Table Of Contents
Configuring 4-Port Gigabit Ethernet WAN Optical Services Modules
Supported Features
Saving your Configuration Before Upgrading from an OSM-4GE-WAN-GBIC to an OSM-2+4GE-WAN+
Gigabit Ethernet WAN Port Configuration
Basic Interface Configuration
Configuring Strict Priority Low Latency Queuing (LLQ) Support on the OSM-2+4GE-WAN+
Examples
Quality of Services
Advanced QinQ Service Mapping
QinQ Translation—Double Tag to Single Tag Translation
QinQ Transparent Tunneling—Double Tag to Double Tag Translation
Out-of-Range and Unspecified In-Range Packets
Per VLAN Load Balancing for Advanced QinQ Service Mapping
Configuring Advanced QinQ Service Mapping
Enabling IEEE 802.1Q-in-802.1Q Translation on a Gigabit Ethernet WAN Interface
Prerequisites
Restrictions
Examples
Enabling IEEE 802.1Q-in-802.1Q Translation on a QinQ Link Bundle
Prerequisites
Restrictions
Examples
Configuring the Service Provider Edge Router
Prerequisites
Examples
Configuring QinQ Translation—Double Tag to Single Tag Translation
Prerequisites
Restrictions
Examples
Configuring QinQ Transparent Tunneling—Double Tag to Double Tag Translation
Prerequisites
Restrictions
Examples
Configuring a Policy Map to Use the Inner COS Bits
Prerequisites
Restrictions
Examples
Disabling IEEE 802.1Q-in-802.1Q Mapping and Translation
Configuration Examples for Advanced QinQ Service Mapping
QinQ Translation Configuration Example—Two-Tag to One-Tag Translation
QinQ Transparent Tunneling Configuration Example
QinQ Translation Using Port-Channel Interfaces Example
Configuring 4-Port Gigabit Ethernet WAN Optical Services Modules
This chapter provides an overview of the features supported on the 4-port Gigabit Ethernet WAN Optical Services Modules (OSM-2+4GE-WAN+ and OSM-4GE-WAN-GBIC) supported on Cisco Catalyst 6500 series switches and Cisco 7600 series routers.
This chapter consists of these sections:
•
Supported Features
•
Saving your Configuration Before Upgrading from an OSM-4GE-WAN-GBIC to an OSM-2+4GE-WAN+
•
Gigabit Ethernet WAN Port Configuration
•
Quality of Services
•
Advanced QinQ Service Mapping
Supported Features
The following Layer 3 features are supported on the Gigabit Ethernet WAN optical services modules (OSMs):
•
Forwarding of distributed IP services
•
Multiprotocol Label Switching (MPLS)
•
Ethernet over Multiprotocol Label Switching (EoMPLS)
•
Frame Relay over MPLS
•
ATM cell relay over MPLS VC-Mode
•
ATM AAL5 over MPLS
•
IOS Modular QoS Command Line Interface (MQC) QoS
•
Flow control
•
802.1Q VLAN trunking
•
Advanced 802.1Q-to-802.1Q (QinQ) Service Mapping
•
Hot Standby Routing Protocol (HSRP)
•
Jumbo frames
•
Support for up to 32,000 MAC addresses per port
•
Support for up to 32,000 simultaneous ACL entries
•
Support for up to 32,000 simultaneous QoS entries
•
SNMP I and II
•
Four RMON groups per port: statistics, history, alarms, and events
•
Online insertion and removal (OIR)
•
Inter-Switch Link (ISL)
Note
The OSM-2+4GE-WAN+ module supports ISL on the Layer 2 Gigabit Ethernet LAN ports but does not support ISL on the Layer 3 Gigabit Ethernet WAN ports.
The Layer 2 Gigabit Ethernet ports on the OSMs are configured from the supervisor engine of the Cisco Catalyst 6500 series switch or the Cisco 7600 series router. For feature support and configuration information for the OSM Layer 2 Gigabit Ethernet ports, refer to the links in the "Layer 2 Software Features" section on page 1-5.
Saving your Configuration Before Upgrading from an OSM-4GE-WAN-GBIC to an OSM-2+4GE-WAN+
When you upgrade from OSM-4GE-WAN-GBIC to an OSM-2+4GE-WAN+, the existing configuration will not be saved and applied to the new OSM-2+4GE-WAN+.
To save your configuration when upgrading from an OSM-4GE-WAN-GBIC to an OSM-2+4GE-WAN+, perform this task:
Step 1
Enter the write memory command before removing the OSM-4GE-WAN-GBIC.
Step 2
Install the new OSM-2+4GE-WAN+.
Step 3
Enter the copy startup-config running-config command.
Step 4
Enter the write memory command.
Warning
The orientation of the GBIC in OSM-4GE-WAN-GBIC ports is reversed (upside down) from those of the LAN ports for OSM-2+4GE-WAN+.
Gigabit Ethernet WAN Port Configuration
The four Gigabit Ethernet WAN ports on the 4-port Gigabit Ethernet WAN OSMs are controlled by Cisco IOS software and support all standard Cisco IOS features. For configuration information for standard Cisco IOS features and routing protocols supported on the GE-WAN ports, refer to the appropriate Cisco IOS configuration guide and command reference publications at this URL:
http://www.cisco.com/univercd/cc/td/doc/product/software/ios121/121cgcr/index.htm
Basic Interface Configuration
After you verify that the 4-port Gigabit Ethernet WAN OSM is installed correctly, use the configure command to configure the Gigabit Ethernet WAN interfaces.
The following procedure is for creating a basic configuration—enabling an interface and specifying IP routing. You might also need to enter other configuration subcommands, depending on the requirements for your system configuration.
Note
Subinterfaces on the 4-port Gigabit Etherent WAN module cannot share HSRP group numbers. As a result, only 16 HSRP groups per Gigabit Ethernet WAN port are supported.
Note
The MTU size you specify on a main Gigabit Ethernet WAN interface will also apply to all subinterfaces you configure on the main interface. It is not possible to specify an MTU size on a subinterface that is different from the MTU size specified for the main interface.
To configure the Gigabit Ethernet WAN interfaces, perform this task:
Step 1
Confirm that the system recognizes the module by entering the show version command:
Step 2
Check the status of each port by entering the show interface command:
Step 3
Enter configuration mode and specify that the console terminal will be the source of the configuration subcommands:
Router# configure terminal
Step 4
Enable IP routing by entering the ip routing command:
Router(config)# ip routing
Step 5
At the prompt, specify the new interface to configure by entering the interface command, followed by the type (ge-wan) and slot/port number. The example that follows is for a Gigabit Ethernet WAN OSM in slot 3:
Router(config)# interface ge-wan 3/0
Step 6
Assign an IP address and subnet mask to the interface with the ip address configuration subcommand, as in the following example:
Router(config-if)# ip address 10.1.2.3 255.255.255.255
By default, a GE-WAN interface is configured for automatic negotiation of link parameters, such as duplex, speed, and flow control. To disable flow control and to force the interface for 1000/full-duplex mode, turn off automatic negotiation with the command:
Router(config-if)# no negotiation auto
Note
Changing the negotiation mode of an active interface flaps the interface by bringing it down and then back up, so as to implement the new negotiation mode. For this reason, we recommend changing the negotiation mode only when the interface is shutdown.
Tip
Use the negotiation auto command to restore the default of automatic negotiation of link parameters.
Step 7
Change the shutdown state to up and enable the interface:
Router(config-if)# no shutdown
The no shutdown command passes an enable command to the Gigabit Ethernet module. It also causes the module to configure itself based on the most recent configuration commands received by the module.
Step 8
Write the new configuration to memory:
Router# copy running-config startup-config
When the configuration is stored, an OK message appears.
Configuring Strict Priority Low Latency Queuing (LLQ) Support on the OSM-2+4GE-WAN+
Starting with Cisco IOS Release 12.2(18)SXE, the Low Latency Queuing feature is changed for the OSM-2+4GE-WAN+ Optical Services Module. With this change, priority queue policing is supported on the module. Using Hiearchical Queuing Framework (HQF), the police command is combined with strict priority in a class on the OSM.
Note
The priority percent % and priority kbps commands from previous releases are no longer supported.
If a second priority police class is included in the policy, police must be configured first.
To configure strict priority LLQ support, perform the following tasks, starting in global configuration mode:
| |
Command or Action
|
Purpose
|
Step 1
|
Router(config)# policy-map policy-name
Example:
Router(config)# policy-map policy11
|
Specifies the name of the policy map to be created or modified.
|
Step 2
|
Router(config-pmap)# class class-name
Example:
Router(config)# class class204
|
Specifies the name of a predefined class included in the service policy.
|
Step 3
|
Router(config-pmap-c)# priority
Example:Router(config)# priority
|
Configures the strict priority class.
|
Step 4
|
Router(config-pmap-c)# police rate
Example:
Router(config-pmap-c) # police 1000000#
|
Sets the policing rate (in bps).
|
Examples
The following example shows a typical configuration and verification for the OSM-2+4GE-WAN+ OSM.
police cir 1000000 bc 31250 be 31250 conform-action transmit exceed-action drop
bandwidth remaining 35 (%)
bandwidth remaining 30 (%)
bandwidth remaining 25 (%)
shape average 6000000 24000 24000
service-policy output parent
ip address 10.10.10.1 255.255.255.0
The following show command verifies the configuration:
Router#show policy interface ge-wan7/1
GE-WAN7/1
Service-policy output: parent
Class-map: vlan-2 (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: vlan 2
Queueing
queue limit 1250 (packets)
(queue depth/total drops/no-buffer drops) 0/0/0
(pkts queued/bytes queued) 0/0
bandwidth 5000 kbps
shape (average) cir 6000000, bc 24000, be 24000
target shape rate 6000000
(shape parameter is rounded to 5952000 bps due to granularity)
Service-policy : child
Class-map: dscp-ef (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: ip dscp ef
Priority: b/w exceed drops: 0
police:
cir 1000000 bps, bc 31250 bytes
(Police cir is rounded to 983040 bps due to granularity)
Class-map: dscp-af21 (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: ip dscp af21
Queueing
queue limit 350 (packets)
(queue depth/total drops/no-buffer drops) 0/0/0
(pkts queued/bytes queued) 0/0
bandwidth remaining 35% (1400 kbps)
(bandwidth parameter is rounded to 1392 kbps due to granularity)
Class-map: dscp-af31 (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: ip dscp af31
Queueing
queue limit 300 (packets)
(queue depth/total drops/no-buffer drops) 0/0/0
(pkts queued/bytes queued) 0/0
bandwidth remaining 30% (1200 kbps)
(bandwidth parameter is rounded to 1196 kbps due to granularity)
Class-map: class-default (match-any)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: any
Queueing
queue limit 250 (packets)
(queue depth/total drops/no-buffer drops) 0/0/0
(pkts queued/bytes queued) 0/0
bandwidth remaining 25% (1000 kbps)
Class-map: class-default (match-any)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: any
queue limit 248750 (packets)
(queue depth/total drops/no-buffer drops) 0/0/0
(pkts queued/bytes queued) 0/0
Quality of Services
The Gigabit Ethernet WAN modules support the following QoS implementations:
•
Differentiated Services Code Point (DSCP) classification
•
IP-precedence classification
•
Class-based traffic shaping
•
Class-based weighted fair queuing (CBWFQ)—Supported on the OSM-2+4GE-WAN+ only
•
Low latency queuing (LLQ)—Supported on the OSM-2+4GE-WAN+ only
•
Weighted Random Early Detection (WRED)—Supported on the OSM-2+4GE-WAN+ only
•
Hierarchical traffic shaping for dot1q encapsulations—Supported for egress traffic on subinterfaces on the OSM-2+4GE-WAN+ only
•
EoMPLS Support with CBWFQ, LLQ and WRED - CBWFQ, LLQ or WRED are applied to the EoMPLS uplink interface. Supported on the OSM-2+4GE-WAN+ only
For QoS configuration information and examples for the WAN OSM ports, see the "Configuring QoS on the OSMs" section on page 9-2.
See Chapter 10, "Configuring Destination Sensitive Services on the Optical Services Modules" for configuration information.
Advanced QinQ Service Mapping
The IEEE 802.1Q VLAN specification provides for a trunking option that tags packets with two VLAN tags:
•
An inner tag that specifies the customer tag
•
An outer tag that specifies the service provider tag—to allow multiple VLANs to be trunked together across an intermediate network.
This type of double-tagged tunnel is referred to as IEEE 802.1Q-in-802.1Q (Q-in-Q) tunneling.
Standard QinQ tunneling, however, is limited. Although double-tagged VLANs can identify different customers, they cannot easily distinguish different service flows for the same customer. You can use separate VLANs for each service flow, but IEEE 802.1Q VLANs are limited to a maximum of 1,024 VLANs. Extended VLANs have a maximum of 4,096 per router, but even this larger number could be exhausted if many customers are using multiple services.
The Advanced QinQ Service Mapping feature solves these problems by enabling the Gigabit Ethernet WAN (GE-WAN) interfaces on the OSM-2+4GE-WAN+ Optical Services Module (OSM) to act as a QinQ access gateway. The access gateway enhances QinQ tunneling by using the combination of inner and outer VLAN tags as a unique identifier for a particular customer's service flows. This allows the interface to perform the following:
•
Translates packets that are tagged with an inner CE VLAN tag and an outer PE VLAN tag to a specifying outgoing trunk VLAN on the basis of the unique combination of CE and PE VLAN tags. Two types of packet translation are supported:
–
QinQ Translation (also known as double-tag to single-tag translation)—The CE and PE tags from the original incoming packet are replaced with a single trunk VLAN tag when the outgoing packet is transmitted.
–
QinQ Transparent Tunneling (also known as double-tag to double-tag translation)—The outer PE tag from the original incoming packet is replaced with an outer trunk VLAN tag when the outgoing packet is transmitted. The inner CE VLAN tag is left unchanged in the outgoing packet.
•
Supports traffic shaping on the basis of the unique combination of CE and PE VLAN tags.
•
Sets the IEEE 802.1P prioritization bits (P bits) in the outgoing trunk VLAN tag by copying the P bits either from the original packet's outer PE VLAN tag or from the original packet's inner CE VLAN tag.
In Cisco IOS Release 12.2(18)SXE and later releases, you can also combine multiple GE-WAN interfaces into a virtual QinQ link bundle (also known as a port-channel). This simplifies configuration and allows the system to automatically load balance the traffic moving across the physical interfaces.
See the following sections for more details on the QinQ translation process and on using QinQ link bundles.
QinQ Translation—Double Tag to Single Tag Translation
In a double-tag-to-single-tag translation, the Advanced QinQ Service Mapping feature replaces both the inner customer edge (CE) VLAN tag and the outer provider edge (PE) VLAN tag with a single trunk VLAN tag. The following shows the format of both the incoming original packet and the outgoing translated packet.
Original Incoming Packet
|
DA
|
SA
|
ETYPE= 0x8100
|
PE VLAN Tag
|
|
|
Data
|
FCS
|
Outgoing Translated Packet
|
|
DA
|
SA
|
ETYPE= 0x8100
|
Trunk VLAN Tag
|
Data
|
FCS
|
|
|
When the interface receives a packet, the following occurs:
•
Examines the inner CE VLAN tag and outer PE VLAN tag, and uses that unique combination to perform the quality of service processing, rate shaping, and switching that is specified by the attached service policy map.
If the packet includes a PE VLAN tag, but no mapping has been configured for this particular CE VLAN tag, or if the incoming packet does not contain any inner CE VLAN tag, the interface drops the packet (unless a subinterface has been configured for out-of-range packets).
•
Removes the inner and outer VLAN tags and replaces them with the trunk VLAN tag that has been configured on the VLAN's subinterface.
•
Sets the 802.1P bits (P bits) on the trunk VLAN tag in one of the following ways, depending on the service policy map being used:
–
Copies the P bits that were in the outer PE VLAN tag to the trunk VLAN tag (default).
–
Copies the P bits that were in the inner CE VLAN tag to the trunk VLAN tag (if the set cos cos-inner command was used in the service policy map).
–
Zeroes out the P bits if the interface or subinterface has been marked as untrusted.
•
Forwards the translated single-tagged packet to the appropriate destination or service.
QinQ Transparent Tunneling—Double Tag to Double Tag Translation
When you configure the Advanced QinQ Service Mapping feature for double-tag-to-double-tag conversion, the Gigabit Ethernet WAN interface replaces the outer PE VLAN tag with the trunk VLAN tag. The inner CE VLAN tag remains unchanged. The following shows the format of both the incoming original packet and the outgoing translated packet:
Original Incoming Packet
|
DA
|
SA
|
ETYPE= 0x8100
|
PE VLAN Tag
|
|
|
Data
|
FCS
|
Outgoing Translated Packet
|
DA
|
SA
|
ETYPE= 0x8100
|
Trunk VLAN Tag
|
ETYPE= 0x8100
|
CE VLAN Tag
|
Data
|
FCS
|
When the interface receives a packet, the following occurs:
•
Examines the inner CE VLAN tag and outer PE VLAN tag, and uses that unique combination to perform the quality of service processing, rate shaping, and switching that is specified by the attached service policy map.
If the packet includes a PE VLAN tag, but no mapping has been configured for this particular CE VLAN tag, or if the incoming packet does not contain any inner CE VLAN tag, the interface drops the packet (unless a subinterface has been configured for out-of-range packets).
•
Removes the outer PE VLAN tag and replaces it with the trunk VLAN tag that is configured on the VLAN's subinterface. The inner CE VLAN tag is left unchanged.
•
Sets the 802.1P bits (P bits) on the trunk VLAN tag in one of the following ways, depending on the service policy map being used:
–
Copies the P bits that were in the outer PE VLAN tag to the trunk VLAN tag (default).
–
Copies the P bits that were in the inner CE VLAN tag to the trunk VLAN tag (if the set cos cos-inner command was used in the service policy map).
–
Zeroes out the P bits if the interface or subinterface has been marked as untrusted.
•
Forwards the translated double-tagged packet to the appropriate destination or service.
Out-of-Range and Unspecified In-Range Packets
Each PE VLAN supports a maximum of 32 CE VLANs, which must be in a contiguous block that starts on a number divisible by 32 (for example: 0, 32, 64, and so on). When you specify the first CE VLAN ID for a PE VLAN (using the bridge-domain command), the Cisco IOS software automatically associates the corresponding block of 32 IDs with that PE VLAN. Any other CE VLANs are considered out-of-range for that particular PE VLAN.
For example, specifying a CE VLAN ID of 131 automatically associates the CE VLAN IDs from 128 to 159 with that particular PE VLAN. Any CE VLANs that are outside of that block (from 1 to 127 and from 160 to 4094) are considered out-of-range. In addition, if a packet arrives without a CE VLAN tag, it is also considered to be out-of-range.
The default behavior is to drop all out-of-range packets that are received on an interface that has been configured for QinQ translation. You can change this behavior by configuring a subinterface to match out-of-range packets.
The QinQ access gateway interface also drops any packets with a CE VLAN that is in-range (within the block of 32 VLAN IDs) but not explicitly mapped on a subinterface. This behavior cannot be changed. For example, if you specify a CE VLAN of 32 and no other CE VLANs for a particular PE VLAN, the interface drops packets for that PE VLAN that have CE VLANs from 33 and 63.
Per VLAN Load Balancing for Advanced QinQ Service Mapping
In Cisco IOS Release 12.2(18)SXE and later releases, you can combine multiple GE-WAN interfaces into a QinQ link bundle, which is a virtual interface that you configure in the same way as the physical GE-WAN interfaces. Using QinQ link bundles has the following advantages:
•
Simplifies configuration because you do not have to configure the individual GE-WAN physical interfaces. Instead, you configure only the one virtual interface with the required QinQ parameters, and those parameters are used for all of the physical interfaces in the bundle.
•
Increases bandwidth by allowing you to aggregate individual physical interfaces into a single logical interface.
•
Increases availability because if one link in the bundle goes down, the traffic is reallocated among the remaining interfaces until the link is reestablished.
•
Enables load-balancing of PE VLANs among the physical interfaces. When the PE VLANs are created, they are automatically distributed among the physical interfaces in the bundle in a round-robin fashion. Adding or removing a physical interface to the QinQ link bundle automatically reallocates the PE VLANs among the physical interfaces, with a minimal interruption of the traffic flows along those VLANs.
Note
The load-balancing algorithm is based only on the number of PE VLANs, where all of the packets for a particular PE VLAN are sent through the same physical interface. The load-balancing does not take into account the bandwidth or the number of the individual CE VLANs that are being transported in each PE tunnel. The assignment of a particular PE VLAN is determined when the PE VLAN is first created, and this assignment does not change unless interfaces are added or removed from the QinQ link bundle.
•
Allows you to logically group physical interfaces according to your management needs, such as application or location. You can obtain aggregate interface statistics by displaying the interface statistics for the bundle's virtual interface, as well as displaying the statistics for each of the individual physical interfaces in the bundle.
•
Simplifies network management by allowing you to perform OIR and other maintenance operations on interfaces and cards in the QinQ link bundle without stopping the traffic flows. Instead, the traffic is automatically redistributed among the remaining physical interfaces. When the card and its interfaces are brought back up, the traffic is again redistributed among all of the slots in the bundles.
•
Allows you to move OSM-2+4GE-WAN+ modules between slots without having to re-enter the complete interface configuration. Instead, you only have to remove the old interfaces from the QinQ link bundle and then add the new interfaces to the bundle. The bundle's configuration is then automatically applied to the card in its new location.
•
Requires a minimal learning curve to learn, because QinQ link bundles are created using the same port-channel and channel-group commands that are used on LAN interfaces to create Ether Channels. The same monitoring and maintenance procedures that are used for Ether Channels can be used for QinQ link bundles.
Configuring Advanced QinQ Service Mapping
This section describes the following configuration tasks that are needed to enable and configure the Advanced QinQ Service Mapping feature:
•
Enabling IEEE 802.1Q-in-802.1Q Translation on a Gigabit Ethernet WAN Interface
•
Enabling IEEE 802.1Q-in-802.1Q Translation on a QinQ Link Bundle
•
Configuring the Service Provider Edge Router
•
Configuring QinQ Translation—Double Tag to Single Tag Translation
•
Configuring QinQ Transparent Tunneling—Double Tag to Double Tag Translation
•
Configuring a Policy Map to Use the Inner COS Bits
•
Disabling IEEE 802.1Q-in-802.1Q Mapping and Translation
Enabling IEEE 802.1Q-in-802.1Q Translation on a Gigabit Ethernet WAN Interface
To use the Advanced QinQ Service Mapping feature, you must first enable IEEE 802.1Q-in-802.1Q translation on the Gigabit Ethernet WAN interface that is connected to the provider edge router through the Metro Ethernet network. You can also optionally configure the interface as trusted, if you want to preserve the IEEE 802.1P bits (P bits) that are in the IEEE 802.1Q header of incoming packets.
To enable IEEE 802.1Q-in-802.1Q translation on a Gigabit Ethernet WAN interface, and optionally configure the interface as trusted, use the following procedure.
Prerequisites
•
This feature requires a Cisco Catalyst 6500 series switch or Cisco 7600 series router with a Cisco Supervisor Engine 2 or Supervisor Engine 720 module that is running Cisco IOS Release 12.2(18)SXD or later.
•
This feature is supported only on the Gigabit Ethernet WAN (GE-WAN) interfaces on the OSM-2+4GE-WAN+ Gigabit Ethernet Enhanced Optical Services Module (OSM). This feature is not supported on other port adapter modules or on LAN Gigabit Ethernet (GE) interfaces.
•
The Cisco IOS software image must support the OSM-2+4GE-WAN+ Gigabit Ethernet Enhanced OSM card.
•
You must remove all IP, MPLS, and other Layer 3 configuration on the main interface before enabling IEEE 802.1Q-in-802.1Q translation.
Note
When a GE-WAN interface is configured for QinQ operation, the Cisco IOS command-line interface (CLI) blocks any IP configuration, but it is still possible to configure other Layer 3 features. All such configuration must be removed from the interface before QinQ can operate successfully.
Restrictions
•
This configuration is supported only on the Gigabit Ethernet WAN interfaces on the OSM-2+4GE-WAN+ enhanced Optical Services Module (OSM).
•
Only the main interface can be configured as an QinQ access gateway. Subinterfaces are then configured to specify the specific VLAN mappings.
•
A Gigabit Ethernet WAN interface that is configured as a QinQ access gateway cannot also be configured with any IP, MPLS, or other Layer 3 configurations. Adding such configuration to the interface can interfere with the QinQ operation.
•
Multiprotocol Label Switching (MPLS) Experimental (EXP) bit mappings and hierarchical QoS are not supported on the Gigabit Ethernet WAN interface being used for QinQ translation.
•
Each provider edge (PE) VLAN (or outer VLAN) supports a a maximum of 32 consecutive customer edge (CE) VLANs (or inner VLANs). This range of CE VLANs must start on a boundary that is divisible by 32 (for example, 32 to 63, 64 to 95, and so on, up to 4000 to 4031, 4032 to 4063, and 4064 to 4094). The invalid or reserved VLANs are excluded from this rule. For example, the first range is 2 to 31 because VLAN 0 is not valid and VLAN 1 is, by default, reserved for a native VLAN. Each PE VLAN also supports one default function that is applied to VLANs that fall outside of this range of 32 VLANs.
•
A PE VLAN can be configured on only one Gigabit Ethernet WAN interface in the router.
•
A PE VLAN cannot have the same ID as the native VLAN that is also being used on any interface in the router. The default native VLAN for interfaces is VLAN ID 1, and we recommend using this default to simplify the use of QinQ tagging on the router.
•
A PE VLAN cannot have the same ID as an MPLS-related VLAN ID being used on the same interface or on its paired interface. GE-WAN interfaces 1 and 2 constitute one pair, and GE-WAN interfaces 3 and 4 constitute another pair.
For example, if interface 1 assigns VLAN ID 200 to an MPLS-based feature (such as MPLS VPN, AToM, or VPLS), you cannot use VLAN 200 as a PE VLAN on either interface 1 or 2. However, you still can use VLAN 200 on interface 3 or 4, because those interfaces are a separate interface pair.
•
VLAN 4095 is reserved and cannot be used as a CE VLAN. Packets that contain a CE VLAN ID of 4095 are automatically dropped by subinterfaces that are configured for QinQ translation. VLAN 4095, however, can continue to be used as a native (non-QinQ) VLAN.
•
VLAN IDs from 1006 to 4094 can be used for either PE VLANs or internal VLANs. Since internal VLANs are automatically allocated for certain features such as Layer 3 LAN ports, WAN interfaces, and WAN subinterfaces, you must coordinate your use of PE VLANs with the system's use of internal VLANs. In particular, be sure to reserve some of the VLAN IDs between 1006 and 4094 for use as internal VLANS, because internal VLANs cannot use IDs between 1 and 1005. If you run out of VLANs for use as internal VLANs, you might not be able to install new cards or use certain software features.
The router, by default, allocates internal VLANs starting with 1006 and ascending sequentially. We recommend that you change this behavior with the vlan internal allocation policy descending global configuration command, so that the router allocates internal VLANs starting with 4094 and descending sequentially.
Note
If you change the allocation method, you must reboot the router before the changes take effect. This is because a number of internal VLANs are automatically allocated at router startup.
Tip
To display the number of internal VLANs that are currently in use, use the show vlan internal usage command.
•
The match vlan command is not supported in this feature.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
vlan internal allocation policy descending
4.
interface ge-wan slot/port
5.
no ip address
6.
mode dot1q-in-dot1q access-gateway
7.
description string
8.
no shutdown
9.
end
Tip
You do not need to configure the mls qos trust command to preserve the CoS bits in the VLAN translation, because this command has no effect on a GE-WAN interface that has been configured with the mode dot1q-in-dot1q access-gateway command. When an interface or port-channel group has been configured for QinQ translation, it always trusts the VLAN Class of Service (CoS) bits.
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
vlan internal allocation policy descending
Example:
Router(config)# vlan internal allocation policy
descending
Router(config)#
|
(Optional) Allocates internal VLANs starting with 4094 and descending sequentially. We recommend this configuration to avoid conflicts with the PE VLAN ID assignment.
Note If you change the allocation method, you must reboot the router before the changes take effect. This is because a number of internal VLANs are automatically allocated at router startup.
|
Step 4
|
interface ge-wan slot/port
Example:
Router(config)# interface ge-wan 5/1
Router(config-if)#
|
Enters interface configuration mode for the specified Gigabit Ethernet WAN interface on the OSM-2+4GE-WAN+ Gigabit Ethernet WAN port.
|
Step 5
|
no ip address
Example:
Router(config-if)# no ip address
Router(config-if)#
|
(Optional) Removes the IP address that might be configured on the interface. This step is required if the interface has been configured previously with an IP address.
|
Step 6
|
mode dot1q-in-dot1q access-gateway
Example:
Router(config-if)# mode dot1q-in-dot1q
access-gateway
Router(config-if)#
|
Enables IEEE 802.1Q-in-802.1Q translation on the interface, enabling the Advanced QinQ Service Mapping feature.
|
Step 7
|
description string
Example:
Router(config-if)# description Connected to ISP
ABC Port SJ-2
Router(config-if)#
|
(Optional) Provides a description of this interface. The string parameter can be any arbitrary text that describes the interface, its neighbor, its purpose, or any other information that might be useful for maintaining and troubleshooting problems with this interface and configuration.
|
Step 8
|
no shutdown
Example:
Router(config-if)# no shutdown
Router(config-if)#
|
Activates the interface and enables it to pass traffic.
|
| |
Note Repeat Step 4 through Step 8 for each Gigabit Ethernet WAN interface to be configured.
|
Step 9
|
end
Example:
Router(config-if)# end
Router#
|
Exits interface configuration mode and returns to privileged EXEC mode.
|
Examples
The following example shows a sample configuration for a Gigabit Ethernet WAN interface:
description connected to SJ QinQ Tunnel
logging event link-status
mode dot1q-in-dot1q access-gateway
Enabling IEEE 802.1Q-in-802.1Q Translation on a QinQ Link Bundle
To use the Advanced QinQ Service Mapping feature on a QinQ link bundle, you must create a virtual port-channel interface and enable IEEE 802.1Q-in-802.1Q translation on that interface. You then must assign Gigabit Ethernet WAN interfaces to the port-channel group. To perform these tasks, use the following procedure.
Prerequisites
•
The QinQ link bundle feature requires a Cisco Catalyst 6500 series switch or Cisco 7600 series router with a Cisco Supervisor Engine 2 or Supervisor Engine 720 module that is running Cisco IOS Release 12.2(18)SXE or later.
•
When using the QinQ link bundle feature, the port-channel group must include only Gigabit Ethernet WAN (GE-WAN) interfaces on the OSM-2+4GE-WAN+ Gigabit Ethernet Enhanced Optical Services Module (OSM). This feature is not supported on other port adapter modules or on LAN Gigabit Ethernet (GE) interfaces.
•
The Cisco IOS software image must support the OSM-2+4GE-WAN+ Gigabit Ethernet Enhanced OSM card.
Restrictions
•
All restrictions listed for the Gigabit Ethernet WAN interfaces also apply to the use of QinQ link bundling. See the "Restrictions" section for a list of those restrictions.
•
Channel groups that are being used for QinQ link bundling can contain only GE-WAN interfaces on the OSM-2+4GE-WAN+ Optical Services Module (OSM) card.
•
Port-channel interfaces that are being used for QinQ link bundling must not be configured for a Maximum Transmission Unit (MTU) value greater than 9170 bytes, which is the maximum MTU that is supported on the OSM-2+4GE-WAN+ OSM card.
•
Only the mode on option is supported when using the channel-group command with GE-WAN interfaces on the OSM-2+4GE-WAN+ Optical Services Module for advanced QinQ translation. The other mode options are not supported on a QinQ link bundle.
•
You cannot use the channel-group command on GE-WAN interfaces if Multiprotocol Label Switching (MPLS) is configured. You must remove all mpls configuration commands from the interface before using the channel-group command.
•
You cannot attach a service policy to the main port-channel interface or to the individual member interfaces of the port-channel group. Instead, you must attach the service policy to the appropriate port-channel subinterfaces. Also, input service policies are not supported on port-channels being used for QinQ link bundling.
•
Service policies for QinQ port-channel interfaces support only the shaping and set cos cos-inner commands. You cannot use other commands, such as the bandwidth command, on QinQ port-channel interfaces.
•
Port-channel interface counters can be displayed with the show interface port-channel {number | number.subif} command. However, the show interface port-channel counters and show counters interface port-channel commands are not supported for channel groups that are using GE-WAN interfaces for QinQ link bundling.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
vlan internal allocation policy descending
4.
interface port-channel number
5.
no ip address
6.
mode dot1q-in-dot1q access-gateway
7.
description string
8.
no shutdown
9.
interface ge-wan slot/port
10.
no ip address
11.
channel-group number mode on
12.
no shutdown
13.
end
Tip
You do not need to configure the mls qos trust command to preserve the CoS bits in the VLAN translation, because this command has no effect on a GE-WAN interface or port-channel group that has been configured with the mode dot1q-in-dot1q access-gateway command. When an interface or port-channel group has been configured for QinQ translation, it always trusts the VLAN Class of Service (CoS) bits.
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
vlan internal allocation policy descending
Example:
Router(config)# vlan internal allocation policy
descending
Router(config)#
|
(Optional) Allocates internal VLANs starting with 4094 and descending sequentially. We recommend this configuration to avoid conflicts with the assignment of IDs for the PE VLANs.
Note If you change the allocation method, you must reboot the router before the changes take effect. This is because a number of internal VLANs are automatically allocated at router startup.
|
Step 4
|
interface port-channel number
Example:
Router(config)# interface port-channel 5
Router(config-if)#
|
Creates a virtual port-channel interface and enters interface configuration mode. The valid range for number is from 1 to 256.
|
Step 5
|
no ip address
Example:
Router(config-if)# no ip address
Router(config-if)#
|
(Optional) Removes the IP address that might be configured on the interface. This step is required if the interface has been configured previously with an IP address.
|
Step 6
|
mode dot1q-in-dot1q access-gateway
Example:
Router(config-if)# mode dot1q-in-dot1q
access-gateway
Router(config-if)#
|
Enables IEEE 802.1Q-in-802.1Q translation on the interface, enabling the Advanced QinQ Service Mapping feature.
Note This command cannot be used on a port-channel that already contains a channel group member that is not a GE-WAN interface on a OSM-2+4GE-WAN+ card.
|
Step 7
|
description string
Example:
Router(config-if)# description QinQ Link Bundle
connected to LA-10/1
Router(config-if)#
|
(Optional) Provides a description of this interface. The string parameter can be any arbitrary text that describes the interface, its neighbor, its purpose, or any other information that might be useful for maintaining and troubleshooting problems with this interface and configuration.
|
Step 8
|
no shutdown
Example:
Router(config-if)# no shutdown
Router(config-if)#
|
Activates the interface and enables it to pass traffic.
|
Step 9
|
interface ge-wan slot/port
Example:
Router(config)# interface ge-wan 5/1
Router(config-if)#
|
Enters interface configuration mode for either the specified Gigabit Ethernet WAN interface on the OSM-2+4GE-WAN+ Gigabit Ethernet WAN port.
|
Step 10
|
no ip address
Example:
Router(config-if)# no ip address
Router(config-if)#
|
(Optional) Removes the IP address that might be configured on the interface. This step is required if the interface has been configured previously with an IP address.
|
Step 11
|
channel-group number mode on
Example:
Router(config-if)# channel-group 5 mode on
Router(config-if)#
|
Adds this physical interface to the specified channel group. The number should be the same as that specified for the port-channel interface in Step 4.
Note The mode on option is the only one allowed for port-channels that are being configured on GE-WAN interfaces for QinQ link bundling.
|
Step 12
|
no shutdown
Example:
Router(config-if)# no shutdown
Router(config-if)#
|
Activates the interface and enables it to pass traffic.
|
| |
Note Repeat Step 9 through Step 12 for each Gigabit Ethernet WAN interface to be added to the port-channel group.
|
Step 13
|
end
Example:
Router(config-if)# end
Router#
|
Exits interface configuration mode and returns to privileged EXEC mode.
|

Note
If after removing the last inner VLAN in a bridge domain, you want to perform a load rebalancing, issue the shutdown and no shutdown commands on the port-channel.
Examples
The following example shows a sample configuration for a port-channel interface that has two GE-WAN physical interfaces as part of its channel group:
logging event link-status
mode dot1q-in-dot1q access-gateway
logging event link-status
logging event link-status
The following sample configuration shows the error message that appears if you attempt to enable QinQ translation on a port-channel interface that contains one or more invalid interfaces:
Router# configure terminal
Router(config)# interface port-channel 30
7600-2(config-if)# mode dot1q-in-dot1q access-gateway
% 'mode dot1q-in-dot1q access-gateway' is not supported on Port-channel30
% Port-channel30 contains 2 Layer 2 Gigabit Ethernet interface(s)
To display the status of the port-channel interface, as well as the members of its channel group, use the show interface command. For example, this command would show the following output for the configuration listed above.
Router# show interface Port-channel 3
Port-channel1 is up, line protocol is up (connected)
Hardware is EtherChannel, address is 0007.8508.474a (bia 000d.edb5.7d7b)
MTU 1500 bytes, BW 2000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
input flow-control is off, output flow-control is unsupported
Members in this channel: GE2/1 Pseudo GE2/3 Pseudo
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output never, output hang never
Last clearing of "show interface" counters never
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Output queue: 0/40 (size/max)
5 minute input rate 0 bits/sec, 0 packets/sec
5 minute output rate 0 bits/sec, 0 packets/sec
0 packets input, 0 bytes, 0 no buffer
Received 0 broadcasts (0 IP multicast)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 0 multicast, 0 pause input
0 input packets with dribble condition detected
0 packets output, 0 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 PAUSE output
0 output buffer failures, 0 output buffers swapped out
To display the inner, outer, and trunk VLANs that are used in a QinQ translation, use the show cwan qinq command. The following examples show the resulting output for the port-channel interface using the show cwan qinq command alone and with each of the following optional keywords:
•
configured—Displays statistics for all configured bridge domains.
•
detail—Displays the details of the inner VLAN configurations for each bridge domain.
•
list—Displays the currently configured assignments.
Caution 
The
show cwan qinq [configured | detail | list] command applies to port-channel interfaces only. Using this command with physical interfaces may provide incorrect results.
Router#show cwan qinq
Bridge-domain Interface Egress-if Inner-start Total Active
3 Po1 GE3/1 0 1 1
Sub-Interface Trunk-vlan Inner-vlan Service State
Po1.2 2 4 dot1q up/up
Router#show cwan qinq configured
Port-channel1 has total 2 bridge-domain vlan(s)
Po1 - GE-WAN3/1 has 1 bridge-domain vlan(s) egress configured
13
Po1 - GE-WAN3/2 has 1 bridge-domain vlan(s) egress configured
3
Router#show cwan qinq detail
Port-channel1 has total 2 bridge-domain vlan(s)
Po1 - GE-WAN3/1 has 1 bridge-domain vlan(s) detail
Bridge-domain Inner Configured Active
------------- ------ ---------- ------
13 active 1 1
Po1 - GE-WAN3/2 has 1 bridge-domain vlan(s) detail
Bridge-domain Inner Configured Active
------------- ------ ---------- ------
3 active 1 1
Router#show cwan qinq list
Port-channel1 has total 2 bridge-domain vlan(s)
Po1 - GE-WAN3/1 has 1 bridge-domain vlan(s) egress active
13
Po1 - GE-WAN3/2 has 1 bridge-domain vlan(s) egress active
3
The related show cwan qinq load-balance commands also apply to port-channel interfaces only.
Router#show cwan qinq load-balance
Port-channel1 has total 2 bridge-domain vlan(s)
Po1 - GE-WAN3/1 has 1 bridge-domain vlan(s)
Po1 - GE-WAN3/2 has 1 bridge-domain vlan(s)
Router#show cwan qinq load-balance detail
Port-channel1 has total 2 bridge-domain vlan(s)
Po1 - GE-WAN3/1 has 1 bridge-domain vlan(s) detail
Bridge-domain Inner Configured Active
------------- ------ ---------- ------
13 active 1 1
Po1 - GE-WAN3/2 has 1 bridge-domain vlan(s) detail
Bridge-domain Inner Configured Active
------------- ------ ---------- ------
3 active 1 1
The following related show commands can be applied to both port-channel and physical interfaces:
Router#show cwan qinq bridge-domain
GE-WAN3/1, group 1, total_rate_active 1
13
GE-WAN3/2, group 1, total_rate_active 1
3
Port-channel1, group 1, total_rate_active 2
Router#show cwan qinq interface
Interface Status Egress op PE CE TRNK Input packets/ Output packets/
Input bytes Output bytes
---------------- --------- ------ -- ---- ---- ---- -------------------- ----------------
Po1.2 up/up GE3/2 1 3 4 2 0 0
0 0
Po1.12 up/up GE3/1 1 13 14 12 0 0
0 0
Note
For additional information regarding these related commands, see the Cisco 7600 Router Cisco IOS Command Reference—Release 12.2SX.
Configuring the Service Provider Edge Router
This section describes the procedure to configure the Gigabit Ethernet interface on the service provider edge router that is connected to the Gigabit Ethernet WAN interface that is acting as the IEEE 802.1Q-in-802.1Q (QinQ) access gateway.
Prerequisites
•
The service provider edge router must be using a Gigabit Ethernet interface.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
vlan vlan-id
4.
interface GigabigEthernet slot/port
5.
no ip address
6.
mls qos trust [cos | dscp | ip-precedence]
7.
switchport
8.
switchport trunk encapsulation dot1q
9.
switch trunk allowed vlan {vlan-list | vlan-range}
10.
switchport mode trunk
11.
description string
12.
no shutdown
13.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
vlan vlan-id
Example:
Router(config)# vlan 22
Router(config)#
|
Add the VLAN ID to be used as the provider edge (PE) VLAN to the router's VLAN database (if not already entered). The valid range for vlan-id is either 1 to 1023, or from 1 to 4094, depending on the Cisco IOS software image being used on the router or switch.
|
Step 4
|
interface GigabitEthernet slot/port
Example:
Router(config)# interface GigabitEthernet3/1
Router(config-if)#
|
Enters interface configuration mode for the specified Gigabit Ethernet interface.
|
Step 5
|
no ip address
Example:
Router(config-if)# no ip address
Router(config-if)#
|
Removes the IP address that might be configured on the interface.
|
Step 6
|
mls qos trust [cos | dscp | ip-precedence]
Example:
Router(config-if)# mls qos trust dscp
Router(config-if)#
|
(Optional) Specifies which quality of service (QoS) bits in incoming frames can be trusted.
• cos—(Optional) Specifies that the CoS bits in incoming frames are trusted and derives the internal DSCP value from the CoS bits.
• dscp—(Optional, default) Specifies that the ToS bits in the incoming packets contain a DSCP value.
• ip-precedence—(Optional) Specifies that the IP precedence bits (found in the ToS bits) of incoming packets are trusted, and derives the internal DSCP value from the IP precedence bits.
Note To configure the interface as untrusted, use the no mls qos trust command. The interface then zeroes out the P bits of all incoming packets.
|
Step 7
|
switchport
Example:
Router(config-if)# switchport
Router(config-if)#
|
Configures the interface for Layer 2 switching.
|
Step 8
|
switchport trunk encapsulation dot1q
Example:
Router(config-if)# switchport trunk
encapsulation dot1q
Router(config-if)#
|
Configures the trunk link to use IEEE 802.1Q encapsulation.
|
Step 9
|
switch trunk allowed vlan {vlan-list |
vlan-range}
Example:
Router(config-if)# switch trunk allowed vlan
3001-4000
Router(config-if)#
|
(Optional) Configures the list of provider edge (PE) VLANs allowed on the trunk. All VLANs are allowed by default. You can either specify a list of individual VLAN IDs separated by commas, or you can specify a range of VLAN IDs separated by a hyphen.
|
Step 10
|
switchport mode trunk
Example:
Router(config-if)# switchport mode trunk
Router(config-if)#
|
Puts the interface into permanent trunking mode.
|
Step 11
|
description string
Example:
Router(config-if)# description Connected to
Metro interface SJ-3
Router(config-if)#
|
(Optional) Provides a description of this interface. The string parameter can be any arbitrary text that describes the interface, its neighbor, its purpose, or any other information that might be useful for maintaining and troubleshooting problems with this interface and configuration.
|
Step 12
|
no shutdown
Example:
Router(config-if)# no shutdown
Router(config-if)#
|
Activates the interface and enables it to pass traffic.
|
| |
Note Repeat Step 4 through Step 12 for each interface to be configured.
|
Step 13
|
end
Example:
Router(config-if)# end
Router#
|
Exits interface configuration mode and returns to privileged EXEC mode.
|
Examples
The following example shows a sample configuration for a Gigabit Ethernet interface that is connected to the Gigabit Ethernet WAN port that is providing IEEE 802.1Q-in-802.1Q translation. VLAN ID 3001 is being used as the PE VLAN.
interface GigabitEthernet3/1
description connected to Metro SJ-3 (QinQ tunnel)
logging event link-status
switchport trunk encapsulation dot1q
switchport trunk allowed vlan 3001-4000
Configuring QinQ Translation—Double Tag to Single Tag Translation
When you configure the Advanced QinQ Service Mapping feature for QinQ translation, also known as double-tag-to-single-tag translation, the outgoing interface replaces both the inner customer edge (CE) VLAN tag and the outer provider edge (PE) VLAN tag with a Trunk VLAN tag. Use the following procedure to configure a subinterface for double-tag-to-single-tag translation.
Note
Cisco IOS Release 12.2(18)SXD used the bridge-vlan command to configure the QinQ translation, but Cisco IOS Release 12.2(18)SXE and later releases have changed this to bridge-domain. Earlier configurations that use bridge-vlan are automatically configured to bridge-domain when the configuration is loaded.
Prerequisites
•
You must have previously enabled IEEE 802.1Q-in-802.1Q VLAN translation on either a Gigabit Ethernet WAN interface, or on a port-channel interface. See either the "Enabling IEEE 802.1Q-in-802.1Q Translation on a Gigabit Ethernet WAN Interface" section and the "Enabling IEEE 802.1Q-in-802.1Q Translation on a QinQ Link Bundle" section.
Restrictions
•
You can configure a maximum of 32 inner CE VLANs for each outer PE VLAN. The inner CE VLANs must be in a contiguous block that starts on a 32-block boundary (32, 64, and so on), excluding invalid or reserved VLANs.
•
You cannot specify an out-of-range configuration for a PE VLAN until you have first configured at least one specific inner CE VLAN ID for that particular PE VLAN. This is required so that the system can determine which VLAN IDs should be considered in-range and out-of-range.
•
VLAN 4095 is reserved and cannot be used as a CE VLAN. Packets that contain a CE VLAN ID of 4095 are automatically dropped by subinterfaces that are configured for QinQ translation. VLAN 4095, however, can continue to be used as a native (non-QinQ) VLAN.
•
A PE VLAN cannot have the same ID as a native (non-QinQ) VLAN that is also being used on the router.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
interface {ge-wan slot/port.subinterface | port-channel number.subinterface}
4.
encapsulation dot1q trunk-vlan-id
5.
bridge-domain vlan-id dot1q inner-vlan-id
or
bridge-domain vlan-id dot1q-tunnel out-range
6.
mls qos trust [cos | dscp | ip-precedence]
7.
service policy input policy-name
8.
service policy output policy-name
9.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
interface {ge-wan slot/port.subinterface |
port-channel number.subinterface}
Example:
Router(config)# interface ge-wan 5/1.64
Router(config-subif)#
|
Enters subinterface mode for the specified subinterface.
|
Step 4
|
encapsulation dot1q trunk-vlan-id
Example:
Router(config-subif)# encapsulation dot1q 2
Router(config-subif)#
|
Configures the subinterface to use the specified IEEE 802.1Q trunk VLAN on outgoing packets:
• trunk-vlan-id—Specifies the trunk VLAN ID to be used for this traffic. The valid range is any VLAN from 1 to 4094, except for the numbers already allocated and the numbers in the range from 1002 to 1005, which are reserved.
|
Step 5
|
bridge-domain vlan-id dot1q inner-vlan-id
or
bridge-domain vlan-id dot1q-tunnel out-range
Example:
Router(config-subif)# bridge-domain 2 dot1q 64
Router(config-subif)#
or
Router(config-subif)# bridge-domain 2
dot1q-tunnel out-range
Router(config-subif)#
|
Creates a table map for the specified outer (provider) VLAN ID to the specified inner (customer) VLAN ID, specifying that these VLAN tags should be replaced by the trunk VLAN tag when the packet is output.
• vlan-id—VLAN ID for the provider edge (PE), or outer, VLAN. The valid range is 1 to 4094, except for the native VLAN (which defaults to 1) and the numbers in the range from 1002 to 1005, which are reserved. This value must match the VLAN that is actually configured on the provider edge router.
• dot1q inner-vlan-id—VLAN ID for the customer edge (CE), or inner, VLAN that is to be mapped to this PE VLAN. The valid range is 1 to 4094, except for the numbers in the range from 1002 to 1005, which are reserved.
• dot1q-tunnel out-range—Creates a table map for all inner (customer) VLAN IDs that are outside of the previously mapped block of 32 VLANs for this particular provider VLAN. If you do not specify an out-range mapping for a PE VLAN, the interface drops all packets for that PE VLAN that either do not have a CE VLAN tag, or that have a CE VLAN outside of the mapped block.
Note You must configure at least one subinterface with a specific CE VLAN ID for a PE VLAN, before you can use the dot1q-tunnel out-range option.
|
| |
Note When you specify the first inner-vlan-id for a PE VLAN, the interface automatically associates the correct block of 32 VLANs with that PE VLAN, and those CE VLANs cannot be used for any other purpose. For example, specifying a CE VLAN of 98 associates the VLANs from 96 to 127 with that PE VLAN. Any other CE VLANs received on that PE VLAN are considered out of range.
|
Step 6
|
mls qos trust [cos | dscp | ip-precedence]
Example:
Router(config-subif)# mls qos trust dscp
Router(config-subif)#
|
(Optional) Specifies which quality of service (QoS) bits in incoming frames can be trusted.
• cos—(Optional) Specifies that the CoS bits in incoming frames are trusted and derives the internal DSCP value from the CoS bits.
• dscp—(Optional, default) Specifies that the ToS bits in the incoming packets contain a DSCP value.
• ip-precedence—(Optional) Specifies that the IP precedence bits (found in the ToS bits) of incoming packets are trusted, and derives the internal DSCP value from the IP precedence bits.
Note To configure the interface as untrusted, use the no mls qos trust command. The Layer 2 interface then zeroes out the P bits of all incoming packets before any QoS processing is done.
|
Step 7
|
service policy input policy-name
Example:
Router(config-subif)# service policy input
policy-in1
Router(config-subif)#
|
(Supported only on physical GE-WAN interfaces, not port-channel interfaces) Specifies a policy map that should be used on incoming packets when they are received on the Gigabit Ethernet WAN interface.
|
Step 8
|
service policy output policy-name
Example:
Router(config-subif)# service policy output
cos-xlat1
Router(config-subif)#
|
Specifies a policy map that should be used on outgoing packets before they leave the Gigabit Ethernet WAN interface.
Note Policy maps that use set cos cos-inner command must be applied as the output policy on the subinterface.
|
| |
Note Repeat Step 3 through Step 8 for each subinterface/VLAN mapping to be configured.
|
Step 9
|
end
Example:
Router(config)# end
Router#
|
Exits global configuration mode and returns to privileged EXEC mode.
|
Examples
The following shows a typical configuration that creates two double-tag-to-single-tag mappings on a subinterface. The first subinterface configuration creates a specific PE/CE mapping, and the second subinterface configuration creates an out-of-range configuration:
mode dot1q-indot1q access-gateway
bridge-domain 133 dot1q 42
bridge-domain 133 dot1q-tunnel out-range
These QinQ mappings operate as follows:
•
The first subinterface matches incoming packets that are tagged with a PE VLAN ID of 133 and a CE VLAN ID of 42, and translates those packets into an outgoing packet with a single trunk VLAN ID of 2. This configuration also automatically associates the block of CE VLANs from 32 to 63 with PE VLAN 133. Any packets with a CE VLAN ID in that range that also have a PE VLAN ID of 133, and are not explicitly mapped by another subinterface, are dropped. Any other CE VLANs that are received on PE VLAN 133 are considered out of range.
•
The second subinterface matches incoming packets that are tagged with a PE VLAN ID of 133, and that either do not have a CE VLAN, or that have a CE VLAN ID that is out of range (that is ranging from 1 to 31 or from 64 to 4094). These packets are translated into an outgoing packet with a trunk VLAN ID of 31 as the outer tag and an unchanged CE VLAN inner tag (if present).
This configuration performs the following mapping on packets that have a PE VLAN ID of 133:
Table 1-1 Example Double-Tag-to-Single-Tag Mappings
PE VLAN ID
|
CE VLAN ID
|
Action
|
133
|
1 to 31
|
Mapped to trunk VLAN 31, CE VLAN 1 to 31 (out of range)
|
133
|
32 to 41
|
Dropped (because not explicitly mapped)
|
133
|
42
|
Mapped to trunk VLAN 2 (explicitly mapped by GE-WAN3/3.42)
|
133
|
43 to 63
|
Dropped (because not explicitly mapped)
|
133
|
64 to 4094
|
Mapped to trunk VLAN 31, CE VLAN 64 to 4094 (out of range)
|
133
|
(none)
|
Mapped to trunk VLAN 31 (out of range)
|
Configuring QinQ Transparent Tunneling—Double Tag to Double Tag Translation
When you configure the Advanced QinQ Service Mapping feature for QinQ transparent tunneling, as known as double-tag-to-double-tag translation, the Gigabit Ethernet WAN interface replaces the outer (provider edge or PE) VLAN tag with the trunk VLAN tag. The inner CE VLAN tag (if present) remains unchanged. Use the following procedure to configure a subinterface for double-tag-to-double-tag translation.
Note
Cisco IOS Release 12.2(18)SXD used the bridge-vlan command to configure the QinQ translation, but Cisco IOS Release 12.2(18)SXE and later releases have changed this to bridge-domain. Earlier configurations that use bridge-vlan are automatically configured to bridge-domain when the configuration is loaded.
Prerequisites
•
You must have previously enabled IEEE 802.1Q-in-802.1Q VLAN translation on either a Gigabit Ethernet WAN interface, or on a port-channel interface. See either the "Enabling IEEE 802.1Q-in-802.1Q Translation on a Gigabit Ethernet WAN Interface" section and the "Enabling IEEE 802.1Q-in-802.1Q Translation on a QinQ Link Bundle" section.
Restrictions
•
You can configure a maximum of 32 inner CE VLANs for each outer PE VLAN. The inner VLANs must be in a contiguous block that starts on a 32-block boundary (0, 32, 64, and so on).
•
VLAN 4095 is reserved and cannot be used as a CE VLAN. Packets that contain a CE VLAN ID of 4095 are automatically dropped by subinterfaces that are configured for QinQ translation. VLAN 4095, however, can continue to be used as a native (non-QinQ) VLAN.
•
You cannot specify an out-of-range configuration for a PE VLAN until you have first configured at least one specific inner CE VLAN ID for that particular PE VLAN. This is required so that the system can determine which VLAN IDs are considered in-range or out-of-range.
•
A PE VLAN cannot have the same ID as a native (non-QinQ) VLAN that is also being used on the router.
•
Input service policies (the service-policy input command) are not supported on port-channels being used for QinQ link bundling.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
interface {ge-wan slot/port.subinterface | port-channel number.subinterface}
4.
encapsulation dot1q trunk-vlan-id
5.
bridge-domain vlan-id dot1q-tunnel {inner-vlan-id | out-range}
6.
mls qos trust [cos | dscp | ip-precedence]
7.
service policy input policy-name
8.
service policy output policy-name
9.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
interface {ge-wan slot/port.subinterface |
port-channel number.subinterface}
Example:
Router(config)# interface ge-wan 5/1.64
Router(config-subif)#
|
Enters subinterface mode for the specified subinterface.
|
Step 4
|
encapsulation dot1q trunk-vlan-id
Example:
Router(config-subif)# encapsulation dot1q 2
Router(config-subif)#
|
Configures the subinterface to use the specified IEEE 802.1Q trunk VLAN on outgoing packets:
• trunk-vlan-id—Specifies the trunk VLAN ID to be used for this traffic. The valid range is any VLAN from 1 to 4094, except for the numbers already allocated and the numbers in the range from 1002 to 1005, which are reserved.
|
Step 5
|
bridge-domain vlan-id dot1q-tunnel
{inner-vlan-id | out-range}
Example:
Router(config-subif)# bridge-domain 2 dot1q 64
Router(config-subif)#
or
Router(config-subif)# bridge-domain 2 dot1q
out-range
Router(config-subif)#
|
Creates a table map for the specified outer (provider) VLAN ID to the specified inner (customer) VLAN ID, specifying that the outer VLAN tag should be replaced by the trunk VLAN tag when the packet is output (leaving the inner tag unchanged):
• vlan-id—VLAN ID for the provider edge (PE), or outer, VLAN. The valid range is 1 to 4094, except for the native VLAN (which defaults to 1) and the numbers in the range from 1002 to 1005, which are reserved. This value must match the VLAN that is actually configured on the provider edge router.
• inner-vlan-id—VLAN ID for the customer edge (CE), or inner, VLAN that is to be mapped to this PE VLAN. The valid range is 1 to 4094, except for the numbers in the range from 1002 to 1005, which are reserved.
• out-range—Matches all inner VLAN IDs that are outside of the previously mapped block of 32 VLANs for this particular provider VLAN. If you do not specify an out-range mapping for a PE VLAN, the interface drops all packets for that PE VLAN with a CE VLAN outside of the mapped block.
Note You must configure at least one subinterface for a specific CE VLAN ID for a PE VLAN, before you can use the out-range command.
|
| |
Note When you specify the first inner-vlan-id for a PE VLAN, the interface automatically associates the correct block of 32 VLANs with that PE VLAN, and those CE VLANs cannot be used for any other purpose. For example, specifying a CE VLAN of 98 associates the VLANs from 96 to 127 with that PE VLAN. Any other CE VLANs received on that PE VLAN are considered out of range.
|
Step 6
|
mls qos trust [cos | dscp | ip-precedence]
Example:
Router(config-subif)# mls qos trust dscp
Router(config-subif)#
|
(Optional) Specifies which quality of service (QoS) bits in incoming frames can be trusted.
• cos—(Optional) Specifies that the CoS bits in incoming frames are trusted and derives the internal DSCP value from the CoS bits.
• dscp—(Optional, default) Specifies that the ToS bits in the incoming packets contain a DSCP value.
• ip-precedence—(Optional) Specifies that the IP precedence bits (found in the ToS bits) of incoming packets are trusted, and derives the internal DSCP value from the IP precedence bits.
Note To configure the interface as untrusted, use the no mls qos trust command. The Layer 2 interface then zeroes out the P bits of all incoming packets before any QoS processing is done.
|
Step 7
|
service policy input policy-name
Example:
Router(config-subif)# service policy input
policy-in1
Router(config-subif)#
|
(Supported only on physical GE-WAN interfaces, not port-channel interfaces) Specifies a policy map that should be used on incoming packets when they are received on the Gigabit Ethernet WAN interface.
|
Step 8
|
service policy output policy-name
Example:
Router(config-subif)# service policy output
cos-xlat1
Router(config-subif)#
|
Specifies a policy map that should be used on outgoing packets before they leave the Gigabit Ethernet WAN interface.
|
| |
Note Repeat Step 3 through Step 8 for each subinterface/VLAN mapping to be configured.
|
Step 9
|
end
Example:
Router(config)# end
Router#
|
Exits global configuration mode and returns to privileged EXEC mode.
|
Examples
The following shows a typical configuration that creates two double-tag-to-double-tag mappings on a subinterface. The first subinterface configuration creates a specific PE/CE mapping, and the second subinterface configuration creates an out-of-range configuration:
bridge-domain 65 dot1q-tunnel 98
bridge-domain 65 dot1q-tunnel out-range
These QinQ mappings operate as follows:
•
The first subinterface matches incoming packets that are tagged with a PE VLAN ID of 65 and a CE VLAN ID of 98, and translates those packets into an outgoing packet with a trunk VLAN ID of 12 and a CE VLAN ID of 98. This configuration also automatically associates the block of CE VLANs from 96 to 127 with PE VLAN 65. Any packets with a CE VLAN ID in that range that also have a PE VLAN ID of 65, and are not explicitly mapped by another subinterface, are dropped. Any other CE VLANs that are received on PE VLAN 65 are considered out of range.
•
The second subinterface matches incoming packets that are tagged with a PE VLAN ID of 65, and that either do not have a CE VLAN tag, or that have a CE VLAN ID that is out of range (that is ranging from 1 to 95 or from 128 to 4094). These packets are translated to an outgoing packet that has a trunk VLAN ID of 31 and an unchanged CE VLAN tag (if present).
This configuration performs the following mapping on packets that have a PE VLAN ID of 65:
Table 1-2 Example Double-Tag-to-Double-Tag Mappings
PE VLAN ID
|
CE VLAN ID
|
Action
|
65
|
1 to 95
|
Mapped to trunk VLAN 31, CE VLAN 1 to 31 (out of range)
|
65
|
96 to 97
|
Dropped (because not explicitly mapped)
|
65
|
98
|
Mapped to trunk VLAN 12, CE VLAN 98 (explicitly mapped by GE-WAN3/3.42)
|
65
|
99 to 127
|
Dropped (because not explicitly mapped)
|
65
|
128 to 4094
|
Mapped to trunk VLAN 31, CE VLAN 128 to 4094 (out of range)
|
65
|
(none)
|
Mapped to trunk VLAN 31 (out of range)
|
Configuring a Policy Map to Use the Inner COS Bits
By default, the IEEE 802.1Q-to-IEEE 802.1Q translation sets the IEEE 802.1P bits (P bits) in the IEEE 802.1Q header of the outgoing packet's trunk VLAN tag by copying the P bits from the outer PE VLAN tag. To change this behavior, create a policy map with a class map that contains the set cos cos-inner command. The system then copies the P bits from the inner CE VLAN tag to the trunk VLAN tag that is put on the outgoing packet.
Prerequisites
•
After creating the policy map, you must apply it to the appropriate VLAN's subinterface by using the service-policy output command in subinterface configuration mode. See the following sections for more details:
–
Configuring QinQ Translation—Double Tag to Single Tag Translation
–
Configuring QinQ Transparent Tunneling—Double Tag to Double Tag Translation
Restrictions
•
The set cos cos-inner command is supported only for subinterfaces that are configured with an inner CE VLAN. The set cos cos-inner command is not supported on subinterfaces that use the out-range option with the bridge-domain command.
•
You cannot use these policy maps on a main Gigabit Ethernet WAN interface or on a main port-channel interface.
•
For the set cos cos-inner command to have any effect, you must configure an interface or subinterface to be a trusted interface, using the mls qos trust command. Otherwise, if the interface or subinterface is untrusted, the interface zeroes out the 802.1P bits of incoming packets before the bits can be copied to the outgoing packets.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
policy-map policy-map-name
4.
class {class-name | class-default}
5.
set cos cos-inner
6.
shape {average | peak} mean-rate [bc [be]]
7.
(other configuration commands as desired)
8.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
policy-map policy-map-name
Example:
Router(config)# policy-map pmap1
Router(config-pmap)#
|
Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy, and enters policy-map configuration mode.
• policy-map-name—Name of the policy map. The name can be a maximum of 40 alphanumeric characters.
|
Step 4
|
class {class-name | class-default}
Example:
Router(config-pmap)# class class-default
Router(config-pmap-c)#
|
Creates or modifies a policy class, and enters policy map class configuration mode.
• class-name—Name of the class to be configured or modified.
• class-default—Specifies the default class that should be used when no other class has been specified.
|
Step 5
|
set cos cos-inner
Example:
Router(config-pmap-c)# set cos cos-inner
Router(config-pmap-c)#
|
(Optional) Sets the IEEE 802.1 prioritization bits (P bits) of the trunk VLAN tag of an IEEE 802.1Q-in-802.1Q translated outgoing packet with the priority value from the incoming packet's inner (customer edge) VLAN tag. The default value is the no form of this command, which uses the P bits from the incoming packet's outer (provider edge) VLAN tag.
|
Step 6
|
shape {average | peak} mean-rate [bc [be]]
Example:
Router(config-pmap-c)# shape average 4000000
16000 16000
Router(config-pmap-c)#
|
(Optional) Specifies the traffic shaping rates to be used with this policy:
• average—(Optional) Maximum number of bits sent out in each interval is equal to the committed burst size (Bc).
• peak—(Optional) Specifies that the maximum number of bits sent out in each interval is equal to the burst size (Bc) plus the excess burst size (Be).
• mean-rate—(Optional) Also called committed information rate (CIR). Indicates the bit rate used to shape the traffic, in bits per second.
• bc—(Optional) The number of bits in a measurement interval burst size (Bc).
• be—(Optional) The number of bits permitted to go over the excess burst size (Be).
|
Step 7
|
end
Example:
Router(config-pmap-c)# end
Router#
|
Exits policy-map class configuration mode and returns to privileged EXEC mode.
|
Examples
The following example shows a typical policy map configuration using the set cos cos-inner command:
Disabling IEEE 802.1Q-in-802.1Q Mapping and Translation
To disable the mapping and translation of IEEE 802.1Q-in-802.1Q double-tagged packets on the Gigabit Ethernet interface or on one of its subinterfaces, use one of the following procedures:
•
Disabling All IEEE 802.1Q-to-802.1Q Translation on An Interface
•
Disabling IEEE 802.1Q-to-802.1Q Translation on One Subinterface
Disabling All IEEE 802.1Q-to-802.1Q Translation on An Interface
To disable all IEEE 802.1Q-to-802.1Q translation on a Gigabit Ethernet WAN interface or a port-channel interface, use the following procedure. This procedure also removes all subinterfaces and their configurations from the interface, which then allows the associated VLANs to be used for other purposes or on other cards.
Tip
Be sure to save the configuration before you begin this procedure if you want to move the configuration to another interface.
Note
Removing the interface card from the router does not remove the interface configuration, because the Cisco IOS software assumes you will be performing an online insertion and removal (OIR) operation. You must disable IEEE 802.1Q-in-802.1Q translation from all interfaces on a card before removing the card from the chassis, before the VLANs that are configured on the card can become available for use by other interfaces.
Note
You should perform online insertion and removal (OIR) in a redundant route processor system only after the standby state reaches a terminal state of the configured High Availability mode.
.Prerequisites
If you have previously attached a service policy that contains a set cos cos-inner command to the interface, you must first remove that service policy before you can use the no mode dot1q-in-dot1q access-gateway command.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
interface {ge-wan slot/port | port-channel number}
4.
shutdown
5.
no mode dot1q-in-dot1q access-gateway
6.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
interface {ge-wan slot/port | port-channel
number}
Example:
Router(config)# interface ge-wan 5/1
Router(config-if)#
|
Enters interface configuration mode for the specified Gigabit Ethernet WAN interface or port-channel interface.
|
Step 4
|
shutdown
Example:
Router(config-if)# shutdown
Router(config-if)#
|
(Optional) Disables the interface and prevents it from passing traffic.
|
Step 5
|
no mode dot1q-in-dot1q access-gateway
Example:
Router(config-if)# no mode dot1q-in-dot1q
access-gateway
Router(config-if)#
|
Disables IEEE 802.1Q-in-802.1Q translation on the interface. This disables the Advanced QinQ Service Mapping feature, and removes all subinterface configuration from the interface.
Note Be sure to save the configuration before giving this command if you plan to move the configuration to another interface.
|
Step 6
|
end
Example:
Router(config-if)# end
Router#
|
Exits global configuration mode and returns to privileged EXEC mode.
|
Disabling IEEE 802.1Q-to-802.1Q Translation on One Subinterface
Use the following procedure to disable IEEE 802.1Q-to-802.1Q translation on an individual subinterface. You can either completely delete the subinterface, or you can remove just the bridge-domain configuration on the subinterface, depending on whether you want to use the subinterface to continue passing other traffic. Both methods release the CE and PE VLANs being used on the subinterface.
Prerequisites
If you have previously attached a service policy that contains a set cos cos-inner command to the interface, you must first remove that service policy before you can use the no bridge-domain command.
SUMMARY STEPS
1.
enable
2.
configure terminal
3.
no interface {ge-wan slot/port.subinterface | port-channel number.subinterface}
or
4.
interface {ge-wan slot/port.subinterface | port-channel number.subinterface}
5.
no bridge-domain vlan-id dot1q {inner-vlan-id | out-range}
6.
end
DETAILED STEPS
| |
Command or Action
|
Purpose
|
Step 1
|
enable
Example:
Router> enable
Router#
|
Enables privileged EXEC mode. Enter your password if prompted.
|
Step 2
|
configure terminal
Example:
Router# configure terminal
Router(config)#
|
Enters global configuration mode.
|
Step 3
|
no interface {ge-wan slot/port.subinterface |
port-channel number.subinterface}
Example:
Router(config)# no interface ge-wan 5/1.64
Router(config-subif)#
|
Completely removes the subinterface and its configuration. All traffic passing through this interface stops.
Note After entering this command, proceed to Step 6.
|
or
|
Step 4
|
interface {ge-wan slot/port.subinterface |
port-channel number.subinterface}
Example:
Router(config)# interface ge-wan 5/1.64
Router(config-subif)#
|
Enters subinterface mode for the specified subinterface.
|
Step 5
|
no bridge-domain vlan-id dot1q {inner-vlan-id |
out-range}
Example:
Router(config-subif)# no bridge-domain 2 dot1q
64
Router(config-subif)#
|
Removes the table mapping for this subinterface, disabling the IEEE 802.1Q-to-IEEE802.1Q translation for this particular combination of VLANs. Traffic continues to pass, depending on the remaining configuration of the subinterface.
|
Step 6
|
end
Example:
Router(config-subif)# end
Router#
|
Exits subinterface configuration mode and returns to privileged EXEC mode.
|
Configuration Examples for Advanced QinQ Service Mapping
This section provides the following sample configurations:
•
QinQ Translation Configuration Example—Two-Tag to One-Tag Translation
•
QinQ Transparent Tunneling Configuration Example
•
QinQ Translation Using Port-Channel Interfaces Example
QinQ Translation Configuration Example—Two-Tag to One-Tag Translation
The following excerpt from a configuration file shows the configuration for a simple QinQ translation, in which incoming packets are received with inner customer edge (CE) and outer provider edge (PE) VLAN tags. The packets are then output, using the configured policy map, with a single trunk VLAN tag.
This configuration configures Gigabit Ethernet WAN interface 4/1 as the QinQ access gateway, and shows two PE-to-CE mappings:
•
The first set of subinterfaces is configured for a PE VLAN ID of 2 and CE VLAN IDs in the range of 32 to 46. These subinterfaces are all configured as trusted (mls qos trust dscp) and use policy maps that use the set cos cos-inner command, so that the 802.1P bits in the customer's original CE VLAN tag are copied to the outgoing trunk VLAN tag.
Subinterface 47 is configured to match any packets that arrive with a PE VLAN ID of 2 and an out-of-range CE VLAN ID (between 47 and 63). Note that the set cos cos-inner command has no effect on out-of-range packets, even when using a policy map that includes this command.
•
The second set of subinterfaces is configured for a trunk VLAN ID of 100 and a PE VLAN ID of 45. These subinterfaces accept incoming CE VLAN IDs in the range of 1237 to 1240. This configuration does not include an out-of-range subinterface, so any packets that arrive with a PE VLAN ID of 45 and an out-of-range CE VLAN ID (from 1216 to 1236 and from 1241 to 1247) are dropped. All subinterfaces use a policy map that does not include the set cos cos-inner command, which means that the trunk VLAN tag uses the 802.1P bits in the original PE VLAN tag.
vlan internal allocation policy descending
shape average 8000000 32000 32000
shape average 20000000 80000 80000
shape average 2000000 16000 16000
interface GigabitEthernet4/1
description connected to SP GE1/3
logging event link-status
switchport trunk encapsulation dot1q
interface GigabitEthernet4/2
!--This is the QinQ Access Gateway interface
description connected to PE-4 GigabitEthernet0/3
logging event link-status
mode dot1q-in-dot1q access-gateway
!--This command configures the interface as trusted, which
!--is required to be able to use the original packet's 802.1P CoS bits.
!--First set of PE/CE mappings
!--note that this bridge-domain command automatically configures the
!--CE VLAN range for this PE VLAN to be from 32 to 63
service-policy output pmap3
service-policy output pmap2
service-policy output pmap1
service-policy output pmap2
service-policy output pmap3
service-policy output pmap1
service-policy output pmap1
service-policy output pmap2
service-policy output pmap3
service-policy output pmap2
service-policy output pmap1
service-policy output pmap2
service-policy output pmap3
service-policy output pmap3
service-policy output pmap1
description out-of-range configuration for CE VLANs 47 to 63
bridge-domain 2 dot1q-tunnel out-range
!-- Although this policy map includes the set cos cos-inner command,
!-- this command is not used for out-of-range packets
service-policy output pmap4
!--Second set of PE/CE mappings
!--note that this bridge-domain command automatically configures the
!--CE VLAN range for this PE VLAN to be from 1216 to 1247
bridge-domain 45 dot1q 1237
service-policy output pmap4
bridge-domain 45 dot1q 1238
service-policy output pmap4
bridge-domain 45 dot1q 1239
service-policy output pmap4
bridge-domain 45 dot1q 1240
service-policy output pmap4
QinQ Transparent Tunneling Configuration Example
The following excerpt from a configuration file shows a typical configuration for a simple QinQ transparent tunneling configuration, in which incoming packets are received with inner customer edge (CE) and outer provider edge (PE) VLAN tags. The packets are then output, using the configured policy map, with a new trunk VLAN tag and the original inner CE VLAN tag. This configuration is called two-tag to one-tag translation.
This configuration configures Gigabit Ethernet WAN interface 4/1 as the QinQ access gateway, and creates a PE/CE mapping with the following characteristics:
•
PE VLAN ID of 152.
•
CE VLAN IDs in the range from 2048 to 2079.
•
Subinterface GE-WAN 4/1.15233 matches any packets that contain CE VLAN IDs that are outside of this range (either from 1 to 2047 or from 2080 to 4094).
•
The interface and all subinterfaces, except for the out-of-range subinterface, are configured as trusted (mls qos trust dscp), which allows them to copy the 802.1P bits in the packet's original PE VLAN tag to the outgoing trunk VLAN tag. (The original CE VLAN tag is unchanged and includes its original 802.1P bits.)
vlan internal allocation policy descending
!--This is an IP LAN interface
interface GigabitEthernet4/1
description QinQ tunnel to Catalyst 3550 Gigabit Ethernet 0/6
switchport trunk encapsulation dot1q
switchport trunk allowed vlan 340
interface GigabitEthernet4/2
!--This is the QinQ Access Gateway interface
description connected to GSR Gigabit Ethernet 4/1
logging event link-status
mode dot1q-in-dot1q access-gateway
interface GE-WAN4/1.15201
!--note that this bridge-domain command automatically configures the
!--CE VLAN range for this PE VLAN to be from 2048 to 2079
bridge-domain 152 dot1q-tunnel 2048
interface GE-WAN4/1.15203
bridge-domain 152 dot1q-tunnel 2049
interface GE-WAN4/1.15204
bridge-domain 152 dot1q-tunnel 2050
interface GE-WAN4/1.15205
bridge-domain 152 dot1q-tunnel 2051
interface GE-WAN4/1.15206
bridge-domain 152 dot1q-tunnel 2052
interface GE-WAN4/1.15207
bridge-domain 152 dot1q-tunnel 2053
interface GE-WAN4/1.15208
bridge-domain 152 dot1q-tunnel 2054
interface GE-WAN4/1.15209
bridge-domain 152 dot1q-tunnel 2055
interface GE-WAN4/1.15210
bridge-domain 152 dot1q-tunnel 2056
interface GE-WAN4/1.15211
bridge-domain 152 dot1q-tunnel 2057
interface GE-WAN4/1.15212
bridge-domain 152 dot1q-tunnel 2058
interface GE-WAN4/1.15213
bridge-domain 152 dot1q-tunnel 2059
interface GE-WAN4/1.15214
bridge-domain 152 dot1q-tunnel 2060
interface GE-WAN4/1.15215
bridge-domain 152 dot1q-tunnel 2061
interface GE-WAN4/1.15216
bridge-domain 152 dot1q-tunnel 2062
interface GE-WAN4/1.15217
bridge-domain 152 dot1q-tunnel 2063
interface GE-WAN4/1.15218
bridge-domain 152 dot1q-tunnel 2064
interface GE-WAN4/1.15219
bridge-domain 152 dot1q-tunnel 2065
interface GE-WAN4/1.15220
bridge-domain 152 dot1q-tunnel 2066
interface GE-WAN4/1.15221
bridge-domain 152 dot1q-tunnel 2067
interface GE-WAN4/1.15222
bridge-domain 152 dot1q-tunnel 2068
interface GE-WAN4/1.15223
bridge-domain 152 dot1q-tunnel 2069
interface GE-WAN4/1.15224
bridge-domain 152 dot1q-tunnel 2070
interface GE-WAN4/1.15225
bridge-domain 152 dot1q-tunnel 2071
interface GE-WAN4/1.15226
bridge-domain 152 dot1q-tunnel 2072
interface GE-WAN4/1.15227
bridge-domain 152 dot1q-tunnel 2073
interface GE-WAN4/1.15228
bridge-domain 152 dot1q-tunnel 2074
interface GE-WAN4/1.15229
bridge-domain 152 dot1q-tunnel 2075
interface GE-WAN4/1.15230
bridge-domain 152 dot1q-tunnel 2076
interface GE-WAN4/1.15231
bridge-domain 152 dot1q-tunnel 2077
interface GE-WAN4/1.15232
bridge-domain 152 dot1q-tunnel 2078
! This creates an out-of-range configuration that matches CE VLANs
! that are out of the configured CE VLAN range of 2048 to 2079
interface GE-WAN4/1.15233
bridge-domain 152 dot1q-tunnel out-range
QinQ Translation Using Port-Channel Interfaces Example
The following shows a sample configuration of a QinQ link bundle that contains two GE-WAN physical interfaces. Note that the bridge-domain commands are configured on the subinterfaces of the port-channel virtual interface.
vlan internal allocation policy ascending
vlan 1, 100-1000, 2976-3008
shape average 8000000 32000 32000
shape average 20000000 80000 80000
logging event link-status
mode dot1q-in-dot1q access-gateway
interface Port-channel1.101
bridge-domain 101 dot1q 101
service-policy output pmap1
interface Port-channel1.102
bridge-domain 102 dot1q 102
service-policy output pmap2
interface Port-channel1.103
bridge-domain 103 dot1q 103
interface Port-channel1.104
bridge-domain 104 dot1q 104
interface Port-channel1.201
bridge-domain 201 dot1q 201
! GigabitEthernet interfaces are not used for QinQ
! link bundling, but can be used for
interface GigabitEthernet4/1
interface GigabitEthernet4/2
logging event link-status
logging event link-status