The Cisco ASR 9000 Series Routers offer these functions:

• Tunnels these protocols:

  • Cisco Discovery Protocol (CDP)

  • Spanning Tree Protocol (STP and its derivatives)

  • Virtual Trunking Protocol (VTP)

• Supports these modes of tunneling

  • Forward

  • Reverse

• L2PT encapsulates and decapsulates protocol frames that have VLAN headers.

• Supports capability of handling enormous frame rates. The Cisco ASR 9000 Series Routers perform L2PT encapsulation and decapsulation at the interface line rates.

Note	There are no dedicated L2PT counters. There are no L2PT-specific adjustments for QoS or other miscellaneous parameters.

Figure below shows L2PT configured in the forward mode.

A Service Provider network (S-network) is depicted in Figure 1. The customer network (C-network) connects to router R1 at the GigabitEthernet subinterface 0/1/0/1.1, and to router R2 at the GigabitEthernet subinterface 0/5/0/2.1. The C-network is not shown in the diagram; however, the C-network sends L2 traffic through the S-network, and the S-network switches the traffic from end to end. The customer traffic also carries L2 protocol frames. The purpose of L2PT is to allow these protocol frames to pass through the S-network. In forward mode, L2PT is applied to the customer facing interfaces of the S-network, R1 GigabitEthernet 0/1/0/1.1 and R2 GigabitEthernet 0/5/0/2.1.

Figure above depicts the configuration for L2PT in forward mode: :

R1:

!
interface GigabitEthernet0/1/0/1
 negotiation auto
!
interface GigabitEthernet0/1/0/1.1 l2transport
 encapsulation default
 l2protocol cpsv tunnel
!
interface GigabitEthernet0/1/0/2
 negotiation auto
!
interface GigabitEthernet0/1/0/2.1 l2transport
 encapsulation default
!
l2vpn
 xconnect group examples
  p2p r1-connect
   interface GigabitEthernet0/1/0/1.1
   interface GigabitEthernet0/1/0/2.1
  !
 !
!

R2:

!
interface GigabitEthernet0/5/0/1
 negotiation auto
!
interface GigabitEthernet0/5/0/1.1 l2transport
 encapsulation default
!
interface GigabitEthernet0/5/0/2
 negotiation auto
!
interface GigabitEthernet0/5/0/2.1 l2transport
 encapsulation default
 l2protocol cpsv tunnel
!
l2vpn
 xconnect group examples
  p2p r2-connect
   interface GigabitEthernet0/5/0/1.1
   interface GigabitEthernet0/5/0/2.1
  !
 !
!

Protocol traffic enters router R1 at the GigabitEthernet subinterface 0/1/0/1.1. Router R1 detects the frames as protocol frames, and performs L2PT encapsulation at the customer facing interface. Inside R1, the local connection r1-connect connects R1's customer-facing and service provider-facing interfaces. The traffic then flows out of router R1 on GigabitEthernet subinterface 0/1/0/2.1 through several other service provider network routers or switches (switch cloud) into router R2 at GigabitEthernet subinterface 0/5/0/1.1. Router R2 connects the customer-facing and service provider-facing interfaces through a local connection r2-connect. Therefore, traffic is sent to the customer-facing interface GigabitEthernet 0/5/0/2.1. At this interface, an L2PT decapsulation occurs and the protocol traffic flows out of router R2 into the customer network.

Without L2PT being configured the customer protocol frames that are sent into R1 are terminated. The customer traffic can consist of a variety of traffic; the protocol frames comprise a small percentage of the overall traffic stream.

The Cisco ASR 9000 Series Routers can perform L2PT encapsulation and decapsulation on supported L2 protocol frames that have VLAN headers. The L2 protocol frames do not have VLAN headers. However, in a service provider (SP) network that transports customer protocol traffic from one customer campus to another, this capability can be put to use within the SP network.

Figure below shows L2PT configured in the reverse mode. Assume that the customer traffic that enters R1 is trunked, that is all traffic is tagged. The only untagged traffic is the protocol traffic, that comes from the customer network.

When L2PT is configured in the reverse mode, the L2PT encapsulation occurs when the frame exits the interface. Likewise, in reverse mode decapsulation is performed when the frame enters the interface. Therefore, the L2PT tunnel is formed between the service provider-facing interfaces, instead of the customer-facing interfaces.

In this example, once the protocol traffic enters router R1, a VLAN tag is added to it. Before the traffic is sent through the service provider network, a second VLAN tag is added (100). The Cisco ASR 9000 Series Routers perform the L2PT encapsulation on a double-tagged protocol frame.

The above figure above shows four customer-facing interfaces (R1: GigabitEthernet subinterface 0/1/0.1.1, GigabitEthernet subinterface 0/1/0/2.1 and R2: GigabitEthernet subinterface 0/5/0/5.1, GigabitEthernet subinterface 0/5/0/6.1) and two service provider-facing interfaces (R1: GigabitEthernet subinterface 0/1/0/3.1 and R2: GigabitEthernet subinterface 0/5/0/4.1).

Figure above depicts the configuration for L2PT in reverse mode:

At R1:

!
interface GigabitEthernet0/1/0/1
 negotiation auto
!
interface GigabitEthernet0/1/0/1.1 l2transport
 encapsulation untagged
 rewrite ingress tag push dot1q 100 symmetric
 ethernet egress-filter strict
!
interface GigabitEthernet0/1/0/2
 negotiation auto
!
interface GigabitEthernet0/1/0/2.1 l2transport
 encapsulation untagged
 rewrite ingress tag push dot1q 200 symmetric
 ethernet egress-filter strict
!
interface GigabitEthernet0/1/0/3
 negotiation auto
!
interface GigabitEthernet0/1/0/3.1 l2transport
 encapsulation dot1q 500
 rewrite ingress tag pop 1 symmetric
 l2protocol cpsv reverse-tunnel
 ethernet egress-filter strict
!
l2vpn
 bridge group examples
  bridge-domain r1-bridge
   interface GigabitEthernet0/1/0/1.1
   !
   interface GigabitEthernet0/1/0/2.1
!
   interface GigabitEthernet0/1/0/3.1
   !
  !
 !
!

At R2:

!
interface GigabitEthernet0/5/0/4
 negotiation auto
!
interface GigabitEthernet0/5/0/4.1 l2transport
 encapsulation dot1q 500
 rewrite ingress tag pop 1 symmetric
 l2protocol cpsv reverse-tunnel
 ethernet egress-filter strict
!
interface GigabitEthernet0/5/0/5
 negotiation auto
!
interface GigabitEthernet0/5/0/5.1 l2transport
 encapsulation untagged
 rewrite ingress tag push dot1q 100 symmetric
 ethernet egress-filter strict
!
interface GigabitEthernet0/5/0/6
 negotiation auto
!
interface GigabitEthernet0/5/0/6.1 l2transport
 encapsulation untagged
 rewrite ingress tag push dot1q 200 symmetric
 ethernet egress-filter strict
!
l2vpn
 bridge group examples
  bridge-domain r2-bridge
   interface GigabitEthernet0/5/0/4.1
   !
   interface GigabitEthernet0/5/0/5.1
   !
   interface GigabitEthernet0/5/0/6.1
   !
  !
 !
!

These assumptions are made:

• Customer traffic entering router R1 is trunked, that is all traffic is tagged. The only untagged traffic is the protocol traffic, which arrives from the customer network.

• The Customer-facing interfaces GigabitEthernet 0/1/0/1 at router R1 and Gigabit Ethernet 0/5/0/5 at router R2 belong to the same customer. Customer-facing interfaces GigabitEthernet 0/1/0/2 at router R1 and GigabitEthernet 0/5/0/6 at router R2 belong to a different customer.

• Traffic from different customers remain segregated.

• Only L2 protocol traffic is sent through the customer-facing interfaces.

• L2 protocol traffic entering the customer-facing interfaces is untagged.

• Traffic must be L2PT encapsulated to successfully pass through the switch cloud.

The purpose of this topology is that router R1 and R2 must receive customer protocol traffic from multiple customer interfaces, and multiplex the traffic across a single service provider interface and link. At the decapsulation end, the reverse is performed. Traffic entering router R1 on the GigabitEthernet subinterface 0/1/0/1.1 exits router R2 from the GigabitEthernet subinterface 0/5/0/5.1 only while traffic entering router R1 at GigabitEthernet subinterface 0/1/0/2.1 exits router R2 from GigabitEthernet subinterface 0/5/0/6.1 only.

A protocol frame entering router R1 on GigabitEthernet interface 0/1/0/1 travels through the network in this manner:

• The protocol frame is directed to GigabitEthernet subinterface 0/1/0/1.1, as the frame is untagged.

• The rewrite statement with GigabitEthernet subinterface 0/1/0/1.1 causes a tag of ID 100 to be added to the frame.

• The frame enters router R1’s bridge domain r1-bridge.

• The bridge (r1-bridge) floods the frame to all attachment circuits (AC) on the bridge domain, except the originating AC (split horizon AC).

• Ethernet egress filtering on GigabitEthernet subinterface 0/1/0/2.1 detects a tag ID mismatch, and drops the frame. In this way, the bridge domain’s flooded traffic is prevented from exiting other customer interfaces.

• A flooded copy of the frame is sent to GigabitEthernet subinterface 0/1/0/3.1.

• GigabitEthernet subinterface 0/1/0/3.1 adds a second tag.

• The frame receives an L2PT encapsulation by GigabitEthernet subinterface 0/1/0/3.1 before it leaves router R1 through the GigabitEthernet interface 0/1/0/3.

  Note	The frame is now double-tagged (100 inner, 500 outer) and has the L2PT MAC DA.

• The frame passes to router R2 GigabitEthernet interface 0/5/0/4 because of the L2PT encapsulation.

• The frame after having entered router R2 on GigabitEthernet interface 0/5/0/4 is directed to GigabitEthernet subinterface 0/5/0/4.1.

• On entering GigabitEthernet subinterface 0/5/0/4.1, an L2PT decapsulation operation is performed on the frame.

• The outer tag ID 500 is removed by GigabitEthernet subinterface 0/5/0/4.1

• Router R2’s bridge (r2-bridge) floods the frames to all ACs.

• Ethernet egress filtering drops the frames on all ACs except the AC through which the frame exits.

• As the frame exits router R2 from GigabitEthernet subinterface 0/5/0/5.1, the tag of ID 100 is removed.

• The frame that exits router R2 from GigabitEthernet interface 0/5/0/5 is identical to the original frame that entered router R1 through GigabitEthernet interface 0/1/0/1.

Keep these points in mind while configuring L2PT:

• The l2protocol command can be configured on either a main or L2 subinterface.

• The l2protocol command can be configured on physical or bundle interfaces.

• When the l2protocol and ethernet filtering commands are configured on the same interface, L2PT encapsulation occurs before ethernet filtering. This means that L2PT prevents the CDP, STP, and VTP protocol frames from being dropped by ethernet filtering.

• When L2PT is configured with other interface features, L2PT encapsulation occurs before the processing for other interface features.

• L2PT encapsulation and decapsulation is supported for untagged protocol frames, single-tagged, and double-tagged frames. Tag Ethertypes of 0x8100, 0x88A8, and 0x9100 are supported, however, 0x9200 is not.