A Layer 2 Virtual Private Network (VPN) emulates a physical sub-network in an IP or MPLS network, by creating private connections between two points. Building a L2VPN network requires coordination between the service provider and customer. The service provider establishes Layer 2 connectivity. The customer builds a network by using the data link resources obtained from the service provider. In a L2VPN service, the service provider does not require information about the customer's network topology and other information. This helps maintain customer privacy, while using the service provider resources to establish the network.

The service provider requires Provider Edge (PE) routers with the following capabilities:

• Encapsulation of L2 protocol data units (PDU) into Layer 3 (L3) packets.
• Interconnection of any-to-any L2 transports.
• Support for MPLS tunneling mechanism.
• Process databases that include all information related to circuits and their connections.

This section introduces Layer 2 Virtual Private Networks (VPNs) and the corresponding Gigabit Ethernet services.

A L2VPN network enables service providers (SPs) to provide L2 services to geographically disparate customer sites. Typically, a SP uses an access network to connect the customer to the core network. This access network may use a mixture of L2 technologies, such as Ethernet and Frame Relay. The connection between the customer site and the nearby SP edge router is known as an attachment circuit (AC). Traffic from the customer travels over this link to the edge of the SP core network. The traffic then tunnels through a pseudowire over the SP core network to another edge router. The edge router sends the traffic down another AC to the customer's remote site.

The L2VPN feature enables the connection between different types of L2 attachment circuits and pseudowires, allowing users to implement different types of end-to-end services.

Note	BOOTP traffic (dst UDP 68) over any type of pseudowire is unsupported.

Cisco IOS XR software supports a point-to-point end-to-end service, where two Ethernet circuits are connected together. An L2VPN Ethernet port can operate in one of two modes:

• Port Mode—In this mode, all packets reaching the port are sent over the pseudowire, regardless of any VLAN tags that are present on the packets. In Port mode, the configuration is performed under the l2transport configuration mode.

• VLAN Mode—Each VLAN on a CE (customer edge) or access network to PE (provider edge) link can be configured as a separate L2VPN connection (using either VC type 4 or VC type 5). To configure L2VPN on VLANs, see The Carrier Ethernet Model chapter in this manual. In VLAN mode, the configuration is performed under the individual sub-interface.

Switching can take place in the following ways:

• AC-to-PW—Traffic reaching the PE is tunneled over a PW (pseudowire) (and conversely, traffic arriving over the PW is sent out over the AC). This is the most common scenario.

• Local switching—Traffic arriving on one AC is immediately sent out of another AC without passing through a pseudowire.

• PW stitching—Traffic arriving on a PW is not sent to an AC, but is sent back into the core over another PW.

Note

If your network requires that packets are transported transparently, you may need to modify the packet’s destination MAC (Media Access Control) address at the edge of the Service Provider (SP) network. This prevents the packet from being consumed by the devices in the SP network. The encapsulation dot1ad vlan-id and encapsulation dot1ad vlan-id dot1q any commands cannot co-exist on the same physical interface or bundle interface. Similarly, the encapsulation dot1q vlan-id and encap dot1q vlan-id second-dot1q any commands cannot co-exist on the same physical interface or bundle interface. If there is a need to co-exist, it is recommended to use the exact keyword in the single tag encapsulation. For example, encap dot1ad vlan-id exact or encap dot1q vlan-id exact. In an interface which already has QinQ configuration, you cannot configure the QinQ Range sub-interface where outer VLAN range of QinQ Range overlaps with outer VLAN of QinQ. Attempting this configuration results in the splitting of the existing QinQ and QinQ Range interfaces. However, the system can be recovered by deleting a recently configured QinQ Range interface. In an interface which already has QinQ Range configuration, you cannot configure the QinQ Range sub-interface where outer VLAN range of QinQ Range overlaps with inner VLAN of QinQ Range. Attempting this configuration results in the splitting of the existing QinQ and QinQ Range interfaces. However, the system can be recovered by deleting a recently configured QinQ Range interface. The inner VLAN ranges of sub-interfaces configured cannot have overlapping values. In such overlapping inner VLAN range cases, the system can be recovered by reloading the LC on Cisco IOS XR Release 6.5.x.

You can use the show interfaces command to display AC and pseudowire information.

This section describes how you can configure Gigabit ethernet interfaces for Layer 2 transport.

RP/0/RP0/CPU0(config)#interface TenGigE 0/0/0/10

/* Configure the ethertype for the 802.1q encapsulation (optional) */
/* For VLANs, the default ethertype is 0x8100. In this example, we configure a value of 0x9100. 
/* The other assignable value is 0x9200 */
/* When ethertype is configured on a physical interface, it is applied to all sub-interfaces created on this interface */

RP/0/RP0/CPU0:router(config-if)#dot1q tunneling ethertype 0x9100 

/* Configure Layer 2 transport on the interface, and commit your configuration */
RP/0/RP0/CPU0:router(config-if)#l2transport 
RP/0/RP0/CPU0:router(config-if-l2)#commit 
Sat May  2 19:50:36.799 UTC
RP/0/RP0/CPU0:router(config-if-l2)#exit  
RP/0/RP0/CPU0:router(config-if)#no shutdown 
RP/0/RP0/CPU0:router(config-if)#exit
RP/0/RP0/CPU0:router(config)#

configure
 interface TenGigE 0/0/0/10
  dot1q tunneling ethertype 0x9100
  l2transport
 !

router# show interfaces TenGigE 0/0/0/10

...
TenGigE0/0/0/10 is up, line protocol is up 
  Interface state transitions: 1
  Hardware is TenGigE, address is 0011.1aac.a05a (bia 0011.1aac.a05a)
  Layer 1 Transport Mode is LAN
  Layer 2 Transport Mode
  MTU 1514 bytes, BW 10000000 Kbit (Max: 10000000 Kbit)
     reliability 255/255, txload 0/255, rxload 0/255
  Encapsulation ARPA,
  Full-duplex, 10000Mb/s, link type is force-up
  output flow control is off, input flow control is off
  Carrier delay (up) is 10 msec
  loopback not set,
  ...

The Ethernet Data Plane Loopback function allows you to run loopback tests to test the connectivity and quality of connections through a Layer 2 cloud. You can run this test on:

• Main interface or sub-interfaces

• Bundle or its sub-interfaces

• Multiple hops through the underlying network

You can use this feature to test the throughput of an Ethernet port remotely. You can verify the maximum rate of frame transmission with no frame loss.

This feature allows for bidirectional or unidirectional throughput measurement, and on-demand or out-of-service (intrusive) operation during service turn-up.

Two types of Ethernet loopback are supported:

• External loopback - Traffic loopback occurs at the Ingress interface. Traffic does not flow into the router for loopback.

• Internal loopback - Traffic loopback occurs at the Egress interface. Traffic loopback occurs after the traffic flows into the router to the other interface.

Ethernet data traffic can be looped back on per port basis. This feature supports a maximum of 100 concurrent Ethernet data plane loopback sessions per system. Filters based on frame header can be used for initiating the loopback session. This ensures that only a subset of traffic that is received on an interface is looped back. You can use Source MAC, Destination MAC, and VLAN Priority (COS bits) as filters.

• N540-28Z4C-SYS-A

• N540-28Z4C-SYS-D

• N540X-16Z4G8Q2C-A

• N540X-16Z4G8Q2C-D

• N540-12Z20G-SYS-A

• N540-12Z20G-SYS-D

• N540X-12Z16G-SYS-A

• N540X-12Z16G-SYS-D