge to core?

what would an example qos config be for edge, trunk, and core?