Cisco ONS 15454 SONET/SDH ML-Series Multilayer Ethernet Card Software Feature and Configuration Guide, Release 4.6

Configuring Ethernet over MPLS

This chapter describes how to configure Ethernet over Multiprotocol Label Switching (EoMPLS) on the ML-Series card.

This chapter includes the following major sections:

•Understanding EoMPLS

•Configuring EoMPLS

•EoMPLS Configuration Example

•Monitoring and Verifying EoMPLS

Understanding EoMPLS

EoMPLS provides a tunneling mechanism for Ethernet traffic through an MPLS-enabled Layer 3 core. It encapsulates Ethernet protocol data units (PDUs) inside MPLS packets and using label stacking forwards them across the MPLS network. EoMPLS is an Internet Engineering Task Force (IETF) standard-track protocol based on the Martini draft, specifically the draft-martini-l2circuit-encap-mpls-01 and draft-martini-l2circuit-transport-mpls-05 sections.

EoMPLS allows service providers to offer customers a virtual Ethernet line service or VLAN service using the service provider's existing MPLS backbone. It also simplifies service provider provisioning, since the provider edge customer-leading edge (PE-CLE) equipment only needs to provide Layer 2 connectivity to the connected customer edge (CE) equipment.

Figure 17-1 shows an example of EoMPLS implemented on a service provider network. In the example, the ML-Series card acts as PE-CLE equipment connecting to the Cisco GSR 12000 Series through an RPR access ring. Point-to-point service is provided to CE equipment in different sites that connect through ML-Series cards to the ML-Series card RPR access ring.

Figure 17-1 EoMPLS Service Provider Network

Implementing EoMPLS on a service provider network requires ML-Series card interfaces to play three major roles. The ML-Series card interface roles must be configured on both sides of the EoMPLS point-to-point service crossing the MPLS core.

•ML-Series card interfaces connect the provider's network directly to the customer edge equipment and are known as the PE-CLE interfaces. This PE-CLE interface on the ML-Series card is FastEthernet or GigabitEthernet and is configured to be an endpoint on the EoMPLS point-to-point session.

•An ML-Series card interface bridges the PE-CLE interface and the RPR network of ML-Series cards. This RPR/SPR interface contains POS ports and is configured for MPLS IP.

•An ML-Series card interface connects to a core MPLS interface. This interface is GigabitEthernet or FastEthernet and connects to the port of a Cisco GSR 12000 Series or similar device that is on the MPLS network. This MPLS cloud-facing interface bridges the SPR interface and the MPLS cloud.

Implementing EoMPLS across a service provider's network requires setting up directed Label Distribution Protocol (LDP) sessions (LSPs) between the ingress and egress PE-CLE routers to exchange information for a virtual circuit (VC). Each VC consists of two LSPs, one in each direction, since an LSP is a directed path to carry Layer 2 frames in one direction only.

EoMPLS uses a two-level label stack to transport Layer 2 frames, where the bottom/inner label is the VC label and the top/outer label is the tunnel label. The VC label is provided to the ingress PE-CLE by the egress PE-CLE of a particular LSP to direct traffic to a particular egress interface on the egress PE-CLE. A VC label is assigned by the egress PE-CLE during the VC setup and represents the binding between the egress interface and a unique and configurative VC ID. During a VC setup, the ingress and egress PE-CLE exchange VC label bindings for the specified VC ID.

An EoMPLS VC on the ML-Series card can transport an Ethernet port or an IEEE 802.1Q VLAN over MPLS. A VC type 5 tunnels an Ethernet port and a VC type 4 transports a VLAN over MPLS. In a VC type 5 session, the user can expect any traffic that is received on an ML-Series card PE-CLE port with an mpls l2transport route command to be tunneled to the remote egress interface on the far-end ML-Series card PE-CLE port. With a VC type 4, a user can expect the tunnel to act as physical extension to that VLAN. The EoMPLS session commands are entered on a VLAN subinterface on the PE-CLE, and only VLAN-tagged traffic received on that port will be tunneled to the remote PE-CLE.

EoMPLS Support

In Software Release 4.6, EoMPLS on the ML-Series card has the following characteristics:

•EoMPLS is only supported on FastEthernet and GigabitEthernet interfaces or subinterfaces.

•MPLS tag switching is only supported on SPR interfaces.

•Class of service (CoS) values are mapped to the experimental (EXP) bits in the MPLS label, either statically or by using the IEEE 802.1p bits (default).

•The ingress PE-CLE ML-Series card sets the time-to-live field to 2 and the tunnel label to a value of 255.

•Ingress PE-CLE ML-Series cards set the S bit of the VC label to 1 to indicate that the VC label is at the bottom of the stack.

•Since EoMPLS traffic is carried over the RPR, whatever load balancing is applicable for the traffic ingressing RPR is also applicable for the EoMPLS traffic.

•The Ethernet over MPLS feature is part of the Cisco Any Transport over MPLS (AToM) product set. The ML-Series card implementation of EoMPLS is based on Cisco IOS 12.1 E.

•The ML-Series card hosting the EoMPLS endpoint ports must be running the MPLS microcode image to support EoMPLS. For more information on multiple microcode images, see the "Multiple Microcode Images" section. Other ML-Series cards in the RPR are not restricted to the MPLS microcode image.

EoMPLS Restrictions

In Software Release 4.6, EoMPLS on the ML-Series card has the following restrictions:

•Packet-based load balancing is not supported. Instead, circuit-ID based load balancing is used.

•Zero hop or hairpin VCs are not supported. A single ML-Series card cannot be both the source and destination for a VC.

•MPLS control word for sequencing of data transmission is not supported. Packets must be received and transmitted without control word.

•Sequence checking or resequencing of EoMPLS traffic is not supported. Both depend on the control word to function.

•Maximum transmission unit (MTU) fragmentation is not supported.

•Explicit-null label for back-to-back LDP sessions is not supported.

Caution Since MTU fragmentation is not supported across the MPLS backbone, the network operator must make sure the MTU of all intermediate links between endpoints is sufficient to carry the largest Layer 2 PDU.

EoMPLS Quality of Service

The EXP is a 3-bit field and part of the MPLS header. It was created by the IETF on an experimental basis, but later became part of the standard MPLS header. The EXP bits in the MPLS header carry the packet priority. Each label switch router along the path honors the packet priority by queuing the packet into the proper queue and servicing the packet accordingly.

By default, the ML-Series card does not map the IEEE 802.1P bits in the VLAN tag header to the MPLS EXP bits. The MPLS EXP bits are set to a value of 0.

There is no straight copy between Layer 2 CoS and MPLS EXP, but the user can use the set mpls experimental action to set the MPLS EXP bit values based on a match to 802.1p bits. This mapping occurs at the entry point, the ingress of the network.

Quality of service (QoS) for EoMPLS traffic on ML-Series cards uses strict priority and/or weighted round robin scheduling in the egress interface of both imposition and disposition router. This requires selection of the service class queue that determines the type of scheduling. In the imposition router, the priority bits EXP or RPR CoS that are marked based on policing are used to select the service class queue and in the disposition router, the dot1p CoS bits (which are copied from EXP bits of the labels) are used to do the same. In addition to scheduling in the egress interface, the output policy action can also include remarking of EXP and RPR CoS bits.

EoMPLS on the ML-Series card uses the Cisco Modular Quality of Service Command-Line Interface (MQC), just like the standard QoS on the ML-Series card. But the full range of MQC commands are not available. Table 17-1 lists the applicable MQC statements and actions for the ML-Series card interfaces.

Configuring EoMPLS

The ML-Series peer cards on both endpoints of the EoMPLS point-to-point service must be configured. Perform the following configuration tasks to enable EoMPLS:

•VC Type 4 Configuration on PE-CLE Port (Either VC type 4 or VC type 5 is required.)

•VC Type 5 Configuration on PE-CLE Port (Either VC type 4 or VC type 5 is required.)

•EoMPLS Configuration on PE-CLE SPR Interface (Required)

•Bridge Group Configuration on MPLS Cloud-facing Port (Required)

•Setting the Priority of Packets with the EXP

EoMPLS Configuration Guidelines

These are the guidelines for configuring EoMPLS:

•Loopback addresses are used to specify the peer ML-Series card's IP address.

•LDP configuration is required. The default Tag Distribution Protocol (TDP) will not work.

•EoMPLS uses LDP targeted session between the ML-Series cards to create the EoMPLS VCs.

•The MPLS backbone must use an Interior Gateway Protocol (IGP) routing protocol, for example, Intermediate System-to-Intermediate System (IS-IS) Protocol or Open Shortest Path First (OSPF).

•Tag switching of IP packets must be enabled on the SPR interface for the PE-CLE ML-Series card.

VC Type 4 Configuration on PE-CLE Port

The customer-facing FastEthernet or GigabitEthernet port must be provisioned with EoMPLS and a VC type 4 or type 5. Interface GigE 0.1 on card A and card C plays the VC type 4 role in Figure 17-2. For more information on the role of a VC type 4, see the "Understanding EoMPLS" section.

To provision a VC type 4, which transport IEEE 802.1Q VLAN packets between two PE-CLE ML-Series cards, perform the following procedure on the customer facing port, beginning in global configuration mode:

VC Type 5 Configuration on PE-CLE Port

The customer-facing FastEthernet or GigabitEthernet port must be provisioned with EoMPLS and a VC type 4 or type 5. Interface GigE 1 on card A and card C plays the VC type 5 role in Figure 17-2. For more information on the role of a VC type 5, see the "Understanding EoMPLS" section.

To provision a VC type 5, which transports the configured port's packets between two PE-CLE ML-Series cards, perform the following procedure on the customer facing port, beginning in global configuration mode:

EoMPLS Configuration on PE-CLE SPR Interface

To enable the RPR to act as an access ring for the MPLS cloud, you must provision the SPR interface on the same ML-Series card that hosts the EoMPLS PE-CLE FastEthernet or GigabitEthernet interfaces. Interface SPR 1 on card A and card C plays this role in Figure 17-2.

To provision the SPR interface for MPLS, perform the following procedure, beginning in global configuration mode:

Router(config)# mpls label protocol 
ldp

Router(config)# interface spr 1

Router(config-if)# mpls ip

Router(config-if)# end

Router# copy running-config 
startup-config

Bridge Group Configuration on MPLS Cloud-facing Port

A FastEthernet or GigabitEthernet port from an ML-Series card in the RPR must connect to the interface of a router that is part of the MPLS cloud. A bridge group must be created that contains this FastEthernet or GigabitEthernet port and the SPR subinterface. Interface GigE 0 on card B and card D plays this role in Figure 17-2.

To provision the MPLS cloud-facing port for EoMPLS, perform the following procedure, beginning in global configuration mode:

Router(config)# bridge bridge-group-number 
protocol {rstp | ieee}

Router(config)# interface {GigabitEthernet 
| FastEthernet} interface-number

Router(config-if)# bridge-group 
bridge-group-number

Router(config-if)# no shutdown

Router(config)# interface 
spr 1.subinterface-number

Router(config-if)# bridge-group 
bridge-group-number

Router(config-if)# end

Router# copy running-config startup-config

Setting the Priority of Packets with the EXP

Ethernet over MPLS provides QoS using the three EXP bits in a label to determine the priority of packets. To support QoS between ML-Series card point-to-point endpoints, set the experimental bits in both the VC and tunnel labels.

Perform the following steps to set the experimental bits:

Router(config)# class-map 
class-name 

Router(config-cmap)# match 
any 

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

Router(config-pmap)# class 
class-name 

Router (config-pmap-c)# set 
mpls experimental value

Router(config)# interface 
GigabitEthernet 
interface-number

or

interface FastEthernet 
interface-number

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

EoMPLS Configuration Example

Figure 17-2 illustrates the sample network that the configuration commands reference. Examples 17-1, 17-2, 17-3, and 17-4 list relevant portions of the configuration files for enabling EoMPLS on ML-Series cards in a sample network.

Figure 17-2 EoMPLS Configuration Example

Example 17-1 ML-Series Card A Configuration

microcode mpls

ip subnet-zero

no ip domain-lookup

!

mpls label protocol ldp

!

interface Loopback0

 ip address 10.10.10.10 255.255.255.255

!

interface SPR1

 ip address 100.100.100.100 255.255.255.0

 no keepalive

spr station-id 1 

 mpls ip

 hold-queue 150 in

!

interface GigabitEthernet0

 no ip address

!

interface GigabitEthernet0.1

 encapsulation dot1Q 10

 mpls l2transport route 3.3.3.3 1 

!

interface GigabitEthernet1

 no ip address

 mpls l2transport route 4.4.4.4 2 

!

interface POS0

 no ip address

 spr-intf-id 1

 crc 32

!

interface POS1

 no ip address

 spr-intf-id 1

 crc 32

router ospf 1

 log-adjacency-changes

 network 1.1.1.0 0.0.0.255 area 0

 network 10.10.10.0 0.0.0.255 area 0

!

ip classless

no ip http server

Example 17-2 ML-Series Card B Configuration

bridge 10 protocol ieee

!

!

interface SPR1

no ip address

no keepalive

 bridge-group 10

 hold-queue 150 in

!

interface GigabitEthernet0

no ip address

bridge-group 10

Example 17-3 ML-Series Card C Configuration

microcode mpls

ip subnet-zero

no ip domain-lookup

!

mpls label protocol ldp

!

interface Loopback0

 ip address 20.20.20.20 255.255.255.255

!

interface SPR1

 ip address 100.100.100.100 255.255.255.0

 no keepalive

 spr station-id 4 

 mpls ip

 hold-queue 150 in

!

interface GigabitEthernet0

 no ip address

!

interface GigabitEthernet0.1

 encapsulation dot1Q 10

 mpls l2transport route 1.1.1.1 1 

!

interface GigabitEthernet1

 no ip address

 mpls l2transport route 2.2.2.2 2 

!

interface POS0

 no ip address

 spr-intf-id 1

 crc 32

!

interface POS1

 no ip address

 spr-intf-id 1

 crc 32

!

router ospf 1

 log-adjacency-changes

 network 1.1.1.0 0.0.0.255 area 0

 network 10.10.10.0 0.0.0.255 area 0

!

ip classless

no ip http server

Example 17-4 ML-Series Card D Configuration

bridge 20 protocol ieee

!

!

interface SPR1

no ip address

no keepalive

 bridge-group 20

 hold-queue 150 in

!

interface GigabitEthernet0

no ip address

bridge-group 20

Monitoring and Verifying EoMPLS

Table 17-2 shows the privileged EXEC commands for monitoring and verifying EoMPLS.

show mpls l2transport vc

show mpls l2transport vc detailed

show mpls l2transport vc summary

show mpls l2transport vc vc-id