Ethernet-over-MPLS (EoMPLS) provides a tunneling mechanism for Ethernet traffic through an MPLS-enabled Layer 3 core, and encapsulates Ethernet protocol data units (PDUs) inside MPLS packets (using label stacking) to forward them across the MPLS network.

The following table summarizes the load balancing behavior for VPLS and VPWS Ethernet bundle attachment circuits from Release 6.3.3 onwards. In the default configuration mode, the parameters used for load balancing through LAG Hashing is provided for disposition traffic flowing from MPLS network, for example, pseudowires to Ethernet attachment circuits.

Note	VLAN tags (Service and Customer) are not considered for load balancing.

The following sections describe the different modes of implementing EoMPLS.

Local switching involves the exchange of L2 data from one attachment circuit (AC) to the other, and between two interfaces of the same type on the same router. The two ports configured in a local switching connection form an attachment circuit (AC). A local switching connection works like a bridge domain that has only two bridge ports, where traffic enters from one port of the local connection and leaves through the other.

These are some of the characteristics of Layer 2 local switching:

• Layer 2 local switching uses Layer 2 MAC addresses instead of the Layer 3 IP addresses.

• Because there is no bridging involved in a local connection, there is neither MAC learning nor flooding.

• Unlike in a bridge domain, the ACs in a local connection are not in the UP state if the interface state is DOWN.

• Local switching ACs utilize a full variety of Layer 2 interfaces, including Layer 2 trunk (main) interfaces, bundle interfaces, and EFPs.

• Same-port local switching allows you to switch Layer 2 data between two circuits on the same interface.

Restrictions

• All sub-interfaces under the given physical port support only two Tag Protocol Identifiers (TPIDs), such as:

  • 0x88a8, 0x8100

  • 0x9100, 0x8100

  • 0x9200, 0x8100

• VLAN and TPID-based ingress packet filtering is not supported.

• Egress TPID rewrite is not supported.

Topology

An Attachment Circuit (AC) binds a Customer Edge (CE) router to a Provider Edge (PE) router. The PE router uses a pseudowire over the MPLS network to exchange routes with a remote PE router. To establish a point-to-point connection in a Layer 2 VPN from one Customer Edge (CE) router to another (remote router), a mechanism is required to bind the attachment circuit to the pseudowire. A Cross-Connect Circuit (CCC) is used to bind attachment circuits to pseudowires to emulate a point-to-point connection in a Layer 2 VPN.

The following topology is used for configuration.

Configuration

To configure an AC-AC local switching, complete the following configuration:

• Enable Layer 2 transport on main interfaces.

• Create sub-interfaces with Layer 2 transport enabled, and specify the respective encapsulation for each.

• Enable local switching between the main interfaces, and between the sub-interfaces.

  • Create a cross-connect group.

  • Create a point-to-point cross connect circuit (CCC).

  • Assign interface(s) to the point-to-point cross connect group.

/* Enter the interface configuration mode and configure 
  L2 transport on the TenGigE interfaces */
Router# configure
Router(config)# interface TenGigE 0/0/0/1 l2transport
Router(config-if-l2)# no shutdown
Router(config-if)# exit
Router(config)# interface TenGigE 0/0/0/9 l2transport
Router(config-if-l2)# no shutdown
Router(config-if-l2)# commit

/* Configure L2 transport and encapsulation on the VLAN sub-interfaces */
Router# configure
Router(config)# interface TenGigE 0/0/0/0.1 l2transport
Router(config-subif)# encapsulation dot1q 5 
Router(config-subif)# exit
Router(config)# interface TenGigE 0/0/0/8.1 l2transport
Router(config-subif)# encapsulation dot1q 5 
Router(config-subif)# commit


/* Configure ethernet link bundles */
Router# configure
Router(config)# interface Bundle-Ether 3 
Router(config-if)# ipv4 address 10.1.3.3 255.0.0.0
Router(config-if)# bundle maximum-active links 32 hot-standby
Router(config-if)# bundle minimum-active links 1
Router(config-if)# bundle minimum-active bandwidth 30000000
Router(config-if)# exit

Router(config)# interface Bundle-Ether 2 
Router(config-if)# ipv4 address 10.1.2.2 255.0.0.0
Router(config-if)# bundle maximum-active links 32 hot-standby
Router(config-if)# bundle minimum-active links 1
Router(config-if)# bundle minimum-active bandwidth 30000000
Router(config-if)# exit

/* Add physical interfaces to the ethernet link bundles */
Router(config)# interface TenGigE 0/0/0/1
Router(config-if)# bundle id 3 mode on
Router(config-if)# no shutdown
Router(config)# exit
Router(config)# interface TenGigE 0/0/0/2
Router(config-if)# bundle id 3 mode on
Router(config-if)# no shutdown
Router(config)# exit
Router(config)# interface TenGigE 0/0/0/9 
Router(config-if)# bundle id 2 mode on
Router(config-if)# no shutdown
Router(config-if)# exit
Router(config)# interface TenGigE 0/0/0/8 
Router(config-if)# bundle id 2 mode on
Router(config-if)# no shutdown
Router(config-if)# exit

/* Configure Layer 2 transport on the ethernet link bundles */
Router(config)# interface Bundle-Ether 3 l2transport
Router(config-if-l2)# no shutdown
Router(config-if)# exit
Router(config)# interface Bundle-Ether 2 l2transport
Router(config-if-l2)# no shutdown
Router(config-if-l2)# commit

/* Configure local switching on the TenGigE Interfaces */
Router(config)# l2vpn
Router(config-l2vpn)# xconnect group XCON1
Router(config-l2vpn-xc)# p2p XCON1_P2P3
Router(config-l2vpn-xc-p2p)# interface TenGigE0/0/0/1
Router(config-l2vpn-xc-p2p)# interface TenGigE0/0/0/9
Router(config-l2vpn-xc-p2p)# commit
Router(config-l2vpn-xc-p2p)# exit

/* Configure local switching on the VLAN sub-interfaces */
Router(config-l2vpn-xc)# p2p XCON1_P2P1
Router(config-l2vpn-xc-p2p)# interface TenGigE0/0/0/0.1
Router(config-l2vpn-xc-p2p)# interface TenGigE0/0/0/8.1
Router(config-l2vpn-xc-p2p)# commit
Router(config-l2vpn-xc-p2p)# exit

/* Configure local switching on ethernet link bundles */
Router(config-l2vpn-xc)# p2p XCON1_P2P4
Router(config-l2vpn-xc-p2p)# interface Bundle-Ether 3
Router(config-l2vpn-xc-p2p)# interface Bundle-Ether 2
Router(config-l2vpn-xc-p2p)# commit

Running Configuration

configure
 interface tenGigE 0/0/0/1 l2transport
 !
 interface tenGigE 0/0/0/9 l2transport
 !
!

interface tenGigE 0/0/0/0.1 l2transport
 encapsulation dot1q 5 
 rewrite ingress tag push dot1q 20 symmetric
 !
interface tenGigE 0/0/0/8.1 l2transport
  encapsulation dot1q 5 
 !
interface Bundle-Ether 3 l2transport
 !
interface Bundle-Ether 2 l2transport
!

l2vpn
 xconnect group XCON1
   p2p XCON1_P2P3
    interface TenGigE0/0/0/1
    interface TenGigE0/0/0/9
    !
   !
 !
l2vpn
 xconnect group XCON1
   p2p XCON1_P2P1
    interface TenGigE0/0/0/0.1
    interface TenGigE0/0/0/8.1
    !
   !
 !
l2vpn
 xconnect group XCON1
   p2p XCON1_P2P4
    interface Bundle-Ether 3
    interface Bundle-Ether 2
    !
   !
 !

Verification

• Verify if the configured cross-connect is UP

  
  router# show l2vpn xconnect brief
  
  Locally Switching
  
    Like-to-Like                        UP       DOWN        UNR
  
      EFP                                1          0          0
  
      Total                              1          0          0
  
  
    Total                                1          0          0
  
  Total: 1 UP, 0 DOWN, 0 UNRESOLVED
  
  

  
  router# show l2vpn xconnect
  
  Legend: ST = State, UP = Up, DN = Down, AD = Admin Down, UR = Unresolved,
          SB = Standby, SR = Standby Ready, (PP) = Partially Programmed
  
  XConnect                   Segment 1                       Segment 2
  Group      Name       ST   Description            ST       Description       ST
  ------------------------   -------------------------------------------------------------
  XCON1      XCON_P2P1 UP   Te0/0/0/1      UP  Te0/0/0/9      UP
  XCON1      XCON_P2P3 UP   Te0/0/0/0.1    UP  Te0/0/0/8.1    UP
  ----------------------------------------------------------------------------------------
  
  

Associated Commands

• interface (p2p)

• l2vpn

• p2p

• xconnect group

Split Horizon Group 2

The Split Horizon Group 2 feature allows you to prevent BUM and known unicast traffic to be flooded from one AC to other AC within the bridge domain. This feature enables efficient bandwidth allocation and resource optimization.

Consider the following topology in which AC1 and AC2 are part of the same VPLS bridge domain. When you configure split horizon group 2 over AC1, AC2 on PE3, BUM and known unicast traffic from AC1 is not flooded to AC2 and vice-versa.

However, BUM traffic coming from the pseduowire on PE3 to AC1 and AC2 that are part of group 2 is flooded. The known unicast traffic is sent to the corresponding AC.

If AC1 is part of group 0 and AC2 is part of group 2, BUM and known unicast traffic is flooded between AC1 and AC2. Similarly, if AC2 is part of group 0 and AC1 is part of group 2, BUM and known unicast traffic is flooded between AC1 and AC2.

Overview

Each Ethernet ring node is connected to adjacent Ethernet ring nodes participating in the Ethernet ring using two independent links. A ring link never allows formation of loops that affect the network. The Ethernet ring uses a specific link to protect the entire Ethernet ring. This specific link is called the ring protection link (RPL). A ring link is bound by two adjacent Ethernet ring nodes and a port for a ring link (also known as a ring port).

Note	The minimum number of Ethernet ring nodes in an Ethernet ring is two.

The fundamentals of ring protection switching are:

• The principle of loop avoidance.

• The utilization of learning, forwarding, and Filtering Database (FDB) mechanisms.

Loop avoidance in an Ethernet ring is achieved by ensuring that, at any time, traffic flows on all but one of the ring links which is the RPL. Multiple nodes are used to form a ring:

• RPL owner—It is responsible for blocking traffic over the RPL so that no loops are formed in the Ethernet traffic. There can be only one RPL owner in a ring.

• RPL neighbor node—The RPL neighbor node is an Ethernet ring node adjacent to the RPL. It is responsible for blocking its end of the RPL under normal conditions. This node type is optional and prevents RPL usage when protected.

• RPL next-neighbor node—The RPL next-neighbor node is an Ethernet ring node adjacent to RPL owner node or RPL neighbor node. It is mainly used for FDB flush optimization on the ring. This node is also optional.

The following figure illustrates the G.8032 Ethernet ring.

Nodes on the ring use control messages called RAPS to coordinate the activities of switching on or off the RPL link. Any failure along the ring triggers a RAPS signal fail (RAPS SF) message along both directions, from the nodes adjacent to the failed link, after the nodes have blocked the port facing the failed link. On obtaining this message, the RPL owner unblocks the RPL port.

Note	A single link failure in the ring ensures a loop-free topology.

Line status and Connectivity Fault Management protocols are used to detect ring link and node failure. During the recovery phase, when the failed link is restored, the nodes adjacent to the restored link send RAPS no request (RAPS NR) messages. On obtaining this message, the RPL owner blocks the RPL port and sends RAPS no request, root blocked (RAPS NR, RB) messages. This causes all other nodes, other than the RPL owner in the ring, to unblock all blocked ports. The ERP protocol is robust enough to work for both unidirectional failure and multiple link failure scenarios in a ring topology.

A G.8032 ring supports these basic operator administrative commands:

• Force switch (FS)—Allows operator to forcefully block a particular ring-port.

  • Effective even if there is an existing SF condition

  • Multiple FS commands for ring supported

  • May be used to allow immediate maintenance operations

• Manual switch (MS)—Allows operator to manually block a particular ring-port.

  • Ineffective in an existing FS or SF condition

  • Overridden by new FS or SF conditions

  • Clears all previous MS commands

• Clear—Cancels an existing FS or MS command on the ring-port

  • Used (at RPL Owner) to clear non-revertive mode

A G.8032 ring can support two instances. An instance is a logical ring running over a physical ring. Such instances are used for various reasons, such as load balancing VLANs over a ring. For example, odd VLANs may go in one direction of the ring, and even VLANs may go in the other direction. Specific VLANs can be configured under only one instance. They cannot overlap multiple instances. Otherwise, data traffic or RAPS packet can cross logical rings, and that is not desirable.

Timers

G.8032 ERP specifies the use of different timers to avoid race conditions and unnecessary switching operations:

• Delay Timers—used by the RPL Owner to verify that the network has stabilized before blocking the RPL

  • After SF condition, Wait-to-Restore (WTR) timer is used to verify that SF is not intermittent. The WTR timer can be configured by the operator, and the default time interval is 5 minutes. The time interval ranges from 1 to 12 minutes.

  • After FS/MS command, Wait-to-Block timer is used to verify that no background condition exists.

  Note	Wait-to-Block timer may be shorter than the Wait-to-Restore timer

• Guard Timer—used by all nodes when changing state; it blocks latent outdated messages from causing unnecessary state changes. The Guard timer can be configured and the default time interval is 500 ms. The time interval ranges from 10 to 2000 ms.

• Hold-off timers—used by underlying Ethernet layer to filter out intermittent link faults. The hold-off timer can be configured and the default time interval is 0 seconds. The time interval ranges from 0 to 10 seconds.

  • Faults are reported to the ring protection mechanism, only if this timer expires.

Single Link Failure

The following figure represents protection switching in case of a single link failure.

The above figure represents an Ethernet ring composed of seven Ethernet ring nodes. The RPL is the ring link between Ethernet ring nodes A and G. In these scenarios, both ends of the RPL are blocked. Ethernet ring node G is the RPL owner node, and Ethernet ring node A is the RPL neighbor node.

These symbols are used:

This sequence describes the steps in the single link failure:

1. Link operates in the normal condition.

2. A failure occurs.

3. Ethernet ring nodes C and D detect a local Signal Failure condition and after the holdoff time interval, block the failed ring port and perform the FDB flush.

4. Ethernet ring nodes C and D start sending RAPS (SF) messages periodically along with the (Node ID, BPR) pair on both ring ports, while the SF condition persists.

5. All Ethernet ring nodes receiving an RAPS (SF) message perform FDB flush. When the RPL owner node G and RPL neighbor node A receive an RAPS (SF) message, the Ethernet ring node unblocks it’s end of the RPL and performs the FDB flush.

6. All Ethernet ring nodes receiving a second RAPS (SF) message perform the FDB flush again; this is because of the Node ID and BPR-based mechanism.

7. Stable SF condition—RAPS (SF) messages on the Ethernet Ring. Further RAPS (SF) messages trigger no further action.

The following figure represents reversion in case of a single link failure.

This sequence describes the steps in the single link failure recovery:

1. Link operates in the stable SF condition.

2. Recovery of link failure occurs.

3. Ethernet ring nodes C and D detect clearing of signal failure (SF) condition, start the guard timer and initiate periodical transmission of RAPS (NR) messages on both ring ports. (The guard timer prevents the reception of RAPS messages).

4. When the Ethernet ring nodes receive an RAPS (NR) message, the Node ID and BPR pair of a receiving ring port is deleted and the RPL owner node starts the WTR timer.

5. When the guard timer expires on Ethernet ring nodes C and D, they may accept the new RAPS messages that they receive. Ethernet ring node D receives an RAPS (NR) message with higher Node ID from Ethernet ring node C, and unblocks its non-failed ring port.

6. When WTR timer expires, the RPL owner node blocks its end of the RPL, sends RAPS (NR, RB) message with the (Node ID, BPR) pair, and performs the FDB flush.

7. When Ethernet ring node C receives an RAPS (NR, RB) message, it removes the block on its blocked ring ports, and stops sending RAPS (NR) messages. On the other hand, when the RPL neighbor node A receives an RAPS (NR, RB) message, it blocks its end of the RPL. In addition to this, Ethernet ring nodes A to F perform the FDB flush when receiving an RAPS (NR, RB) message, due to the existence of the Node ID and BPR based mechanism.

Forcing a Manual Switchover to the Backup Pseudowire

To force the router to switch over to the backup or switch back to the primary pseudowire, use the l2vpn switchover command in EXEC mode.

A manual switchover is made only if the peer specified in the command is actually available and the cross-connect moves to the fully active state when the command is entered.

Configuration

This section describes the configuration for pseudowire redundancy.

/* Configure a cross-connect group with a static point-to-point 
cross connect */
Router# configure
Router(config)# l2vpn
Router(config-l2vpn)# xconnect group A
Router(config-l2vpn-xc)# p2p xc1
ROuter(config-l2vpn-xc-p2p)# interface tengige 0/0/0/0.2
Router(config-l2vpn-xc-p2p)# neighbor 10.1.1.2 pw-id 2

/*Configure the pseudowire segment for the cross-connect group */
Router(config-l2vpn-xc-p2p-pw)#pw-class path1

/*Configure the backup pseudowire segment for the cross-connect group */
Router(config-l2vpn-xc-p2p-pw)# backup neighbor 10.2.2.2 pw-id 5
Router(config-l2vpn-xc-p2p-pw-backup)#end

/*Commit your configuration */
Router(config-l2vpn-xc-p2p-pw-backup)#commit
Uncommitted changes found, commit them before exiting(yes/no/cancel)?
[cancel]: yes

Running Configuration

Router# show-running configuration
...
l2vpn
 encapsulation mpls
 !
 xconnect group A
  p2p xc1
   interface tengige 0/0/0/0.2
   neighbor ipv4 10.1.1.2 pw-id 2
    pw-class path1
    backup neighbor 10.2.2.2 pw-id 5
    !
   !
...