Scheduler policies

General Information

Scheduler policies determine the order in which queues are serviced. All ingress and egress queues operate within the context of a scheduler. Multiple queues share the same scheduler. Schedulers control the data transfer between the following queues and destinations:

• service ingress queues to switch fabric destinations

• service egress queues to access egress ports

• network ingress queues to switch fabric destinations

• network egress queues to network egress interfaces

The are two types of scheduler policies:

• single-tier, in which queues are scheduled based on the forwarding class of the queue and the operation state of the queue relative to the queue CIR and PIR

• hierarchical or multi-tier, which allow the creation of a hierarchy of schedules where queues or other schedulers are scheduled by superior schedulers

Scheduler policies are applied to access ingress and access egress interfaces.

Single tier schedulers

Single-tier scheduling is the default method of scheduling queues. Queues are scheduled with single-tier scheduler policies when no explicit hierarchical scheduler policy is defined or applied. In single-tier scheduling, queues are scheduled based on the forwarding class of the queue and the operational state of the queue relative to the queue CIR and PIR.

Hierarchical schedulers

Hierarchical scheduler policies are used for access ingress and access egress queues. Hierarchical scheduler policies allow you to create a hierarchy of schedulers where queues and other schedulers are scheduled by superior schedulers.

Virtual schedulers are created within the context of a hierarchical scheduler policy. A hierarchical scheduler policy defines the hierarchy and parameters for each scheduler. A scheduler is defined in the context of a tier. The tier level determines the scheduler's position in the hierarchy. Three tiers of virtual schedulers are supported, as shown in the following figure:

Figure 50-2: Hierarchical scheduler and queue association

Tier 1 schedulers are defined without a parent scheduler. A scheduler can enforce a maximum rate of operation for all child queues and associated schedulers.

You can create a tier 2 scheduler without a parent tier 1 scheduler or a tier 3 scheduler without a parent tier 2 scheduler.

During configuration of the hierarchical scheduler, you can assign numerical level and weight values to the tier 2 or tier 3 child scheduler, to specify a priority level in comparison with other child schedulers of the same parent scheduler. The level value determines relative importance when tier 2 or tier 3 schedulers are contending for bandwidth. The higher the number, the higher the priority level of the child scheduler bandwidth request.

Child schedulers with a level value lower than other child schedulers do not receive bandwidth until all child schedulers with a higher level have reached their maximum bandwidth allocation, or have no packets to pass.

When two child schedulers have the same level value, the weight value determines which scheduler first receives bandwidth.

Policy override

You can override some or all settings associated with an ingress or egress scheduler policy on an L2 or L3 access interface or subscriber profile. See To configure QoS policy overrides on an L2 or L3 access interface for more information.