Re: MLS QOS

This is a good reference diagram for the process:

Step 3

*mls qos trust* [*cos* |*dscp* | *ip-precedence*]

Configure the port trust state.

By default, the port is not trusted. If no keyword is specified, the
default is*dscp*.

The keywords have these meanings:

*cos*Classifies an ingress packet by using the packet CoS value. For an
untagged packet, the port default CoS value is used. The default port CoS
value is 0.

*dscp*Classifies an ingress packet by using the packet DSCP value. For a
non-IP packet, the packet CoS value is used if the packet is tagged; for an
untagged packet, the default port CoS is used. Internally, *the switch maps
the CoS value to a DSCP value by using the CoS-to-DSCP map*.

*ip-precedence*Classifies an ingress packet by using the packet
IP-precedence value. For a non-IP packet, the packet CoS value is used if
the packet is tagged; for an untagged packet, the default port CoS is used.
Internally, the switch maps the CoS value to a DSCP value by using the
CoS-to-DSCP map.

On Sat, May 26, 2012 at 9:22 AM, Carlos G Mendioroz <tron_at_huapi.ba.ar>wrote:

> Marc,
> switch don't do "correlation" in any of the meanings of that word I know.
> If you trust COS, then the incoming COS value is used to derive the
> internal DSCP. If you trust DSCOm then incoming DSCP value is used instead.
>
> Is it hard for me to offer "evidence" of that in written form. What you
> need to see it is to generate frames with different markings and see what
> happen.
> Medianet QoS docs may help showing it:
>
> http://www.cisco.com/en/US/**docs/solutions/Enterprise/WAN_**
>
and_MAN/QoS_SRND_40/QoSCampus_**40.html#wp1098549<http://www.cisco.com/en/US/
docs/solutions/Enterprise/WAN_and_MAN/QoS_SRND_40/QoSCampus_40.html#wp1098549
>
>
> -Carlos
>
>
> marc edwards @ 26/05/2012 12:36 -0300 dixit:
>
>> /This is the whole point of the previous question! If you use "trust
>> DSCP" then the switch will indeed use DSCP and not COS./
>> /
>>
>> /
>> Carlos, if trusting DSCP, the switch will use it's mappings to correlate
>> the packet's DSCP value to the frame's COS value. Switches deal with
>> frames, hense the use of COS. Please show me evidence that supports
>> otherwise.
>>
>> Marc
>>
>> On Sat, May 26, 2012 at 8:25 AM, Carlos G Mendioroz <tron_at_huapi.ba.ar
>> <mailto:tron_at_huapi.ba.ar>> wrote:
>>
>> marc edwards @ 26/05/2012 11:55 -0300 dixit:
>>
>> 1) doesn't matter because the switch will use it's mappings to
>> take dscp
>> value and convert to associated cos more importantly the
>> mappings have
>> to be adjusted. either command is good but I would also add a
>> 'switchpot
>> voice vlan XX' putting the phone into another network and 'mls
>> qos trust
>> device cisco-phone'. These commands will put cisco phones on
>> another
>> vlan and use macros to ensure .1p is set up properly.
>>
>>
>> It does matter.
>> It is how you tell the switch to decide which marking is for good,
>> and ignore the other. cos-to-dscp is only used if you decide to
>> trust COS, otherwise the DSCP goes straight into internal DSCP.
>> dscp-to-cos is used on output to create the L2 marking if needed.
>>
>> But you also can use trust extend, which instructs the phone
>> (switch) to remark PC's packets (or should I say frames) with a
>> given COS. That is the remarking that I was referring to.
>>
>>
>> 2) The Cisco phones will most likely mark both but switch will
>> only use
>> CoS. The DSCP values are adjusted based on the internal switch
>> mappings.
>> This mapping is not ideal left to default values and is
>> typically fixed
>> manually. I don't know why cisco just didn't make it default.
>> ... This
>> command : 'mls qos map cos-dscp 0 8 16 24 32 46 48 56' is used
>> in global
>> config to map EF (IP precedence 5) to the dscp value 46.
>>
>>
>> What do you mean by "switch will only use COS" ?? This is the whole
>> point of the previous question! If you use "trust DSCP" then the
>> switch will indeed use DSCP and not COS.
>>
>> The mapping is older than DSCP, at least older than any DSCP usage
>> standarization, I would guess.
>>
>> -Carlos
>>
>>
>> You are taking the right steps for Catalyst QoS good luck!
>>
>> HTH
>>
>> Marc
>>
>> On Sat, May 26, 2012 at 5:04 AM, Carlos Mendioroz
>> <tron_at_huapi.ba.ar <mailto:tron_at_huapi.ba.ar>
>> <mailto:tron_at_huapi.ba.ar <mailto:tron_at_huapi.ba.ar>>> wrote:
>>
>> COS fpr sure.
>> Phones mark COS & DSCP but only remark COS, so if you want
>> tp extend the
>> trust border, you have tp trust COS.
>> -Carlos
>>
>> --
>> Carlos Mendioroz <tron_at_acm.org <mailto:tron_at_acm.org>
>> <mailto:tron_at_acm.org <mailto:tron_at_acm.org>>>
>>
>>
>>
>> -----Original message-----
>> From: Vincent Tay <vtay.75_at_gmail.com
>> <mailto:vtay.75_at_gmail.com> <mailto:vtay.75_at_gmail.com
>> <mailto:vtay.75_at_gmail.com>>>
>> To: "Ccielab_at_groupstudy.com <mailto:Ccielab_at_groupstudy.com**>
>> <mailto:Ccielab_at_groupstudy.com <mailto:Ccielab_at_groupstudy.com**
>> >__>"
>> <ccielab_at_groupstudy.com <mailto:ccielab_at_groupstudy.com**>
>> <mailto:ccielab_at_groupstudy.com <mailto:ccielab_at_groupstudy.com**
>> >__>>
>> Sent: Sat, 26 May 2012, 05:36:38 GMT+00:00
>> Subject: MLS QOS
>>
>> Hi all,
>>
>> I got the below questions to clarify.
>>
>> 1) When a cisco VoIP phone is connected to a switchport,
>> which is
>> better?
>> mls qos trust dscp or mls qos trust cos?
>>
>> 2) The cisco phone will mark cos or dscp value or both? If
>> mls qos trust
>> cos is use, must i map the cos 5 to dscp 46 manually?
>>
>> Regards
>> Vincent Tay
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>>
>> ______________________________**______________________________**
>> _______________
>>
>> Subscription information may be found at:
>>
http://www.groupstudy.com/__**list/CCIELab.html<http://www.groupstudy.com/__l
ist/CCIELab.html>
>>
>>
<http://www.groupstudy.com/**list/CCIELab.html<http://www.groupstudy.com/list
/CCIELab.html>
>> >
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>>
>> ______________________________**______________________________**
>> _______________
>>
>> Subscription information may be found at:
>>
http://www.groupstudy.com/__**list/CCIELab.html<http://www.groupstudy.com/__l
ist/CCIELab.html>
>>
<http://www.groupstudy.com/**list/CCIELab.html<http://www.groupstudy.com/list
/CCIELab.html>
>> >
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> --
>> Carlos G Mendioroz <tron_at_huapi.ba.ar <mailto:tron_at_huapi.ba.ar>> LW7
>> EQI Argentina
>>
>>
>>
> --
> Carlos G Mendioroz <tron_at_huapi.ba.ar> LW7 EQI Argentina

Blogs and organic groups at http://www.ccie.net
Received on Sat May 26 2012 - 09:55:27 ART