This chapter provides information to configure network QoS policies using the CLI.
6.1. Overview of Network QoS Policies
Network QoS policies have an ingress and egress component, which define the QoS processing behavior provided for packets that ingress the access-uplink port and egress the access-uplink port, respectively.
The ingress component of the port policy defines how the dot1p bits are mapped to the internal FC and profile state. The FC and profile state define the PHB and the QoS handling through the system. The mapping on each access uplink port defaults to the mappings defined in the default network QoS policy until an explicit policy is defined for the access uplink ports. It also defines the bandwidth-limiting parameters for the traffic mapped to each FC. Traffic mapped to each FC can be limited to configurable bandwidth values using separate meters for unicast traffic and multipoint traffic.
| Note: 7210 SAS platforms provide different mechanisms to limit the bandwidth per FC. On 7210 SAS-D and 7210 SAS-Dxp, the user needs to use policers/meters to rate-limit the traffic per FC. |
The egress component of the network QoS ip-interface policy type defines the marking values associated with each FC.
Access uplink port egress marking support consists of the following:
For packets sent out of a access-uplink port, the network QoS policy defines the marking values (for example, IEEE 802.1p bits, and so on) to use based on the FC and the profile state.
The default map of FC to marking values (for example, 802.1p bits) is as shown in the default network QoS policy, policy id 1.
All non-default network QoS policies inherit the default map and can be modified by the user.
On 7210 SAS-D and 7210 SAS-Dxp, remarking can be enabled or disabled on access uplink ports. An option is provided to map FC to dot1p and DEI bits and IP DSCP values.
Non-default network policy parameters can be modified. The no form of the command reverts to the default values.
Changes made to a policy are applied immediately to all access uplink ports where the policy is applied. For this reason, when a policy requires several changes, Nokia recommends that you copy the policy to a work area policy-id. The work-in-progress copy can be modified, and then the original policy-id can be overwritten using the config qos copy command.
See “CLI Usage” in the 7210 SAS-D, Dxp, K 2F1C2T, K 2F6C4T, K 3SFP+ 8C Basic System Configuration Guide for information about the tasks and commands necessary to access the command line interface, and to configure and maintain your devices.
6.1.1. Resource Allocation for Network QoS Policy
This section describes the allocation of QoS resources for network QoS policies. When the port mode is changed to access-uplink, a default network QoS policy is applied. For the default policy, two meters and eighteen classification entries in hardware are allocated.
| Note: The number of resources used per network QoS policy determines the number of access-uplink ports that can be configured. If more resources are used, fewer access-uplink ports can be used, and vice versa. |
For every FC in use, the system allocates two classification entries in hardware, provided the FC is configured to use both the unicast meter and the multicast meter, or provided the default meter 9 is configured in the policy. If multiple match criteria entries map to the same FC, each of these are allocated two classification entries in hardware. For example, if there are two match-criteria entries that map to FC “af” a total of four classification entries are allocated in hardware, and if there are four match-criteria entries that map to FC “af” a total of eight classification entries are allocated in hardware.
For every meter or policer in use, the system allocates one meter in hardware. A meter or policer is considered in use when it is associated with an FC in use.
To calculate the number of QoS resources used by an access-uplink port, determine the following items:
number of match-criteria entries used to identify the FC
number of FCs to use
Only the FCs used by the match-criteria classification entries are considered for the “number of FCs” and are therefore are referred to as “FCs in use”. Also, in network policies of the type “ip-interface", the default multipoint meter 9 is created in a policy; whereas, for policy of type “port”, the default multipoint meter needs to be explicitly configured by the user, if required.
Use the following rules to compute the number of classification entries per FC in use.
If an FC is in use and is created without explicit meters, use default meter 1 for unicast traffic and default meter 9 (if configured) for all other traffic types (that is, broadcast, multicast, and unknown-unicast). This requires two classification entries in hardware. If default multipoint meter 9 is not configured, the FC uses the unicast meter for all traffic types. In this case, the FC requires a single classification entry in hardware.
If an FC is in use and is created with an explicit unicast meter, use that meter for unicast traffic and use default meter 9 (if configured) for all other traffic types. This requires two classification entries in hardware. If default multipoint meter 9 is not configured, the FC uses the unicast meter for all traffic types. In this case, the FC requires a single classification entry in hardware.
If an FC is in use and is created with an explicit unicast meter and explicit multicast meter, use the unicast meter for unicast traffic and multicast meter for all other kinds of traffic. This requires two classification entries in hardware.
Given the number of match criteria and the number of FCs in use, use the following equation to calculate the total number of classification entries per policy (for example TC):
TC= ? 2 * E(i)
i=nc,h1,ef,h2,l1,af,l2,be
where:
E(i) is the number of match- criteria entries that classify packets to FCi. For 7210 SAS platforms, the maximum number of classification entries per policy can be 64 (including default).
2 is the number of classification entries that are required by FCi.
| Note: In the worst case, only 2 classification entries are used per FC in a network policy, because only two traffic types are supported. |
Determine number of policers or meters to use (for example TP). A maximum of 16 meters per network policy are available.
Only those meters that are associated with FCs need to be considered for number of meters; only FCs in use are considered.
6.1.1.1. Network QoS Policies Resource Usage Examples
This section provides usage examples for network QoS policies.
6.1.1.1.1. Example 1
network 1 create
description “default QoS policy”
ingress
default-action fc be profile out
meter 1 create
exit
meter 9 multipoint create
exit
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
The number of classification entries (TC) used is calculated as follows.
(2 * 0)nc + (2 * 0)h1 + (2 * 0)ef + (2 * 0)h2 + (2 * 0)l1 + (2 * 0)af + (2 * 0)l2 + (2 * 1)be = 18
The number of meters (TP) used is 2 (meter 1 and 9).
6.1.1.1.2. Example 2
network 2 create
description “network-policy-2”
ingress
default-action fc be profile out
meter 1 create
exit
meter 2 create
exit
meter 9 multipoint create
exit
meter 12 multipoint create
exit
fc "af" create
meter 2
multicast-meter 12
exit
dot1p 2 fc af profile out
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
exit
The number of classification entries (TC) used is calculated as follows.
(2 * 0)nc + (2 * 0)h1 + (2 * 0)ef + (2 * 0)h2 + (2 * 0)l1 + (2 * 1)af + (2 * 0)l2 + (2 * 1)be = 4
The number of meters (TP) used is 4 (meters 1,2,9,12).
6.1.1.1.3. Example 3
network 3 create
description “network-policy-3”
ingress
default-action fc be profile out
meter 1 create
exit
meter 2 create
exit
meter 9 multipoint create
exit
meter 12 multipoint create
exit
fc "af" create
meter 2
multicast-meter 12
exit
fc "be" create
meter 2
multicast-meter 12
exit
dot1p 2 fc af profile out
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
exit
The number of classification entries (TC) used is calculated as follows.
(2 * 0)nc + (2 * 0)h1 + (2 * 0)ef + (2 * 0)h2 + (2 * 0)l1 + (2 * 1)af + (2 * 0)l2 + (2 * 1)be = 4
The number of meters (TP) used is 2 (meters 2 and 12).
6.1.1.1.4. Example 4
network 4 create
description “network-policy-4”
ingress
default-action fc be profile out
meter 1 create
exit
meter 9 multipoint create
exit
dot1p 1 fc l2 profile in
dot1p 2 fc af profile out
dot1p 3 fc af profile in
dot1p 4 fc h2 profile in
dot1p 5 fc ef profile in
dot1p 6 fc h1 profile in
dot1p 7 fc nc profile in
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
exit
The number of filter entries (TC) used is calculated as follows.
(2 * 1)nc + (2 * 1)h1 + (2 * 1)ef + (2 * 1)h2 + (2 * 0)l1 + (2 * 2)af + (2 * 1)l2 + (2 * 1)be = 16
The number of meters (TP) used is 2 (meters 1 and 9).
6.1.1.1.5. Example 5
network 5 create
description “network-policy-5”
ingress
default-action fc be profile out
meter 1 create
exit
meter 2 create
exit
meter 9 multipoint create
exit
meter 12 multipoint create
exit
fc "af" create
exit
fc "be" create
exit
fc "ef" create
exit
fc "h1" create
exit
fc "h2" create
exit
fc "l2" create
exit
fc "nc" create
exit
dot1p 1 fc l2 profile in
dot1p 2 fc af profile out
dot1p 3 fc af profile in
dot1p 4 fc h2 profile in
dot1p 5 fc ef profile in
dot1p 6 fc h1 profile in
dot1p 7 fc nc profile in
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
The number of classification entries (TC) used is calculated as follows.
(2 * 1)nc + (2 * 1)h1 + (2 * 1)ef + (2 * 1)h2 + (2 * 0)l1 + (2 * 2)af + (2 * 1)l2 + (2 * 1)be = 16
The number of meters (TP) used is 2 (meters 1 and 9). Note that meters 2 and 12 are not accounted for, because they are not associated with any FC.
6.1.1.1.6. Example 6
network 6 create
description “network-policy-6”
ingress
default-action fc be profile out
meter 1 create
exit
meter 2 create
exit
meter 3 create
exit
meter 9 multipoint create
exit
meter 12 multipoint create
exit
fc "af" create
meter 2
multicast-meter 12
exit
fc "be" create
exit
fc "ef" create
exit
fc "h1" create
meter 3
exit
fc "h2" create
exit
fc "l2" create
exit
fc "nc" create
meter 3
exit
dot1p 1 fc l2 profile in
dot1p 2 fc af profile out
dot1p 3 fc af profile in
dot1p 4 fc h2 profile in
dot1p 5 fc ef profile in
dot1p 6 fc h1 profile in
dot1p 7 fc nc profile in
exit
egress
fc af
exit
fc be
exit
fc ef
exit
fc h1
exit
fc h2
exit
fc l1
exit
fc l2
exit
fc nc
exit
exit
exit
The number of classification entries (TC) used is calculated as follows.
(2 * 1)nc + (2 * 1)h1 + (2 * 1)ef + (2 * 1)h2 + (2 * 0)l1 + (2 * 2)af + (2 * 1)l2 + (2 * 1)be = 16
The number of meters (TP) used is 5 (meters 1,2,3,9, and 12).
6.1.1.1.7. Example 7
network 2 create
description "network-policy 2"
scope template
ingress
default-action fc be profile out
meter 1 create
mode trtcm
adaptation-rule cir closest pir closest
rate cir 0 pir max
mbs default
cbs default
exit
meter 9 multipoint create
mode trtcm
adaptation-rule cir closest pir closest
rate cir 0 pir max
mbs default
cbs default
exit
network-policy 2 0 fc be profile out
network-policy 2 1 fc l2 profile in
network-policy 2 2 fc af profile out
network-policy 2 3 fc af profile in
network-policy 2 4 fc h2 profile in
network-policy 2 5 fc ef profile in
network-policy 2 6 fc h1 profile in
network-policy 2 7 fc nc profile in
exit
egress
no remarking
The number of classification entries (TC) used is 18.
The number of meters (TP) used is 2.
6.1.1.1.8. Example 8
network 8 create
description “network-policy-8”
ingress
default-action fc nc profile in
meter 1 create
exit
meter 2 create
exit
meter 3 create
exit
meter 4 create
exit
meter 5 create
exit
meter 7 multipoint create
exit
meter 8 multipoint create
exit
meter 9 multipoint create
exit
meter 12 multipoint create
exit
fc "af" create
meter 2
multicast-meter 12
exit
fc "ef" create
meter 4
multicast-meter 8
exit
fc "h2" create
exit
fc "l2" create
meter 3
multicast-meter 7
exit
fc "nc" create
meter 4
multicast-meter 8
exit
dot1p 1 fc l2 profile in
dot1p 3 fc af profile in
dot1p 5 fc ef profile in
dot1p 7 fc nc profile in
exit
egress
The number of classification entries (TC) used is calculated as follows.
(2 * 2)nc + (2 * 0)h1 + (2 * 1)ef + (2 * 0)h2 + (2 * 0)l1 + (2 * 1)af + (2 * 1)l2 + (0 * 0)be = 10
The numbers of meters (TP) used is 6 (meters 2, 3, 4, 7, 8, and 12).
6.1.2. Basic Configuration
A basic network QoS policy must:
have a unique policy ID
specify the default-action
have a QoS policy scope of template or exclusive
depending on the 7210 SAS platform, have at least one default unicast FC meter/queue
depending on the 7210 SAS platform, have at least one multipoint FC meter/queue
6.1.3. Create a Network QoS Policy
Configuring and applying QoS policies other than the default policy is optional. A default network policy of the appropriate type is applied to each uplink port.
To create an network QoS policy, define the following.
Specify a network policy ID value. The system does not dynamically assign a value.
Include a description that provides a brief overview of policy features.
Use egress marking and remarking to specify the egress FC to marking value (for example, IEEE 802.1p) map. Otherwise, the default values are applied.
remarking
If enabled, this command remarks ALL packets that egress on the specified access uplink port. The remarking is based on the FC to marking values mapping defined under the egress node of the network QoS policy. Remarking can be enabled or disabled.
FC criteria
The FC name represents an egress queue. Specify FC criteria to define the marking criteria of packets flowing through it.
marking value
The marking (for example, IEEE 802.1p) value is used for all packets requiring marking that egress on this FC queue that are in-profile or out-of-profile.
Specify ingress criteria using the following:
default action
Defines the default action to be taken for packets that have an undefined dot1p bit set. The default-action specifies the FC to which such packets are assigned.
dot1p
This specifies the dot1p to FC mapping for all packets. Ingress traffic that matches the specified criteria are assigned to the corresponding FC.
Use the following syntax to create a network QoS policy.
config>qos#
network policy-id [network-policy-type network-policy-type]
description description-string
scope {exclusive|template}
egress
remarking
fc {be|l2|af|l1|h2|ef|h1|nc}
dot1p-in-profile dot1p-priority
dot1p-out-profile dot1p-priority
default-action fc {fc-name} profile {in|out}
dot1p dot1p-priority fc {fc-name} profile {in|out}
fc {fc-name}
meter {meter-id}
multicast-meter {id}
meter meter-id [multipoint]
adaptation-rule cir {closest | max | min} pir {closest | max | min}
cbs {size-in-kbits}
mbs {size-in-kbits}
mode {trtcm | srtcm}
rate cir cir-rate-in-kbps [pir pir-rate-in-kbps]
The following commands associate a network QoS policy with the access-uplink port.
config>port
ethernet
access
uplink
qos network-policy-id
The following is a sample configuration output for uplink port 1/1/1 with network policy 600 applied to the interface.
A:ALA-7>config# info
#--------------------------------------------------
echo "Port Configuration"
#--------------------------------------------------
port 1/1/1
shutdown
description "port 1/1/1"
ethernet
mode access uplink
access
uplink
qos 600
exit
exit
exit
exit
...
#--------------------------------------------------
A:ALA-7>config#
6.1.4. Default Network Policy Values
The default network policy access uplink port is identified as policy-id “1”. Default policies cannot be modified or deleted. Table 29 lists default network policy parameters.
Table 29:
Default Network Policy 1
Field | Default |
description | Default network QoS policy |
Ingress | |
default-action | fc be profile out |
meter | 1 |
mode | trtcm1 |
adaptation-rule | cir closest |
pir closest |
rate | cir 0 |
pir max |
mbs | default kbits |
cbs | default kbits |
6.1.5. DSCP Marking for CPU-Generated Traffic
DSCP marking for CPU-generated traffic is not configurable by the user. Table 30 lists the default values.
| Note: DSCP and dot1p values in the table are applicable when remarking is disabled at the port level. |
Table 30:
DSCP and dot1p Marking
Protocol | IPv4 | DSCP Marking | dot1p Marking | Default
FC | DSCP Values | dot1p Values |
SNMP | Yes | Yes | Yes | H2 | 34 | 4 |
NTP | Yes | Yes | Yes | NC | 48 | 7 |
TELNET | Yes | Yes | Yes | H2 | 34 | 4 |
FTP | Yes | Yes | Yes | H2 | 34 | 4 |
TFTP | Yes | Yes | Yes | H2 | 34 | 4 |
SYSLOG | Yes | Yes | Yes | H2 | 34 | 4 |
TACACS | Yes | Yes | Yes | H2 | 34 | 4 |
RADIUS | Yes | Yes | Yes | H2 | 34 | 4 |
SSH | Yes | Yes | Yes | H2 | 34 | 4 |
ICMP Req | Yes | Yes | Yes | NC | 0 | 7 |
ICMP Res | Yes | Yes | Yes | NC | 0 | 7 |
ICMP Unreach | Yes | Yes | Yes | NC | 0 | 7 |
SCP | Yes | Yes | Yes | H2 | 34 | 4 |
CFM | NA | NA | Yes | NC | - | 7 |
ARP | NA | NA | Yes | NC | - | 7 |
SNMP trap/log | Yes | Yes | Yes | H2 | 34 | 4 |
ICMP ping | Yes | Yes | Yes | NC | 0 | 7 |
Trace route | Yes | Yes | Yes | NC | 0 | 7 |
TACPLUS | Yes | Yes | Yes | H2 | 34 | 4 |
IGMP | Yes | Yes | Yes | NC | 48 | 7 |
PTP (see note 1) | Yes | Yes | Yes | see note 1 | see note 1 | 7 |
Note:
Based on the type of the PTP message, that is, PTP event messages (for example, Sync message) and PTP non-event messages (for example, Announce, Follow-up), the DSCP value used is either 0x30 (h1) or 0x38 (nc), and the dot1p value is always 7.
6.1.6. Default DSCP Mapping Table
Table 31 lists default DSCP mapping values.
Table 31:
Default DSCP Mapping Table
DSCP Name | DSCP Value Decimal | DSCP Value Hexadecimal | DSCP Value Binary | Label |
Default | 0 | 0x00 | 0b000000 | be |
nc1 | 48 | 0x30 | 0b110000 | h1 |
nc2 | 56 | 0x38 | 0b111000 | nc |
ef | 46 | 0x2e | 0b101110 | ef |
af11 | 10 | 0x0a | 0b001010 | assured |
af12 | 12 | 0x0c | 0b001100 | assured |
af13 | 14 | 0x0e | 0b001110 | assured |
af21 | 18 | 0x12 | 0b010010 | l1 |
af22 | 20 | 0x14 | 0b010100 | l1 |
af23 | 22 | 0x16 | 0b010110 | l1 |
af31 | 26 | 0x1a | 0b011010 | l1 |
af32 | 28 | 0x1c | 0b011100 | l1 |
af33 | 30 | 0x1d | 0b011110 | l1 |
af41 | 34 | 0x22 | 0b100010 | h2 |
af42 | 36 | 0x24 | 0b100100 | h2 |
af43 | 38 | 0x26 | 0b100110 | h2 |
default 1 | 0 | |
Note:
The default FC mapping is used for all DSCP names or values for which there is no explicit FC mapping.
6.2. Service Management Tasks
This section provides information about service management tasks.
6.2.1. Deleting QoS Policies
A network policy is associated by default with access uplink ports.
You can replace the default policy with a non-default policy, but you cannot remove default policies from the configuration. When you remove a non-default policy, the policy association reverts to the appropriate default network policy.
6.2.2. Remove a Policy from the QoS Configuration
Use the following syntax to delete a network policy.
config>qos# no network network-policy-id
6.2.3. Copying and Overwriting Network Policies
You can copy an existing network policy to a new policy ID value or overwrite an existing policy ID. The overwrite option must be specified or an error occurs if the destination policy ID exists.
Use the following syntax to overwrite a network policy.
config>qos# copy network source-policy-id dest-policy-id [overwrite]
The following is a sample configuration output.
A:ALA-12>config>qos# info detail
---------------------------------------------
...
network 1 create
description "Default network QoS policy."
scope template
ingress
default-action fc be profile out
...
network 600 create
description "Default network QoS policy."
scope template
ingress
default-action fc be profile out
...
network 700 create
description "Default network QoS policy."
scope template
ingress
default-action fc be profile out
...
---------------------------------------------
A:ALA-12>config>qos#
6.2.4. Editing QoS Policies
You can change existing policies, except the default policies, and entries in the CLI. The changes are applied immediately to all access uplink ports where the policy is applied. To prevent configuration errors, use the copy command to make a duplicate of the original policy to a work area, make the edits, and then overwrite the original policy.