[c-nsp] QoS per ethernet subif
Frédéric Jutzet
frederic.jutzet at cablecom.ch
Sun Sep 30 11:07:44 EDT 2007
> Does priority traffic compete with traffic from the same subinterface
only
> (where the policy is applied), or does it compete with
> the whole traffic passing through all subinterfaces?
if you apply it on a subif, it will compete only on this subif
> Does it make sense to use priority on the default class or should i
use
> another class for matching all traffic?
> Does it make sense to use priority when there is only one class
defined for
> the subif's policy?
the priority will only guaranty you that you have 3000k for the pk that
match your criteria, and only when the queue is full (in that case the
virtual queue created by the shaper), which mean in that case only when
you reach the 4M you have define in the shaper. That mean as soon as you
reach 4M on your shaper, you will then garanty 3M for all
traffic...Based on the 1st answer, you see that make no sense.
maybe you can apply your police-map on the main intf instead of subif,
but based on different criteria,
router7200(config-cmap)#match ?
access-group Access group
any Any packets
class-map Class map
cos IEEE 802.1Q/ISL class of service/user priority
values <<<<<<<<<
destination-address Destination address <<<<<<<<<<<<<
discard-class Discard behavior identifier
dscp Match DSCP in IP(v4) and IPv6 packets
fr-de Match on Frame-relay DE bit
fr-dlci Match on fr-dlci
input-interface Select an input interface to match
<<<<<<<<<<<
ip IP specific values
mpls Multi Protocol Label Switching specific values
not Negate this match result
packet Layer 3 Packet length
precedence Match Precedence in IP(v4) and IPv6 packets
protocol Protocol
qos-group Qos-group
source-address Source address <<<<<<<<<<<<<<
But in my opinion, it's even better if you set a cos/dscp in the pk on
your egress router/switch, then it's easier to match which pk get
prioritised on your congested intf.
Rgds.
-Frédo
>>> On 27.09.2007 at 17:59, in message <46FBD346.30900 at forthnet.gr>,
Tassos
Chatzithomaoglou <achatz at forthnet.gr> wrote:
> Hi,
>
> After reading
>
http://www.cisco.com/en/US/tech/tk543/tk545/technologies_tech_note09186a00801
> 14326.shtml
> i'm trying the following config on a 7200-G1 (12.2(31)SB7) and i was
> wondering if there is any actual reason to use it.
>
> !---------------------------------
> policy-map TEST-POLICY
> class class-default
> shape average 4000000
> service-policy TEST-POLICY-CHILD
> !
> policy-map TEST-POLICY-CHILD
> class class-default
> priority 3000
> set cos 5
> !
> interface GigabitEthernet0/2.2
> encapsulation dot1Q 2
> ip address y.y.y.y
> !
> interface GigabitEthernet0/2.3
> encapsulation dot1Q 3
> ip address x.x.x.x
> service-policy output TEST-POLICY
> !---------------------------------
>
> Gi0/2 is connected to a 6500 and has ~10 subinterfaces like the
above.
> I want all traffic passing from specific subinterfaces to have
priority over
> traffic passing from the other subinterfaces.
>
> So...
> Does it make sense to use priority on the default class or should i
use
> another class for matching all traffic?
> Does it make sense to use priority when there is only one class
defined for
> the subif's policy?
> Does priority traffic compete with traffic from the same subinterface
only
> (where the policy is applied), or does it compete with
> the whole traffic passing through all subinterfaces?
>
More information about the cisco-nsp
mailing list