Hardware-assisted hierarchical QoS

Hardware-assisted HQoS is an alternative to traditional HQoS which is described in Scheduler QoS policies.

This model improves reaction time and provides more accurate control of downstream bursts for aggregate rate enforcement. It is based on:

  • hardware aggregate shapers (hw-agg-shaper)

  • additional hardware scheduling priorities (6 scheduling classes)

  • fair bandwidth distribution algorithm

  • software-based algorithm for enforcing HQoS at the vport level

Hardware aggregate shapers

Hardware-assisted HQoS is based on hardware aggregate shapers, which provide enforcement of aggregate rates for a group of queues in the hardware. In the FP4-based and later chipset, hardware aggregate shapers are available at egress only.

Hardware aggregate shapers are controlled by four parameters:

rate
sets the aggregate rate of the hardware shaper
burst-limit
sets the aggregate burst for the specific hardware aggregate shaper
adaptation-rule
sets how the administrative rates are translated into hardware operational values
queue-set size
sets the number of queues allocated as a block to the hardware aggregate shaper

Traffic is scheduled from the queues in a strict priority order based on the hardware scheduling priority assigned to each queue. This can be controlled using the sched-class command. There are 6 scheduling-classes available.

Hardware aggregate shapers do not need software-based HQoS to modify the oper-PIR of the queues to enforce an aggregate rate. This results in a significantly faster reaction time compared to traditional HQoS based on virtual schedulers.

The following examples show the configuration of a hardware aggregate shaper on a subscriber profile. The configuration is similar to traditional QoS configuration.

MD-CLI

[ex:/configure subscriber-mgmt]
A:admin@node-2# info
    sub-profile "trialSub" {
        egress {
            qos {
                agg-rate {
                    rate 1500000
                    burst-limit auto
                }
            }
        }
    }
    sla-profile "trialSLA" {
        egress {
            qos {
                sap-egress {
                    policy-name "trial-egr"
                }
            }
        }
        ingress {
            qos {
                sap-ingress {
                    policy-name "trial-ingr"
                }
            }
        }

classic CLI

A:node-2>config>subscr-mgmt# info
----------------------------------------------
        sla-profile "trialSLA" create
            ingress
                qos 2
                exit
            exit
            egress
                qos 2
                exit
            exit
        exit
        sub-profile "trialSub" create
            egress
                agg-rate-limit 1500000
            exit
        exit
----------------------------------------------

Fair bandwidth-distribution algorithm

Hardware-assisted HQoS introduces a new way to implement queue-weight based bandwidth distribution.

The queue hardware-enforced FIR rate is used to provide weighted fair access between queues assigned to the same hardware scheduling priority (scheduling class) within a hardware aggregate shaper, by dynamically changing the queue's hardware scheduling priority as shown in Hardware FIR enforcement for dynamic fairness control.

Figure 1. Hardware FIR enforcement for dynamic fairness control

Every queue in a hardware aggregate shaper is described by the following parameters:

scheduling class
sets assignment to hardware priorities
aggregate shaper weight
sets the queue fair-share of bandwidth at the specific scheduling class level
burst limit
sets the individual queue burst limit at PIR
FIR burst limit
sets the individual queue burst limit at FIR

The aggregate shaper weight algorithm operates similarly to the virtual-scheduler HQoS algorithm, meaning that the offered load of the queues at the same scheduling class is weighted against the aggregate shaper weight to determine the FIR, which then controls hardware-based scheduling.

A FIR-based algorithm is more efficient and more responsive to dynamic changes in offered load. The PIR is not adjusted, which prevents aggregate-rate underruns in case of a sudden change in an incoming load for some queues.

When hardware aggregate shapers are enabled, CIR as a queue parameter has no impact on the BW allocation algorithm, as in traditional HQoS. The CIR is still used for calculation of the default CBS. The queue-frame-based-accounting setting is also ignored when hardware aggregate shaping is configured.

The following examples show a SAP egress policy configuration. Weight-based bandwidth distribution in this example is only between queue 1 and queue 2. These queues are both assigned to the same scheduling class (sched-class 3), while queue 3 has a higher priority scheduling-class (sched-class 5) and is served with strict priority.

MD-CLI

[ex:/configure qos]
A:admin@node-2# info
    sap-egress "trial-egr" {
        queue 1 {
            agg-shaper-weight 5
            burst-limit 500
            mbs 102400
            sched-class 3
        }
        queue 2 {
            agg-shaper-weight 15
            burst-limit 500
            mbs 102400
            sched-class 3
        }
        queue 3 {
            burst-limit 500
            mbs 102400
            sched-class 5
        }
        fc be {
            queue 1
        }
        fc af {
            queue 2
        }
        fc ef {
            queue 3
        }

classic CLI

A:node-2>config>qos>sap-egress# info
----------------------------------------------
            queue 1 create
                mbs 100 kilobytes
                burst-limit 500 bytes
                agg-shaper-weight 5
                sched-class 3
            exit
            queue 2 create
                mbs 100 kilobytes
                burst-limit 500 bytes
                agg-shaper-weight 15
                sched-class 3
            exit
            queue 3 create
                mbs 100 kilobytes
                burst-limit 500 bytes
                sched-class 5
            exit
            fc af create
                queue 2
            exit
            fc be create
                queue 1
            exit
            fc ef create
                queue 3
            exit
----------------------------------------------

Hardware aggregate shaper scheduler policy

A hardware aggregate shaper scheduler policy is an alternative to a port scheduler. A hardware aggregate shaper policy can be assigned at the Vport level or at the port level for subscribers or SAPs.

A virtual Vport scheduler is added to implement a bandwidth-distribution software algorithm to handle traffic oversubscription by subscriber or SAP hardware aggregate shapers at a Vport.

The Vport represents a downstream bandwidth constraint and, therefore, the algorithm enforces a maximum rate, which is an Ethernet on-the-wire rate. This algorithm also implements the logic to distribute the available bandwidth at the Vport to its child hardware aggregate shapers.

Similarly, a port scheduler enforces bandwidth distribution between SAPs attached to the specific hardware aggregate shaper policy assigned to the specific port. For the bandwidth-distribution algorithm to operate effectively, all SAPs attached to it must have an explicit aggregate rate defined.

The algorithm performs a similar function to the existing H-QoS port scheduler, however, instead of modifying the oper-PIRs of queues, policers, or schedulers, it modifies the oper-PIR of the hardware aggregate shapers connected to it, and consequently triggers recalculation of the individual queue’s FIR, if required.

Vport or port bandwidth distribution shows the functionality of a hardware aggregate shaper scheduler, including all parameters which the user can set. This architecture allows the user full flexibility to define SLA models.

Figure 2. Vport or port bandwidth distribution

The generic design in Vport or port bandwidth distribution can be mapped into two SLA models as shown in SLA models using subscriber management as an example. Similar examples can be constructed using SAPs

Both SLA models show two subscriber categories for simplicity, but this model can be extended to any number of subscriber categories.

Figure 3. SLA models

The SLA model (a) in SLA models shows the overall Vport bandwidth distribution between gold and white categories in a ratio of 8:1 regardless of the number of subscribers in either category, and assuming that traffic passing through queue 7 for both subscriber categories is delay-critical but not significant in terms of volume.

The SLA model (b) in SLA models guarantees that the ratio between any gold and any white subscriber is 8:1. The overall bandwidth distribution between gold and white categories, however, it depends on the number of subscribers in each category.

SLA models are a design goal and must be tested to guarantee the expected behavior. SR OS configuration allows a user to construct other models by combining concepts and parameters in other ways. The resulting behavior is subject to a specific configuration and offered load.

When alternative SLA models are configured, the following considerations apply:

  • If a group is used to enforce weighting between different scheduling classes, all queues attached to a specific hardware aggregate shaper must have the same scheduling class. This is because the software algorithm enforcing the hardware aggregate scheduler policy only controls the rate of the hardware aggregate shaper and not the rate of individual queues. If this recommendation is not followed, weights enforced at the group level may contradict weights enforced between queues within a single hardware aggregate shaper.

  • Elevation-weight and groups for any set of queues are mutually exclusive.

The hardware aggregate scheduler policy implements a congestion threshold that is used to trigger software recalculation. If the overall offered load for a specific Vport is below this threshold (expressed as a percentage of the Vport aggregate rate), the software recalculation of all parameters is not performed to save CPU resources because it is assumed that all offered traffic passes through.

Similarly to a port scheduler policy, a hardware aggregate rate scheduler policy supports a monitor threshold that detects congestion. The configuration and operation are identical as described in Scheduler QoS policies.

The following example hardware aggregate shaper scheduler policy describes the SLA model from SLA models (a).

MD-CLI

[ex:/configure qos hw-agg-shaper-scheduler-policy "trial"]
A:admin@node-2# info
    max-percent-rate 80.0
    group "g1" { }
    sched-class 3 {
        group "g1"
        weight 1
    }
    sched-class 4 {
        group "g1"
        weight 8
    }

[ex:/configure port 1/1/1/c1/1 ethernet access egress]
A:admin@ode-2# info
    virtual-port "dslam-a" {
        hw-agg-shaper-scheduler-policy "trial"
    }

classic CLI

A:node-2>config>qos>hw-agg-shap-sched-plcy# info
----------------------------------------------
            max-rate percent 80.00
            group "g1" create
            exit
            sched-class 3 group "g1" weight 1
            sched-class 4 group "g1" weight 8
----------------------------------------------
*A:node-2# /configure port 1/1/1/c1/1 ethernet access egress
*A:node-2>config>port>ethernet>access>egress# info
----------------------------------------------
                    vport "dslam-a" create
                        hw-agg-shaper-scheduler-policy "trial"
                    exit
----------------------------------------------

Interactions with other QoS features

The FP4-based and later chipsets support all existing QoS features including software-driven HQoS and hardware-assisted HQoS. The following sections describe how these features interact.

Activating hardware aggregate shapers

Hardware aggregate shapers are applicable to Ethernet and PXC port subscriber management and SAPs; for other objects, such queue groups, only software-driven H-QoS can be used. Hardware aggregate shapers and software-driven H-QoS can coexist on the same FP.

  1. Use the following command to enable hardware aggregate shapers on a specific FP:
    • MD-CLI
      configure qos fp-resource-policy aggregate-shapers queue-sets size allocation-weight auto-creation true
    • classic CLI
      configure qos fp-resource-policy aggregate-shapers queue-sets size allocation-weight

    There are seven queue set sizes available for queues two to eight. Queue sets with different sizes are allowed. The distribution of hardware aggregate shapers between different queue sets is controlled through the allocation-weight command. The default queue set size is 8.

  2. Use the following command to configure the number of queues for use outside of the hardware aggregate shapers:
    • MD-CLI
      configure qos fp-resource-policy aggregate-shapers reserved-non-shaper-queues
    • classic CLI
      configure qos fp-resource-policy aggregate-shapers reserved-non-shaper-queues

    When using a queue set with less than eight queues, a certain number of queues will be unused by the hardware aggregate shapers. These queues can be used for objects not using hardware aggregate shapers. To use these queues for this purpose, the reserved-non-shaper-queues command must be configured. The minimum number of reserved non-shaper queues is set to 2048, which covers all system-related requirements.

  3. Use the following command to configure a default queue set for subscribers or SAPs:
    • MD-CLI
      configure qos fp-resource-policy aggregate-shapers queue-sets default-size saps <size>
      
      configure qos fp-resource-policy aggregate-shapers queue-sets default-size subscribers
    • classic CLI
      configure qos fp-resource-policy aggregate-shapers queue-sets default-size saps
      
      configure qos fp-resource-policy aggregate-shapers queue-sets default-size subscribers

    The default queue set can be set to the queue set size which has an allocation weight equal to 100. It can also be set to use non-shaper queues, by using the reserved-non-shaper-queues keyword.

    This step may appear to be redundant, since there is only a single queue set supported. However, it is important to configure this to ensure forward compatibility of the configuration file for future releases.

  4. Use the following command to enable hardware aggregate shapers.
    1. Use the following command to enable hardware aggregate shapers for subscribers:
      • MD-CLI
        configure qos fp-resource-policy aggregate-shapers subscribers true
      • classic CLI
        configure qos fp-resource-policy aggregate-shapers hw-agg-shaping subscribers
    2. Use the following command to enable hardware aggregate shapers for SAPs:
      • MD-CLI
        configure qos fp-resource-policy aggregate-shapers saps true
      • classic CLI
        configure qos fp-resource-policy aggregate-shapers hw-agg-shaping saps
  5. Use the following command to enable hardware aggregate shapers.
    1. Use the following command to enable the use of hardware aggregate shaping on all ports of the FP:
      • MD-CLI
        configure qos fp-resource-policy ports hqos-mode hw-agg-shaping true
      • classic CLI
        configure qos fp-resource-policy ports hqos-mode hw-agg-shaping
      MD-CLI
      [ex:/configure qos fp-resource-policy "test"]
      A:admin@node-2# info
      
          aggregate-shapers {
              
              hw-agg-shapers {
                  subscribers true
                  saps true
              }
          }
          ports {
              hqos-mode hw-agg-shaping
          }
      classic CLI
      A:node-2>config>qos>fp-resource-policy# info 
      ---------------------------------------------- 
          aggregate-shapers 
                  hw-agg-shapers subscribers saps 
          exit
          ports 
                  hqos-mode hw-agg-shaping 
          exit 
      ----------------------------------------------
    2. Use the following command to enable hardware aggregate shapers on a specific FP:
      • MD-CLI
        configure qos fp-resource-policy aggregate-shapers auto-creation true
      • classic CLI
        configure qos fp-resource-policy aggregate-shapers auto-creation

      For chassis-based systems, this causes a reboot of the card on which the FP resides. For non-chassis-based systems, the user must save the configuration and reboot the system manually. When an FP resource policy is in operation, the allocation of queue sets is in blocks as defined by queue set size. This occurs for all objects already configured on the FP, regardless of how many queues they effectively use. If there are not enough queues to perform the allocation, the commit command fails in MD-CLI.

      Given that the auto-creation command effectively forces the reset of FPs when it is applied, it is recommended that all configuration steps defined above are performed in a single commit, when using model-driven configuration mode. In classic configuration mode, steps 1 to 4 must be performed before the auto-creation command is activated.

      The auto-creation command does not force subscriber hosts to use a hardware aggregate shaper. Even if this command is enabled, subscriber hosts on the specified FP still use traditional H-QoS if it is configured. Any hardware aggregate shaper-related configuration in the SAP egress profile, SLA profile, and sub-profile is ignored except in the scheduler class configuration, where, by default, expedited and non-expedited classes are mapped into scheduling classes as follows:

      • non-expedited class queues map to sched-class 4

      • expedited class queues map to sched-class 6

  6. Use the following command to assign a specific queue set size to a specified SAP or subscriber SLA profile.
    The specific queue set size or the use of non-shaper queues is defined within SAP egress as follows:
    • MD-CLI
      configure qos sap-egress hw-agg-shaper-queues queue-set-size {number | non-shaper-queues}
    • classic CLI
      configure qos sap-egress hw-agg-shaper-queues {queue-set-size number | non-shaper-queues}
    The appropriate SAP egress is associated with a SAP or SLA profile. Each SAP or SLA profile instance then draws from the pool of the queue set sizes (or non-shaper queues) as defined in Step 1, at the FP level.

After the preceding commands are applied to a specific FP, all subscriber hosts created on this FP use hardware aggregate shaping and any traditional H-QoS-related configuration is ignored.

SAP and queue group objects always use traditional H-QoS configuration whether they are created on the same or different ports as subscriber hosts. Any hardware aggregate shaping configuration associated with a SAP or queue group is ignored.

Hardware aggregate shapers and LAG

LAGs using link or port-fair adapt-qos modes support hardware aggregate shapers; LAGs in distribute mode do not.

When a LAG spans multiple FPs, only one set of subscriber queues is created per FP (on the elected port); consequently, there is only one hardware aggregate shaper created per FP for subscribers on a LAG. Egress subscriber traffic is hashed only to the active FP for that subscriber.

For SAPs, the queues and hardware aggregate shapers are instantiated on each member port. The operational hardware aggregate shaper rate is allocated to individual instances based on the LAG mode:

  • in link mode all instances have the same rate
  • in port-fair mode the rate is proportional to the port rates

The Vport bandwidth distribution algorithm manages only the hardware aggregate shaper associated with the elected LAG port on the active FP. All hardware aggregate shapers not associated with the active FP have their oper-PIRs set to their admin-PIRs. This eliminates the need for the Vport bandwidth distribution algorithm to manage the nonactive hardware aggregate shapers and reduces the processing required.

Nokia recommends that for LAGs spanning different FPs, all participating FPs are configured to use hardware aggregate shapers. However, there are no checks to flag a mismatch of configuration between all member FPs.

Encapsulation offset

In traditional HQoS for subscribers it is possible to define the subscriber aggregate rates, which should be adjusted with respect to last-mile encapsulation using the following command:

  • MD-CLI
    configure subscriber-mgmt sub-profile egress qos encap-offset type
  • classic CLI
    configure subscriber-mgmt sub-profile egress encap-offset type

There are multiple types of encapsulations available.

This causes the offered-load measurements and operational-rate values to be normalized with respect to encap-offset and avg-frame-overhead values. Based on these values, the corresponding operational rates are calculated, and the adjustment is transparent to users.

For hardware aggregate shapers, the shapers and FIR distribution rely on hardware enforcement while the hardware aggregate shaper scheduler policy is software-based. Because of this, queue weights in FIR calculations are normalized with respect to the encap-offset value. Therefore, in show commands related to hardware aggregate shapers the queue weights may differ from the configured values due to such adjustments.

However, weights in hardware aggregate shaper scheduler policy remains as configured, because offered-load and rate calculations in software consider last-mile overhead in the same way as in software-driven HQoS.

If no encap-offset is set for the hardware aggregate shaper, the on-the-wire-rate is enforced. Considering that offered-rates do not account for on-the-wire Ethernet overhead, the QoS software adjusts the aggregate rate of the shaper based on the average frame overhead calculated during HQoS calculation.

Advanced QoS configuration policy

The following configuration parameters are not applicable when hardware aggregate shaping is used:

  • configure qos adv-config-policy name child-control bandwidth-distribution

    • above-offered-allowance delta-consumed-agg-rate

    • above-offered-allowance delta-consumed-agg-rate percent

    • above-offered-allowance delta-consumed-higher-tier-rate

    • above-offered-allowance delta-consumed-higher-tier-rate percent

    • above-offered-allowance unconsumed-agg-rate

    • above-offered-allowance unconsumed-agg-rate percent

    • above-offered-allowance unconsumed-higher-tier-rate

    • above-offered-allowance unconsumed-higher-tier-rate percent

    • above-offered-cap percent

    • above-offered-cap rate

    • enqueue-on-pir-zero

    • granularity percent

    • granularity rate

    • internal-scheduler-weight-mode

    • limit-pir-zero-drain

    • lub-init-min-pir

  • configure card 1 virtual-scheduler-adjustment

    • internal-scheduler-weight-mode