Say I have the following style configuration for a head office router connecting into an MPLS IP WAN:
class-map match-all SITE-A
match access-group name SITE-A
!
class-map match-all SITE-B
match access-group name SITE-B
!
class-map match-all SITE-C
match access-group name SITE-C
!
!
policy-map SHAPE
class SITE-A
shape average 5000000
service-policy SCHEDULE
class SITE-B
shape average 5000000
service-policy SCHEDULE
class SITE-C
shape average 5000000
service-policy SCHEDULE
!
interface Hssi1/0
bandwidth 10000
service-policy output SHAPE
The access-lists simply matches traffic destined to each site.
The child policy SCHEDULE is a standard CBWFQ policy (with priority queuing and min BW classes).
The SHAPE policy is meant to shape traffic down to 5Mbps for each site. However, the interface to which this policy is applied is clocked at 10Mbps.
The question is what happens from a QOS scheduling perspective if all 3 classes under the SHAPE policy attempt to transmit at an aggregate rate of over 10Mbps (the access speed of the interface)? What QOS policy controls the combined traffic entering the hardware buffers (TX-Ring) of the physical interface?
Daniel
