12-04-2010 12:01 PM - edited 03-04-2019 10:41 AM
(There is a small into which you might want to ignore)
<intro>This is the first of a series of topics i will post over the next week or so as im during
a QOS over MPLS VPNs project which will actually focus on the customer edge mostly
and the way traffic is treated.I have a few things that need answering because i am simply
not covered by the Cisco docs which sometimes seem to be very laconic,thanks for your time </intro>
I have configured a traffic policer as a part of the customer's output policy to (de)mark or drop traffic
destined for the provider edge and i am stuck over two things:
1.
I am using a percentage-based policer which,according to Cisco documentation,has got a
default bc burst value of 4 ms (Cisco QOS command reference). Not according to this though:
police:
cir 40 %
cir 102000 bps, bc 3187 bytes
conformed 0 packets, 0 bytes; actions:
set-dscp-transmit af21
exceeded 0 packets, 0 bytes; actions:
set-dscp-transmit af23
conformed 0 bps, exceed 0 bps
If you do the maths,the bc value is 250 ms, not 4.Am i missing something here? (i must be)
2.
The policer runs inside classes being a part of a policy map attached on a FR interface as output.
The interface runs Frame Relay Traffic Shaping and with fragmentation at 320 bytes inside the map-class,
where the policy map is applied.
When applied though,the policer automatically sets the default bc in bytes to 1500,which is the Ethernet MTU value.
police:
cir 10 %
cir 25500 bps, bc 1500 bytes
conformed 0 packets, 0 bytes; actions:
set-dscp-transmit af41
exceeded 0 packets, 0 bytes; actions:
drop
conformed 0 bps, exceed 0 bps
According to the 250 ms bc i explained above which i think it takes as default, the bc in bytes should be
796 (using the formula: bc in bytes * 8 / class BW = bc in ms).
But i get a message saying: "burst bc increased to 1500 bytes".I can only assume it does this because the ingress
interface is an Ethernet interface and the policer runs before FRTS? (again,according to Cisco diagrams).
This stops me from setting the bc value to lower than 1500 by lowering the value in ms, by the way you cannot specify
bc in bytes, only in ms.
Now, what's a logical solution?I have tried lowering the interface MTU ( a serial interface) but nothing.
It wouldn't make any sense to lower the MTU on the ingress Ethernet interface,would it?
FRTS config:
!
map-class frame-relay 2PE
frame-relay cir 256000
frame-relay bc 2560
frame-relay be 0
frame-relay mincir 256000
frame-relay fragment 320
service-policy output ce2pe
policy-map:
!
policy-map ce2pe
class voice
priority percent 33
police cir percent 33 bc 20 ms
conform-action transmit
exceed-action drop
class data
bandwidth percent 40
random-detect dscp-based
police cir percent 40
conform-action set-dscp-transmit af21
exceed-action set-dscp-transmit af23
class video
bandwidth percent 10
police cir percent 10
conform-action set-dscp-transmit af41
exceed-action drop
class control
bandwidth percent 5
class vocontrol
bandwidth percent 5
set ip dscp cs3
class class-default
bandwidth percent 5
set ip dscp default
interface:
!
interface Serial0/0
no ip address
encapsulation frame-relay
frame-relay traffic-shaping
max-reserved-bandwidth 100
!
interface Serial0/0.1 point-to-point
ip address 10.0.1.2 255.255.255.0
no cdp enable
frame-relay class 2PE
frame-relay interface-dlci 100
Thanks