Cisco ONS 15454 and Cisco ONS 15454 SDH Ethernet Card Software Feature and Configuration Guide, Releases 9.0, 9.1, 9.2, and 9.2.1

Configuring Ethernet Virtual Circuits and QoS on the ML-MR-10 Card

This chapter provides information about configuring Ethernet virtual circuits (EVC) for the Cisco ONS 15454 ML-MR-10 card.

This chapter contains the following major sections:

•Understanding EVC

•Configuring EVC

The ML-MR-10 card employs the Cisco IOS Modular QoS command-line interface (CLI), known as the MQC. For more information on general MQC configuration, refer to the following Cisco IOS documents:

•Cisco IOS Quality of Service Solutions Configuration Guide, Release 12.2

•Cisco IOS Quality of Service Solutions Command Reference, Release 12.2

Understanding EVC

Ethernet virtual connection services (EVCS) uses the concepts of EVCs and service instances to provide Layer 2 switched Ethernet services. An EVC is an end-to-end representation of a single instance of a Layer 2 service being offered by a provider to a customer. It embodies the different parameters on which the service is being offered. A service instance is the instantiation of an EVC on a given port on a given ML-MR-10 card.

Configuring EVC

Establishing EVC on the ML-MR-10 card involves three basic operations:

•Configuring Layer 2 Ethernet services

•(Optional) Configuring quality of service (QoS)

•Configuring resilient packet ring (RPR) interfaces

Layer 2 Ethernet Services

IEEE 802.1Q (QinQ) mapping and service awareness on the ML-MR-10 card provides the following functionality:

•Only point-to-point EVC services are supported in Software Release 9.0 and above

•MAC address learning is not supported in Software Release 9.0 and above

•QinQ (1-to-2 translation) Layer 2 switching—QinQ adds an outer tag to the received dot1q traffic and then performs Layer 2 switching.

•Local VLAN significance—VLAN tags are significant only to the port.

•VLAN translation is supported

Restrictions and Usage Guidelines

When configuring QinQ Mapping and Service Awareness on ML-MR-10 cards, follow these restrictions and usage guidelines:

•Service Scalability:

–Service Instances: 8,000

–Bridge-domains: 4,000

•MQC actions supported include:

–Bandwidth/Weighted Deficit Round Robin (WDRR) queuing

–One priority queue per policy

–Police, set Class of Service (CoS) (marks 802.1p bits)

–Police, set QoS-group (egress queue number)

–Police, set discard-class (typically on non-edge nodes)

–Police, set rpr-ieee service-class {A| B| C}

–Valid for traffic out of RPR interface

–Qos-group is ignored for service-class A and B traffic

–Qos-group will be considered for service-class C traffic (C0, C1, C2, C3 queues are used for this case)

–Police, set discard-class {0| 1| 2} (0 = Green, 1 = Yellow, 2 = Red)

Configuring Layer 2

Use the following commands to configure Layer 2.

Router> enable

Examples

The following section provides examples for Configuring Ethernet Vitrual Circuits and QoS on ML-MR-10 Card

QinQ Point-to-Point EVC

In this example, an incoming frame with a dot1q tag of 10 enters GigabitEthernet1 and exits with a dot1q tag of 11. No MAC learning is involved.

!Customer facing port

Router(config)# interface GigabitEthernet1

Router(config-if)# service instance 100 ethernet

Router(config-if-srv)# encapsulation dot1q 10

Router(config-if-srv)# rewrite egress tag pop 1

Router(config-if-srv)# bridge-domain 10

!RPR IEEE 802.17 Ring facing port

Router(config)# interface rpr-ieee0

Router(config-if)# service instance 101 ethernet

Router(config-if-srv)# encapsulation dot1q 11

Router(config-if-srv)# rewrite egress tag push 1 dot1q 11

Router(config-if-srv)# bridge-domain 10

VLAN Translation

!Customer facing port

Router(config)# interface GigabitEthernet1

Router(config-if)# service instance 100 ethernet

Router(config-if-srv)# encapsulation dot1q 10

Router(config-if-srv)# rewrite egress tag translate 1-to-1 dot1q 10

Router(config-if-srv)# bridge-domain 10

!RPR IEEE 802.17 Ring facing port

Router(config)# interface rpr-ieee0

Router(config-if)# service instance 101 ethernet

Router(config-if-srv)# encapsulation dot1q 11

Router(config-if-srv)# rewrite egress tag translate 1-to-1 dot1q 11

Router(config-if-srv)# bridge-domain 10

Untagged Service Instance

!Customer facing port

Router(config)# interface GigabitEthernet1

Router(config-if)# service instance 11 ethernet

Router(config-if-srv)# encapsulation untagged

Router(config-if-srv)# rewrite egress tag pop 1

Router(config-if-srv)# bridge-domain 11

!RPR IEEE 802.17 Ring facing port

Router(config)# interface rpr-ieee0

Router(config-if)# service instance 11 ethernet

Router(config-if-srv)# encapsulation dot1q 10

Router(config-if-srv)# rewrite egress tag push dot1q 10 

Router(config-if-srv)# bridge-domain 11

Default Service Instance

The default service instance aggregates all the traffic except for the previously configured service instances on that interface.

Note The Default Service instance cannot be configured with any other Service instances. The default service instance captures all of the traffic on that interface including tagged, priority tagged, and untagged traffic.

!Customer facing port

Router(config)# interface GigabitEthernet2

Router(config-if)# service instance 12 ethernet

Router(config-if-srv)# encapsulation default

Router(config-if-srv)# rewrite egress tag pop 1

Router(config-if-srv)# bridge-domain 12

!RPR IEEE 802.17 Ring facing port

Router(config)# interface rpr-ieee0

Router(config-if)# service instance 12 ethernet

Router(config-if-srv)# encapsulation dot1q 12

Router(config-if-srv)# rewrite egress tag push dot1q 12

Router(config-if-srv)# bridge-domain 12

Verification

Use the following commands to verify operation.

EFP status has three possible states.

•UP

–Both the ingress and egress EFP physical interfaces IDB state should be UP. (One exception is that when CPP is enabled look at CPP state instead of IDB state)

–Hardware programming (TCAM) is intended for both the ingress and egress EFPs, and configuring both of the EFP encapsulation, rewrite, and bridge domains is a prerequisite to hardware programming

–When Service Advertisements are enabled for one of the EFP, the RPR destination should be in the Resolved status

•DOWN

–If any of the above criteria (for declaring the EFP status UP) is not met then the EFP status is declared DOWN.

•FLAPPING

–The ML-MR-10 card supports only point-to-point services. The S-VLAN (VLAN is the S-VLAN or outer tag) must be configured on exactly two RPR stations. If more than two RPR stations are configured to have the same service/S-VLAN, the Service Advertisement scheme will advertise two or more destinations for the same service. In this scenario the EVC platform can detect, and will declare, EFP status as FLAPPING.

Note An error message automatically displays on the console when there is a FLAPPING service. This is useful to determine which service is FLAPPING. When the EFP is in FLAPPING status it is service affecting because for a few seconds traffic goes to one destination and then another destination, and follows a circular path as multiple destinations keep advertising the other - the same S-VLAN.

The RPR destination field is valid only for the EFPs that are configured on RPR interfaces. It provides three possible types of output:

•Unresolved

–When the Service Advertisement cannot resolve the RPR destination for the point-to-point service because it is misconfigured, or due to operational errors (like interface down) then the RPR destination are displayed as Unresolved

•<mac-address> (learnt)

–When the Service Advertisement scheme is able to resolve the RPR destination, the MAC address and the keyword learnt, are displayed.

•<mac-address> (static)

–When the RPR destination is configured statically (using rpr-destination under service instance mode) the MAC address together with the keyword static, are displayed.

Sample Output

!show ethernet service instance plat

Router# show ethernet service instance platform 

NOTE: EFP status UP/DOWN is determined based on both ingress and egress interface states 
and RPR destination resolving status. EFP status FLAPPING means more than one RPR station 
is advertising this specific P2P service and need to check the network level config.

(*) RPR-destination field is valid only for EFPs configured on RPR interfaces

EFP-ID     Intf  EFP-Status RPR-Destination        

1          Gi0   DOWN       Not applicable         

1          Gi8   DOWN       Not applicable         

30         Gi8   DOWN       Not applicable         

30         RP0   UP 							 aabb.bbbb.cccc (static)

Configuring EtherChannel on ML-MR Card

You can configure an EtherChannel on the ML-MR card by creating an EtherChannel interface (port channel). All interfaces that are members of an EtherChannel should have the same link parameters, such as duplex and speed.

To create an EtherChannel interface, perform the following procedure beginning in global configuration mode:

Router(config)# interface 
port-channel channel-number

Router(config-if)# end

Router# copy running-config 
startup-config

To assign Ethernet interfaces to the EtherChannel perform the following procedure, beginning in global configuration mode:

Router(config)# interface 
gigabitethernet number

Router(config-if)#channel-gro
up 

channel number

Router(config-if)# end

Router# copy running-config 
startup-config

EtherChannel Configuration Example

Figure 32-1 shows an example of EtherChannel. The associated commands are provided in Example 13-1 (Switch A) and Example 13-2 (Switch B).

Figure 32-1 EtherChannel Configuration

Example 32-1 Switch A Configuration

hostname MLMR-A

!

interface Port-channel10

 no ip address

 no negotiation auto

 load-balance src-dst-mac

 service instance 20 ethernet

  encapsulation dot1q 20

   bridge-domain 20

!

interface GigabitEthernet0

 no ip address

 speed auto

 duplex auto

 negotiation auto

 channel-group 10

 no keepalive

!

interface GigabitEthernet1

 no ip address

 speed auto

 duplex auto

 negotiation auto

 channel-group 10

 no keepalive

!

interface POS0

 no ip address

 pos mode gfp

 service instance 20 ethernet

  encapsulation dot1q 20

   bridge-domain 20

 !

Example 32-2 Switch B Configuration

hostname MLMR-B

!

interface Port-channel10

 no ip address

 no negotiation auto

 load-balance src-dst-mac

 service instance 20 ethernet

  encapsulation dot1q 20

   bridge-domain 20

 !

interface GigabitEthernet0

 no ip address

 speed auto

 duplex auto

 negotiation auto

 channel-group 10

 no keepalive

!

interface GigabitEthernet1

 no ip address

 speed auto

 duplex auto

 negotiation auto

 channel-group 10

 no keepalive

!

interface POS0

 no ip address

 pos mode gfp

 service instance 20 ethernet

  encapsulation dot1q 20

   bridge-domain 20

 !

Configuring LACP on ML-MR

To configure LACP over the EtherChannel perform the following procedure beginning in global configuration mode:

Router(config)# int 
port-channel 
<interface-number>.

Router(config-if)# int 
gigabitEthernet 
<facility-number>

Router(config-if)# 
channel-group 

<channel-number> mode ?

Router(config-if)# 
channel-group 

<channel-number> mode active

Router(config-if)# exit

Router(config-if)# int 
gigabitEthernet 
<facility-number>

Router(config-if)# lacp-port

Router(config-if)# lacp 
port-priority 

<priority number>

Router(config-if)# exit

Router(config)#lacp sys

Router(config)#lacp 
system-priority <system 

priority>

Router(config-if)# exit

Router# copy running-config 
startup-config

Refer to Example 12-8 for LACP configuration example on the ML-MR-10 card.

EVC QoS Support

Refer to "Configuring Quality of Service" for information on QoS and for additional details on configuring the ML-Series cards.

Restrictions and Usage Guidelines

When configuring QoS with EVCS on the ML-MR-10 card, follow these restrictions and usage guidelines:

•Service instances use MQC.

•QoS supports 4,000 service instances.

•For ingress and egress QoS, only flat policy maps are supported.

•As service instance configurations change, Cisco recommends that you remove and reapply the policy because the policy map configuration may no longer be valid.

•Ethernet Flow Points (EFP) are supported on port channels.

EVC QoS supports:

•EVC QoS, classification is based on the following filters, which can be combined:

–Inner VLAN tag

–Outer VLAN tag

–CoS

•Egress Qos Policy supports the bandwidth and priority actions

–Bandwidth command - assigns the queue in Weighted Deficit Round Robin (WDRR) mode, and the bandwidth value can either be absolute value or a percentage value.

–Priority command - assigns the queue in strict priority mode; no bandwidth value needs to be associated with this queue.

–For RPR interfaces, only bandwidth in percentage values is supported for the bandwidth command.

–For port-channel interfaces it is recommended that you use percentage values for the bandwidth command, as the total bandwidth of a port-channel varies, based on member availability.

•EVC QoS, actions supported:

–Supports up to 4 queues per port scheduled with a WDRR scheduler.

–One egress strict priority queue per port.

–There are two modes of operation for the egress queues: all four queues operate in WDRR mode, or one queue is in strict priority mode and the other three are in WDRR mode.

–Mappings between the cos index values and the egress queue are allowed and achieved using the qos-group value.

–Re-marking of the IEEE 802.1p VLAN priority bits in a frame is supported.

•Policing:

–A 2-rate 3-color policer is supported.

–The two rates are cir (committed information rate) and pir (peak information rate).

–The two burst sizes corresponding to the rates are cbs (committed burst size) and pbs (peak burst size).

–The 3 colors correspond to the possible outcomes of the policer (conform, exceed and violate).

–Conform actions are transmit and set cos-transmit.

–Exceed actions are drop and set cos-transmit.

–Violate actions are drop and set cos-transmit.

•MOQ (Modular QOS CLI)

–Service policy is configurable on a physical Ethernet interface in the ingress direction.

–Service policy is configurable on a physical Ethernet interface in the egress direction.

–Service policy is configurable on an EtherChannel/ port-channel bundle interface in the ingress direction.

–Service policy is configurable on an EtherChannel/ port-channel bundle interface in the egress direction.

–If a physical interface is a member of an EtherChannel/ link aggregation bundle it will not allow configuration of any MQC service policy (service policies mst be configured on the logical etherchannel/ link aggregation bundle interface only).

–MQC service policy is configurable on an Ethernet Flow Point (EFP) in the ingress direction.

–Match cos 0-7 is supported on ingress service policies attached to physical Ethernet interfaces, and to EFPs, as long as the type of the EFP is not untagged (supports packet classification based on the IEEE 802.1p priority bits in the outermost 802.1Q VLAN header in the packet).

–Match ip dscp 0-63 is supported on physical interfaces in the ingress direction.

–Match ip precedence 0-7 is supported on physical interfaces in the ingress direction.

–Match ip dscp 0-63 is supported on EFPs in the ingress direction, as long as the type of the EFP is untagged.

–Match ip precedence 0-7 is supported on EFPs in the ingress direction as long as the type of the EFP is untagged.

–Set cos 0-7 is supported as a possible action in an ingress service policy (interpreted to mean setting of the 802.1p bits in the outermost 802.1Q header of the packet (if there is one) when the packet goes out the egress port).

–pir or pbs are set equal to cir and cbs.

–All actions in an egress service policy configured on an etherchannel bundle interface are applied equally across all members of the bundle (and not in an aggregate as on the ingress side). For example, it is not correct to calculate the egress bandwidth that can be reserved on an etherchannel bundle to be equal to the sum of the bandwidths of the individual members interfaces, instead, it is the minimum of the bandwidths of any of the members.

Port Channel QoS

QoS is supported on port-channel interfaces, and service policies are applied on the port-channel interfaces. Instead, of a member interface. At the ingress, the QoS actions are applied on the aggregate interface. At the egress, the bandwidth guaranteed on a port- channel interface is limited to one member interface's bandwidth. Policing can be configured up to the maximum port channel bandwidth.

QoS Classification

Use the QoS classification features to select your network traffic and categorize it into classes for further QoS processing based on matching certain criteria. The default class, named "class-default," is the class to which traffic is directed for any traffic that does not match any of the selection criteria in the configured class maps.

Note When class-default is applied on a physical interface, any traffic on the interface, regardless of the EVC, matches this class. When "class-default" is applied on an EFP, any traffic on this EFP matches the class.

Restrictions and Usage Guidelines

When configuring traffic classes on an ML-MR-10 card, follow these restrictions and usage guidelines:

•You can define up to four unique class maps per output Service Policy.

•You can define up to 65 class maps on ingress.

QoS Classifiers Supported on Various Frames on ML-MR-10 Card

The following are QoS classifiers supported on various frames on the ML-MR-10 card:

Configuring QoS Traffic Class

To create a user-defined QoS traffic class, use the following commands, beginning in global configuration mode:

Router(config)# class-map [match-all] 
class-name

Router(config-cmap)# match type

Configuring Policing

This section describes information for configuring QoS traffic policing policies.

Restrictions and Usage Guidelines

The ML-MR-10 card supports different forms of policing using the police command.

When configuring ingress policing on interfaces, and VLANs, follow these restrictions and usage guidelines:

•Policing on physical interfaces is supported.

•Policing on service instances is supported.

•Policing supports three actions:

–Transmit

–Set-cos transmit

–Drop

•Set-dscp-transmit is not supported

See Table 32-4 for Policer action details.

Table 32-4 Policer Actions Supported

Configuring QoS Traffic Policies

To create QoS traffic policies with policing, use the following commands beginning in global configuration mode:

Router(config-pmap-c)# police bps 
[burst-normal] [burst-max] 
conform-action action exceed-action 
action violate-action action.

Router(config-pmap-c)# police {cir 
cir} [bc conform-burst] {pir pir} [be 
peak-burst] [conform-action action 
[exceed-action action [violate-action 
action]]] 

Examples

Input service policy can be applied to a physical interface, or to the service instance.

This example shows classmap configuration;

Router(config)# class-map match-all voip_class

Router(config-cmap)# match cos 5

!

Single Rate Two color Policer

Router(config)# policy-map example_policy

Router(config-pmap)# class voip_class

Router(config-pmap-c)# police 1000000 25000 conform-action transmit exceed-action drop

!

Two rate three color marker example

Router(config)# policy-map remark_policy

Router(config-pmap)# class voip_class

Router(config-pmap-c)# police 50000000 1250000 2500000 pir 100000000 conform-action 
transmit exceed-action set-cos-transmit 3 violate-action drop

Verification

Use the following commands to verify policing:

This example shows an ingress service police with two color policer action specified.

Router# show policy-map interface 7

service-policy input: in

	class-map: i1(match-all)

	33239921 packets, 3058072732 bytes

	5 minute rate 70861000 bps, drop rate 776000 bps

	match: cos 1

	police:

	2000000 bps, 50000 limit

	conformed 120206 packets, 11895232 bytes; actions: transmit

	exceeded 33110625 packets, 3046177500 bytes; action: drop

	class-map: class-default (match-all)

	0 packets, 0 bytes

	5 minute offered rate 0 bps, drop rate 0 bps

	match: any

	20 packets, 0 bytes

	5 minute 0 bps

This example displays an egress service policy.

Router# show policy-map interface 9

GigabitEthernet9

  Service-policy output: 1

  Counters last updated 00:00:00 ago

	class-map: 1 (match-all)

	3 packets, 1014 bytes

	5 minute offered rate 0 bps, drop rate 0 bps

	match: qos-group 2

	class-map: 2(match-all)

	0 packets, 0 bytes

	5 minute offered rate 0 bps, drop rate 0 bps

	match: qos-group 1

	class-map: class-default (match-all)

	129296 packets, 12925268 bytes

	5 minute offered rate 270000 bps, drop rate 0 bps

	match: any

	3 packets, 990 bytes

	5 minute 0 bps

Associating a QoS Traffic Policy with an Interface or Service Instance

Before a traffic policy can be enabled for a class of traffic, it must be configured on an interface. A traffic policy also can be associated with Ethernet service instances.

Traffic policies can be applied for traffic coming into an interface (input) and for traffic leaving that interface (output).

Traffic policies can not be applied for output on service instance.

Associating a QoS Traffic Policy with an Input Interface

When you associate a traffic policy with an input interface, the policy is applied to traffic coming into that interface. To attach a traffic policy for an input interface, use the following command beginning in interface configuration mode:

Router(config-if)# service-policy 
input policy-map-name

Associating a QoS Traffic Policy with an Output Interface

When you associate a traffic policy with an output interface, the policy is applied to traffic leaving that interface. To attach a traffic policy to an output interface, use the following command beginning in interface configuration mode:

Configuring Marking

After you have created your traffic classes, you can configure traffic policies to configure marking features to apply certain actions to the selected traffic in those classes.

In most cases, the purpose of a packet mark is identification. After a packet is marked, downstream devices identify traffic based on the marking and categorize the traffic according to network needs. This categorization occurs when the match commands in the traffic class are configured to identify the packets by the mark (for example, match ip precedence, match ip dscp, match cos, and so on). The traffic policy using this traffic class can then set the appropriate QoS features for the marked traffic.

Restrictions and Usage Guidelines

When configuring class-based marking on an ML-MR-10 cards, follow these restrictions and usage guidelines:

•Marking is supported two ways:

–policing

–class-based marking

Configuring QoS Class-based Marking

To configure a QoS traffic policy with class-based marking, use the following commands beginning in global configuration mode:

Router(config)# policy-map 
policy-map-name

Router (config-pmap)# class 
{class-name | class-default}

Router(config-pmap-c)# set cos [0-7]

Examples

This example shows the creation of a service policy called policy1. This service policy is associated to a previously defined classification policy through the use of the class command. This example assumes that a classification policy called class1 was previously configured.

Router(config)# policy-map policy1

Router(config-pmap)# class class1

Router(config-pmap-c)# set cos 3

Verification

Use the following commands to verify marking:

For more detailed information about configuring class-based marking features, refer to the Class-Based Marking document located at the following URL:

http://www.cisco.com/en/US/docs/ios/12_2/qos/configuration/guide/qcfcbmrk.html

Configuring EVC on RPR-IEEE

Refer to Chapter 29 "Configuring IEEE 802.17b Resilient Packet Ring on the ML-MR-10 Card," for information on RPR-IEEE. Two commands are available: configure service advertisements, and configure static RPR destinations. Begin in global command mode and use the commands provided in the table below.

Restrictions and Usage Guidelines

When configuring RPR features on an ML-MR-10 card, follow these restrictions and usage guidelines:

•All packets use bandwidth only between the source and destination nodes on the ring associated with the EVC.

•Flooding of broadcast packets or unknown destination packets is not supported. The RPR connections are point-to-point only for each frame associated with a given EVC.

•An EVC will have two EFPs on the RPR ring: both EFPs share a unique VLAN tag.

•The EFP VLAN tag is added to the frame (along with the source and destination RPR node MAC addresses) when the frame is sent to the RPR ring. The EFP VLAN is removed from the frame (along with the source and destination RPR node MAC addresses) when the frame is received from the RPR ring.

•Flooding of broadcast packets or unknown destination packets is not supported on point-to-point EVCs.

•RPR connections are point to point only for each frame associated with a given point-to-point EVC.

•Attribute discovery frames (ATD) are used to advertise remote EFPs for each EVC on the RPR ring.

•Source MAC address learning of destination RPR nodes is not supported. ATD frames are used to determine the EFP mapping to the remote RPR node.

•EVC frames associated with an EFP are dropped and not sent to the RPR ring if there has been no ATD frame received for the remote EPF associated with the EVC.

•Traffic can be put in bandwidth queues based on QoS classification via MQC. This along with assigning an MQC class to an 802.17 Class of Service provides unequal local fairness within an 802.17 Class.

Router(config-if-srv)# bridge-domain 1

Router(config-if-srv)# encapsulation

Router(config-if-srv)# rewrite

Router(config-if-srv)# rpr-destination 
service-advertisement 

Router(config-if-srv)# service-policy 
policy_name

Router(config-if-srv)# rpr-destination 
static xxxx.xxxx.xxxxx

Configuring Service Domains

The following example shows a service domain configuration.

Note All of the existing services, except services with static RPR destinations configured, will be interrupted and re-assigned to the new domain-id

Examples

!

Router(config)interface RPR-IEEE0

Router(config-if) no ip address

Router(config-if) no ip route-cache

Router(config-if) no rpr-ieee sas

Router(config-if) rpr-ieee vlan-service-domain 10

Router(config-if-srv) service instance 10

Router(config-if-srv) ethernet encapsulation dot1q 10

Router(config-if-srv) bridge-domain 10

! 

EFP Configuration Combinations on the ML-MR-10 Card

There are three Types of EFP encapsulation:

•Untagged

•Default

•Tagged - Either 1 tag or 2 tag (Q-in-Q)

Three types of Egress Rewrite Operations are supported:

•POP tags

•Push tags

•Translate tags

Configuration Examples

Table 32-5 provides configuration examples for EFP combinations on the ML-MR-10 card.

1. The above mentioned combinations are the possible EFP configurations, but it has to be noted that not all the combinations would work with dynamic Service advertisements.

2. Also, some of the combinations would need to be used carefully based on network planning.

3. The above mentioned combinations hold good for the following EFP configurations - Gig-to-Gig, Gig--to--POS, Gig--to--RPR, POS--to--Gig, POS--to--POS, POS--to--RPR, RPR--to--Gig, RPR--to--POS.

4. Unsupported Configuration Combinations: Tag push is not supported - i.e. untagged --> Q-in-Q type of traffic.