Foundry NetIron M2404C and M2404F Metro Access Switches
Configuring HQoS (Rev.03)
QoS/HQoS Implementation
© 2008 Foundry Networks, Inc.
Page 28 of 98
instance. In this example, there is no direct relation between the number of Queue Policies and the
number of WRED Profiles.
Please note that in the current release of the product, only the HQoS resources implemented on the
‘ES Processor’ are managed using Policies.
For management of the QoS resources implemented on the ‘Packet Processor’ please refer to the
”
Configuring QoS”
and “
Configuring ACLs
” chapters.
Service-related Policies
NetIron M2404x supports the following Service-related HQoS policies:
•
Service Ingress Policy
o
Applied per SAP.
o
Defines mapping of VPT / DSCP values to FC and Color.
o
Defines mapping of FC to unicast, multicast and broadcast service ingress queues.
o
Defines the parent L2 scheduler as well as WFQ and WRED profiles for each queue.
•
Service Egress Policy
o
Applied per SAP.
o
Defines mapping of FC to unicast, multicast and broadcast service egress queues.
o
Defines the parent L2 scheduler as well as WFQ and WRED profiles for each queue.
•
Ingress Scheduling Policy
o
Applied per SAP or customer-site (a group of SAPs).
o
Defines the configuration of service ingress L1 and L2 schedulers, including their WFQ
and shaping profiles.
•
Egress Scheduling Policy
o
Applied per SAP or customer-site (a group of SAPs).
o
Defines the configuration of service ingress L1 and L2 schedulers, including their WFQ
and shaping profiles.
These policies, when applied to SAPs or customer-sites, govern how queues and schedulers are
allocated. By managing configuration of these polices and applying them to SAPs and customer-
sites the user can control how these resources are allocated.
Both shared and dedicated policy instantiation is supported.
Sharing Resources between Multiple SAPs
The same service ingress (or egress) policy can be applied to multiple SAPs (from different or the
same service). In this case the defined queues are shared between all those SAPs. Packets from
different SAPs that are mapped to the same shared queue are en-queued on a first-come-first-
served basis.
This also means that traffic from all SAPs sharing the queues is counted when showing statistics
for a specific SAP.
The same is supported for ingress (or egress) scheduling policies. When a scheduling policy is
assigned to multiple SAPs/customer-sites the traffic from all associated SAPs is aggregated by
relevant shapers for L1 & L2 schedulers in the scheduling policy.