Catalyst 4500 Series Switch Software Configuration Guide, 12.2(53)SG

Configuring Ethernet CFM and OAM

Ethernet Operations, Administration, and Maintenance (OAM) is a protocol for installing, monitoring, and troubleshooting Ethernet networks to enhance management within the Ethernet infrastructure. The Catalyst 4500 series switch supports IEEE 802.1ag Connectivity Fault Management (CFM) and IEEE 802.3ah Ethernet OAM discovery, link monitoring, remote fault detection, and remote loopback. The Ethernet OAM manager controls the interaction between CFM and OAM.

Note For complete command and configuration information for CFM, see the Cisco IOS feature module at this URL:

http://www.cisco.com/en/US/docs/ios/cether/configuration/guide/ce_cfm.html

This chapter contains these sections:

•Command List

•Ethernet CFM Overview

•Configuring Ethernet CFM

•Displaying Ethernet CFM Information

•Ethernet OAM Protocol Overview

•Setting Up and Configuring Ethernet OAM

•Displaying Ethernet OAM Protocol Information

•Ethernet CFM and Ethernet OAM Interaction

Command List

This table lists the commands most commonly used with Ethernet CFM and OAM.

Switch(config-if)# no ehernet cfm 
enable

Switch(config)# ethernet cfm 
traceroute cache [size entries | 
hold-time minutes]

Switch(config)# ethernet cfm 
domain domain-name level level-id 
{direction outward}

Switch(config-ether-cfm)# [no] 
service csi-id vlan vlan-id

Switch(config-ether-cfm)# mep 
archive-hold-time minutes 

Switch(config-if)# ethernet cfm 
mip level level-id 

Switch(config-if)# ethernet cfm 
mep level level-id {[inward] | 
outward} mpid id vlan {vlan-id | 
any | vlan-id-vlan-id 
[,vlan-id-vlan-id]}

Switch(config)# ethernet cfm cc 
{[enable] level {level-id | any} 
vlan {vlan-id | any}}

Switch(config)# snmp-server enable 
traps ethernet cfm cc [mep-up] 
[mep-down] [config] [loop] 
[cross-connect]

Switch(config)# snmp-server enable 
traps ethernet cfm crosscheck 
[mep-unknown] [mep-missing] 
[service-up]

Ethernet CFM Overview

Ethernet CFM is an end-to-end per-service-instance (per VLAN) Ethernet layer OAM protocol. It includes proactive connectivity monitoring, fault verification, and fault isolation. End-to-end can be provider-edge-to provider-edge (PE-to-PE) device or customer-edge-to-customer-edge (CE-to-CE) device. Ethernet CFM, as specified by IEEE 802.1ag, is the standard for Layer 2 ping, Layer 2 traceroute, and end-to-end connectivity verification of the Ethernet network.

Unlike CFM, other metro-Ethernet OAM protocols are not end-to-end technologies. For example, IEEE 802.3ah OAM is a single-hop and per-physical-wire protocol and is not end-to-end or service aware.

These sections contain conceptual information about Ethernet CFM:

•Definition List

•CFM Domain

•CFM Maintenance Points

•General Packet Forwarding Rules

•CFM Messages

•Crosscheck Function

•SNMP Traps

•IP SLAs Support for CFM

Definition List

CFM Domain

A CFM maintenance domain is a management space on a network that is owned and operated by a single entity and defined by a set of internal boundary ports. You assign a unique maintenance level (from 0 to 7) to define the domain hierarchy. The larger the domain, the higher the level. For example, as shown in Figure 55-1, a service-provider domain would be larger than an operator domain and might have a maintenance level of 6, while the operator domain maintenance level would be 3 or 4.

As shown in Figure 55-2, domains cannot intersect or overlap because that would require management by more than one entity, which is not allowed. Domains can touch or nest (if the outer domain has a higher maintenance level than the nested domain). Nesting domains can be useful when a service provider contracts with one or more operators to provide Ethernet service. Each operator has its own maintenance domain and the service provider domain is a superset of the operator domains. Maintenance levels of nesting domains should be communicated among the administrating organizations. CFM exchanges messages and performs operations on domains individually.

Figure 55-1 CFM Maintenance Domains

Figure 55-2 Allowed Domain Relationships

CFM Maintenance Points

Operating with a maintenance domain, a maintenance point demarcates an interface that participates in an CFM. Maintenance points drop all lower-level frames and forward all higher-level frames. There are two types of maintenance points:

•Maintenance end points (MEPs) are points at the edge of the domain that define the boundary and confine CFM messages within the boundary. MEPs are inward facing by default. Inward facing means that they communicate through the relay function side, not the wire side (connected to the port), whereas MEPs that can be configured as outward facing communicate through the wire side, and not through the relay function side.

An inward-facing MEP sends and receives CFM frames through the relay function. It drops all CFM frames at its level or lower that come from the wire side. For CFM frames from the relay side, it processes the frames at its level and drops frames at a lower level. The MEP transparently forwards all CFM frames at a higher level, whether they are received from the relay or wire side. CFM runs at the provider maintenance level (UPE-to-UPE), specifically with inward-facing MEPs at the user network interface (UNI).

An outward facing MEP (OFM) sends and receives CFM frames on the wire side. It drops all CFM frames at its level or lower that come from the relay function side. For CFM frames from the wire side, it processes the frames at its level and drops frames at a lower level. OFM transparently forwards all CFM frames at a higher level, whether they are received from the relay or wire side.

•Maintenance intermediate points (MIPs) are inside a domain, not at the boundary, and respond to CFM only when triggered by traceroute and loopback messages. They forward CFM frames received from MEPs and other MIPs, drop all CFM frames at a lower level, and forward all CFM frames at a higher level, whether they are received from the relay or wire side.

If a port on which the inward facing MEP is configured is blocked by Spanning-Tree Protocol (STP), the MEP cannot receive or transmit CFM messages. If a port on which the outward facing MEP is configured is blocked by STP, the OFM can only receive CFM messages from and transmit them towards the wire. If a port on which a MIP is configured is blocked by STP, the port cannot receive or respond to messages from the relay function side, but can receive CFM messages and respond to them from the wire side.

General Packet Forwarding Rules

Ethernet CFM frames should be forwarded or dropped based on the strict rules of hierarchical maintenance domains. MEPs and MIPs configured on bridge ports act as filters that confine CFM frames within the bounds of the correct domain(s) by dropping frames that do not belong to the correct Level.

Topics include:

•Inward-Facing MEPs

•Outward-Facing MEPs

•Transparent Ports

Inward-Facing MEPs

An inward-facing MEP does the following:

•Sends and receives CFM frames at its level through the relay function, not through the wire connected to the port on which the MEP is configured.

•Drops all CFM frames at its level (or lower level) that come from the wire side.

•Processes all CFM frames at its level that come from the direction of the relay function.

A packet arriving on a wire at a port is coming from the wire side.

A packet arriving internally from the CPU or by the bridging action in hardware (or software) is coming from the relay function side.

•Drops all CFM frames at a lower level that come from the direction of the relay function.

•Transparently forwards all CFM frames at a higher level, whether they come from the relay function or the wire side.

Note A MEP of level L (where L != 7) requires a MIP of level M > L on the same port. So, CFM frames at a higher level than the MEP are cataloged by this MIP.

•If the port on which the MEP is configured is blocked by STP, the MEP can no longer transmit or receive CFM messages.

Note On Catalyst 4500 Supervisor Engine 6-ME, outward MEPs are only supported on the supervisor uplink ports.

Outward-Facing MEPs

An outward-facing MEP does the following:

•Sends and receives CFM frames at its level through the wire connected to the port through which the MEP is configured.

•Drops all CFM frames at its level (or lower level) that come from the relay function side.

•Processes all CFM frames at its level coming from the direction of the wire.

•Drops all CFM frames at a lower level coming from the direction of the wire.

•Transparently forwards all CFM frames at a higher level, whether they come in from the relay function side or the wire side.

A MEP of level L (where L != 7) requires a MIP of level M > L on the same port. So, CFM frames at a higher level than the MEP are catalogued by this MIP.

•If the port on which the MEP is configured is blocked by STP, the MEP can still transmit and receive CFM messages through the wire.

•A MIP catalogues and forwards CFM frames at its level both through the wire and through the relay function.

•A MIP stops and drops all CFM frames at a lower level whether they come from the wire or relay function side.

•A MIP transparently forwards CFM frames at a higher level whether they come from the wire or relay function side.

•If the port on which the MIP is configured is blocked by STP, the MIP cab no longer receive or relay CFM messages towards the relay function side; however, it can still receive and respond to CFM messages from the wire.

Transparent Ports

A transparent port has neither a MEP nor MIP configured, and forwards CFM frames like regular data traffic.

STP blocking applies to CFM frames on transparent ports just as it applies to ports with inward-facing MEPs; if the port is blocked by STP, CFM frames are dropped as they attempt to ingress or egress that port.

CFM Messages

CFM uses standard Ethernet frames distinguished by EtherType or (for multicast messages) by MAC address. All CFM messages are confined to a maintenance domain and to a service-provider VLAN (S-VLAN). Four CFM messages are supported:

•Continuity Check (CC) messages—multicast heartbeat messages exchanged periodically between MEPs that allow MEPs to discover other MEPs within a domain and allow MIPs to discover MEPs. CC messages are confined to a domain or VLAN.

•Loopback messages—unicast frames transmitted by a MEP at administrator request to verify connectivity to a particular maintenance point, indicating whether a destination is reachable. A loopback message is similar to an Internet Control Message Protocol (ICMP) ping message.

•Traceroute messages—multicast frames transmitted by a MEP at administrator request to track the path (hop-by-hop) to a destination MEP. Traceroute messages are similar in concept to UDP traceroute messages.

•AIS messages—described in Chapter 55 "Configuring Ethernet CFM and OAM."

Crosscheck Function

The crosscheck function verifies a post-provisioning timer-driven service between dynamically configured MEPs (using crosscheck messages) and expected MEPs (by configuration) for a service. It verifies that all endpoints of a multipoint service are operational. The crosscheck function is performed only one time and is initiated from the command-line interface (CLI).

SNMP Traps

The MEPs generate two types of SMNP traps: CC traps and crosscheck traps.

Supported CC traps include:

•MEP up

•MEP down

•cross-connect (a service ID does not match the VLAN)

•loop

•configuration error

Supported crosscheck traps include:

•service up

•MEP missing (an expected MEP is down)

•unknown MEP

IP SLAs Support for CFM

The Metro switch supports CFM with IP Service Level Agreements (SLAs), which gathers Ethernet layer network performance metrics. Available statistical measurements for the IP SLAs CFM operation include round-trip time, jitter (interpacket delay variance), and packet loss. You can schedule multiple IP SLA operations and use Simple Network Management Protocol (SNMP) trap notifications and syslog messages to monitor threshold violations proactively.

IP SLA integration with CFM gathers Ethernet layer statistical measurements by sending and receiving Ethernet data frames between CFM MEPs. Performance is measured between the source MEP and the destination MEP. Unlike other IP SLA operations that provide performance metrics for only the IP layer, IP SLAs with CFM provide performance metrics for Layer 2.

You can manually configure individual Ethernet ping or jitter operations. You can also configure an IP SLA automatic Ethernet operation that queries the CFM database for all MEPs in a given maintenance domain and VLAN. The operation then automatically creates individual Ethernet ping or jitter operations based on the discovered MEPs.

For more information about IP SLA operation with CFM, see the IP SLAs for Metro-Ethernet feature module at this URL:

http://www.cisco.com/en/US/docs/ios/12_2sr/12_2srb/feature/guide/sr_meth.html

Configuring Ethernet CFM

To configure Ethernet CFM you must prepare the network and configuring services. You can optionally configure and enable crosschecking. These sections are included:

•Default Ethernet CFM Configuration

•Ethernet CFM Configuration Guidelines

•Disabling CFM on a Port

•Configuring the Ethernet CFM Service over VLANs

•Configuring Ethernet CFM Crosscheck for VLANs

•Configuring IP SLAs CFM Operation

•Example: Switchport/VLAN CFM with an Inward-Facing MEP

Default Ethernet CFM Configuration

CFM is globally disabled.

CFM is enabled on all interfaces. A port can be configured as a flow point (MIP/MEP), a transparent port, or disabled (CFM disabled). By default, ports are transparent until configured as MEP, MIP, or disabled.

There are no MEPs or MIPs configured.

Ethernet CFM Configuration Guidelines

Here are some configuration guidelines and restrictions for CFM:

•A MIP should be configured on a port before a MEP, unless the MEP is at level 7 or the MEP is an outward-facing MEP (OFM). Similarly, all the MEPs have to be removed before a MIP on a port.

•CFM Unicast packets (Loopback Messages and Traceroute Reply), are not allowed on an OFM on STP blocked ports. So, the blocked port cannot respond to ping and traceroute.

•CFM is not supported and cannot be configured on routed ports.

•CFM is not supported and cannot be configured on dot1q-tunnel ports.

•CFM is supported on EtherChannel port channels. You can configure an EtherChannel port channel as MEP or MIP. However, CFM is not supported on individual ports that belong to an EtherChannel and you cannot add a CFM port to an EtherChannel group.

•You cannot configure CFM on VLAN interfaces.

•You cannot configure CFM on an EoMPLS port.

•CFM is not supported and cannot be configured on a PVLAN isolated host port, community host port or promiscuous access port.

•CFM is supported only on regular VLANs for inward-facing MEP on PVLAN trunks. Whereas OFM is supported on regular VLANs and isolated VLANs on PVLAN secondary trunk, similarly OFM is supported on regular VLANs and primary VLANs on promiscuous trunk ports.

The CFM service on a PVLAN ends at the PVLAN trunk. The translation of CFM service from one PVLAN to another PVLAN is not supported between the PVLAN trunks.

Disabling CFM on a Port

When CFM is globally enabled, you might want to disable CFM selectively on a port (or port channel).

Note By default, CFM is globally disabled and enabled at every port.

To configure the network for Ethernet CFM over VLANs, follow these steps:

Switch# configure terminal

Switch(config)# interface interface-id 

Switch(config-if)# no ethernet cfm enable

Switch(config-if)# exit

Switch(config)# end

Configuring the Ethernet CFM Service over VLANs

To configure the network for Ethernet CFM over VLANs, follow these steps:

Switch# configure terminal

Switch(config)# ethernet cfm enable

Switch(config)# vlan vlan-id

Switch(config)# ethernet cfm traceroute cache 
[size entries | hold-time minutes]

Switch(config)# ethernet cfm domain 
domain-name level level-id {direction outward}

Switch(config-ether-cfm)# [no] service csi-id 
vlan vlan-id

Switch(config-ether-cfm)# mep 
archive-hold-time minutes 

Switch(config-ether-cfm)# exit

Switch(config)# interface interface-id 

Switch(config-if)# ethernet cfm mip level 
level-id 

Switch(config-if)# ethernet cfm mep level 
level-id {[inward] | outward} mpid id vlan 
{vlan-id | any | vlan-id-vlan-id 
[,vlan-id-vlan-id]}

Switch(config-ether-cfm)# exit

Switch(config)# ethernet cfm cc {[enable] 
level {level-id | any} vlan {vlan-id | any}}

Switch(config)# snmp-server enable traps 
ethernet cfm cc [mep-up] [mep-down] [config] 
[loop] [cross-connect]

Switch(config)# snmp-server enable traps 
ethernet cfm crosscheck [mep-unknown] 
[mep-missing] [service-up]

Switch(config)# end

Switch# show ethernet cfm domain brief

Switch# show ethernet cfm maintenance-points 
local

Switch# show ethernet cfm traceroute-cache

Switch# show running-config

Switch# copy running-config startup-config

Use the no versions of the commands to remove the configuration or return to the default configurations.

Configuring Ethernet CFM Crosscheck for VLANs

To configure Ethernet CFM crosscheck for VLANs, follow these steps:

Switch# configure terminal

Switch(config)# ethernet cfm mep crosscheck 
start-delay delay

Switch(config)# ethernet cfm domain 
domain-name level level-id {direction outward}

Switch(config-ether-cfm)# mep crosscheck mpid 
identifier vlan vlan-id [mac remote MAC address]

Switch(config)# end

Switch# ethernet cfm mep crosscheck {enable | 
disable} level level-id vlan {vlan-id | any}

Switch# show ethernet cfm maintenance-points 
remote crosscheck

Switch# show ethernet cfm errors

Switch# copy running-config startup-config

Use the no form of each command to remove a configuration or to return to the default settings.

Configuring IP SLAs CFM Operation

You can manually configure an IP SLA's Ethernet ping or jitter echo operation, or you can configure
IP SLAs Ethernet operation with endpoint discovery. You can also configure multiple operation scheduling. For accurate one-way delay statistics, the clocks on the endpoint switches must be synchronized. You can configure the endpoint switches with Network Time Protocol (NTP) so that the switches are synchronized to the same clock source.

Note When you configure the Catalyst 4500 series switch for a class of service (CoS) probe, you must first globally enable QoS by entering the mls qos global configuration command.

For detailed information about configuring IP SLAs operations, see the Cisco IOS IP SLAs Configuration Guide, Release 12.4T at this URL:

http://www.cisco.com/en/US/products/ps6441/products_installation_and_configuration_guides_list.html

For detailed information about IP SLAs commands, see the command reference at this URL:

http://www.cisco.com/en/US/docs/ios/ipsla/command/reference/sla_book.html

This section includes these procedures:

•Manually Configuring an IP SLAs CFM Probe or Jitter Operation

•Configuring an IP SLAs Operation with Endpoint Discovery

Manually Configuring an IP SLAs CFM Probe or Jitter Operation

To manually configure an IP SLAs Ethernet echo (ping) or jitter operation, follow these steps:

configure terminal

ip sla operation-number

ethernet echo mpid identifier domain 
domain-name vlan vlan-id

or

ethernet jitter mpid identifier domain 
domain-name vlan vlan-id [interval 
interpacket-interval] [num-frames number-of 
frames transmitted]

cos cos-value

frequency seconds

history history-parameter

owner owner-id

request-data-size bytes 

tag text 

threshold milliseconds 

timeout milliseconds 

exit

ip sla schedule operation-number [ageout 
seconds] [life {forever | seconds}] 
[recurring] [start-time {hh:mm {:ss} [month 
day | day month] | pending | now | after 
hh:mm:ss}]

end

show ip sla configuration [operation-number]

copy running-config startup-config

To remove an IP SLAs operation, enter the no ip sla operation-number global configuration command.

Configuring an IP SLAs Operation with Endpoint Discovery

To use IP SLAs to automatically discover the CFM endpoints for a domain and VLAN ID, follow these steps. You can configure ping or jitter operations to the discovered endpoints.

configure terminal

ip sla ethernet-monitor operation-number

type echo domain domain-name vlan vlan-id 
[exclude-mpids mp-ids]

or

type jitter domain domain-name vlan vlan-id 
[exclude-mpids mp-ids] [interval 
interpacket-interval] [num-frames number-of 
frames transmitted]

cos cos-value

owner owner-id

request-data-size bytes 

tag text 

threshold milliseconds 

timeout milliseconds 

exit

ip sla schedule operation-number [ageout 
seconds] [life {forever | seconds}] 
[recurring] [start-time {hh:mm {:ss} [month 
day | day month] | pending | now | after 
hh:mm:ss}]

end

show ip sla configuration [operation-number]

copy running-config startup-config

To remove an IP SLAs operation, enter the no ip sla operation-number global configuration command.

Example: Switchport/VLAN CFM with an Inward-Facing MEP

The following is an example of a VLAN-based CFM configuration between two switches. In this example, a Supervisor Engine II+10GE switch called g6-1 is connected to a Metro Ethernet Supervisor Engine 6-E switch called Switch. Gi 6/5 of g6-1 is connected to the gi 3/5 of the Switch through an ethernet cable.

Configuration on the Supervisor Engine II+10GE ("g6-1")
-----------------------------------------------------

!

ethernet cfm domain customer2 level 6

ethernet cfm domain PROVIDER2 level 5

 service customerX vlan 102

ethernet cfm enable

!

!

vlan 102

!

interface GigabitEthernet6/2

 switchport access vlan 102

 ethernet cfm mip level 6

 ethernet cfm mep level 5 mpid 2101 vlan 102

!

interface GigabitEthernet6/5

 switchport trunk encapsulation dot1q

 switchport trunk allowed vlan 102

 switchport mode trunk

 ethernet cfm mip level 5

!

ethernet cfm cc enable level 5 vlan 102

!

Screen dumps from "g6-1"
-----------------------------------------------------

g6-1# show ethernet cfm main rem

Can only Ping/Traceroute to remote MEPs marked with *

-------------------------------------------------------------------------------

MPID Level Mac Address    Vlan PortState InGressPort     Age(sec)        Service ID                                                          

-------------------------------------------------------------------------------

2111*  5     001b.d550.90fd  102   UP            Gi6/5                 6                 
customerX

Total Remote MEPs: 1

g6-1#

g6-1# show ethernet cfm main local

sh ethernet cfm main local

-------------------------------------------------------------------------------

MPID Level Type  VLAN Port       CC-Status MAC                           DomainName                                                          

-------------------------------------------------------------------------------

2101    5      MEP I 102     Gi6/2      Enabled   000a.4172.df3d              PROVIDER2

-------------------------------------------------------------------------------

Level Type  Port                      MAC                                      

-------------------------------------------------------------------------------

6     MIP   Gi6/2                     000a.4172.df3d

5     MIP   Gi6/5                     000a.4172.df3d

g6-1#

Configuration on the Metro Ethernet Supervisor Engine 6-E Switch ("Switch")
-----------------------------------------------------

!

ethernet cfm domain customer2 level 6

ethernet cfm domain PROVIDER2 level 5

 service customerX vlan 102

ethernet cfm enable

!

vlan 102

!

interface GigabitEthernet3/1

 switchport mode trunk

 ethernet cfm mip level 6

 ethernet cfm mep level 5 mpid 2111 vlan 102

!

interface GigabitEthernet3/5

 switchport mode trunk

 ethernet cfm mip level 5

!

ethernet cfm cc enable level 5 vlan 102

!

Screen dumps on "Switch"
-----------------------------------------------------

Switch# show ethernet cfm main rem

Can only Ping/Traceroute to remote MEPs marked with *

MPID  Level Mac Address    Vlan PortState InGressPort   Age(sec) Service ID

2101*    5     000a.4172.df3d 102   UP          Gi3/5             1              customerX

Total Remote MEPs: 1

Switch# show ethernet cfm main local

MPID DomainName     Level Type  VLAN   Port       CC-Status    MAC

2111 PROVIDER2         5        MEP 102         Gi3/1   Enabled        001b.d550.90fd

Level Type  Port                      MAC

6     MIP   Gi3/1                     001b.d550.90fd

5     MIP   Gi3/5                     001b.d550.90fd

Displaying Ethernet CFM Information

To display Ethernet CFM information, you can use the privileged EXEC commands in Table 55-1.

show ethernet cfm domain brief

show ethernet cfm errors

show ethernet cfm maintenance-points local

show ethernet cfm maintenance-points 
remote [detail | domain | level]

show ethernet cfm maintenance-points 
remote crosscheck

show ethernet cfm traceroute-cache

To display IP SLAs Ethernet CFM information, you can use the privileged EXEC commands in Table 55-2.

show ip sla configuration [entry-number] 

show ip sla ethernet-monitor configuration 
[entry-number] 

show ip sla statistics [entry-number | 
aggregated | details] 

Ethernet OAM Protocol Overview

The Ethernet OAM protocol for installing, monitoring, and troubleshooting Metro Ethernet networks and Ethernet WANs relies on an optional sublayer in the data link layer of the OSI model. Normal link operation does not require Ethernet OAM. You can implement Ethernet OAM on any full-duplex point-to-point or emulated point-to-point Ethernet link for a network or part of a network (specified interfaces).

OAM frames, called OAM protocol data units (OAM PDUs) use the slow protocol destination MAC address 0180.c200.0002. They are intercepted by the MAC sublayer and cannot propagate beyond a single hop within an Ethernet network. Ethernet OAM is a relatively slow protocol, with a maximum transmission rate of 10 frames per second, resulting in minor impact to normal operations. However, because the CPU must poll error counters frequently, when you enable link monitoring, the number of required CPU cycles is proportional to the number of interfaces that must be polled.

Ethernet OAM has two major components:

•The OAM client establishes and manages Ethernet OAM on a link and enables and configures the OAM sublayer. During the OAM discovery phase, the OAM client monitors OAM PDUs received from the remote peer and enables OAM functionality. After the discovery phase, it manages the rules of response to OAM PDUs and the OAM remote loopback mode.

•The OAM sublayer presents two standard IEEE 802.3 MAC service interfaces facing the superior and inferior MAC sublayers. It provides a dedicated interface for the OAM client to pass OAM control information and PDUs to and from the client. The sublayer includes these components:

–The control block provides the interface between the OAM client and other OAM sublayer internal blocks.

–The multiplexer manages frames from the MAC client, the control block, and the parser and passes OAM PDUs from the control block and loopback frames from the parser to the subordinate layer.

–The parser classifies frames as OAM PDUs, MAC client frames, or loopback frames and sends them to the appropriate entity: OAM PDUs to the control block, MAC client frames to the superior sublayer, and loopback frames to the multiplexer.

OAM Features

These OAM features are defined by IEEE 802.3ah:

•Discovery identifies devices in the network and their OAM capabilities. It uses periodic OAM PDUs to advertise OAM mode, configuration, and capabilities; PDU configuration; and platform identity. An optional phase allows the local station to accept or reject the configuration of the peer OAM entity.

•Link monitoring detects and indicates link faults under a variety of conditions and uses the event notification OAM PDU to notify the remote OAM device when it detects problems on the link. Error events include when the number of symbol errors, the number of frame errors, the number of frame errors within a specified number of frames, or the number of error seconds within a specified period exceeding a configured threshold.

•Remote failure indication conveys a slowly deteriorating quality of an OAM entity to its peers by communicating these conditions: Link Fault means a loss of signal, Dying Gasp means an unrecoverable condition, and Critical Event means an unspecified vendor-specific critical event. The switch can receive and process but not generate Link Fault or Critical Event OAM PDUs. It can generate Dying Gasp OAM PDUs to show that Ethernet OAM is disabled, the interface is shut down, the interface enters the error-disabled state, or the switch is reloading. It can respond to, but not generate, Dying Gasp PDUs based on loss of power.

•Remote loopback mode ensures link quality with a remote peer during installation or troubleshooting. In this mode, when the switch receives a frame that is not an OAM PDU or a pause frame, it sends it back on the same port. The link appears to the user to be functioning. You can use the returned loopback acknowledgement to test delay, jitter, and throughput.

OAM Messages

Ethernet OAM messages or PDUs are standard length, untagged Ethernet frames between 64 and 1518 bytes. They do not go beyond a single hop and have a maximum transmission rate of 10 OAM PDUs per second. Message types are information, event notification, loopback control, or vendor-specific OAM PDUs.

Setting Up and Configuring Ethernet OAM

This section includes this information:

•Default Ethernet OAM Configuration

•Ethernet OAM Configuration Guidelines

•Enabling Ethernet OAM on an Interface

•Enabling Ethernet OAM Remote Loopback

•Configuring Ethernet OAM Link Monitoring

•Configuring Ethernet OAM Remote Failure Indications

•Configuring Ethernet OAM Templates

Default Ethernet OAM Configuration

The default configuration is as follows:

•Ethernet OAM is disabled on all interfaces.

•When Ethernet OAM is enabled on an interface, link monitoring is automatically turned on.

•Remote loopback is disabled.

•No Ethernet OAM templates are configured.

Ethernet OAM Configuration Guidelines

Follow these guidelines when configuring Ethernet OAM:

•The switch does not support monitoring of egress frames sent with cyclic redundancy code (CDC) errors. The ethernet oam link-monitor transmit crc interface-configuration or template-configuration commands are visible but are not supported on the switch. The commands are accepted but are not applied to an interface.

•For a remote failure indication, the switch does not generate Link Fault or Critical Event OAM PDUs. However, if these PDUs are received from a link partner, they are processed. The switch generates and receives Dying Gasp OAM PDUs when Ethernet OAM is disabled, the interface is shut down, the interface enters the error-disabled state, or the switch is reloading. It can respond to, but not generate, Dying Gasp PDUs based on loss of power.

•The switch does not support Ethernet OAM loopback on ports that belong to an EtherChannel, ISL trunk, and promiscuous trunk.

Enabling Ethernet OAM on an Interface

To enable Ethernet OAM on an interface, follow these steps:

configure terminal

interface interface-id

ethernet oam 

ethernet oam [max-rate oampdus | min-rate 
seconds | mode {active | passive} | timeout 
seconds] 

end

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

Enter the no ethernet oam interface configuration command to disable Ethernet OAM on the interface.

This example shows how to set basic OAM parameters on the switch:

Switch(config)# int gi1/3

Switch(config-if)# ethernet oam

Switch(config-if)# ethernet oam max-rate 9

Switch(config-if)# ethernet oam mode passive

Switch(config-if)# end

Switch# show ethernet oam status int gi1/2

GigabitEthernet1/2

General

-------

  Admin state:           enabled

  Mode:                  passive

  PDU max rate:          9 packets per second

  PDU min rate:          1 packet per 1 second

  Link timeout:          5 seconds

  High threshold action: no action

  Link fault action:     no action

  Dying gasp action:     no action

  Critical event action: no action

Link Monitoring

---------------

  Status: supported (on)

  Symbol Period Error

    Window:              100 x 1048576 symbols

    Low threshold:       1 error symbol(s)

    High threshold:      none

Frame Error

    Window:              10 x 100 milliseconds

    Low threshold:       1 error frame(s)

    High threshold:      none

  Frame Period Error

    Window:              1000 x 10000 frames

    Low threshold:       1 error frame(s)

    High threshold:      none

  Frame Seconds Error

    Window:              100 x 100 milliseconds

    Low threshold:       1 error second(s)

    High threshold:      none

  Receive-Frame CRC Error

    Window:              10 x 100 milliseconds

    Low threshold:       10 error frame(s)

    High threshold:      none

  Transmit-Frame CRC Error: Not Supported

Enabling Ethernet OAM Remote Loopback

You must enable Ethernet OAM remote loopback on an interface for the local OAM client to initiate OAM remote loopback operations. Changing this setting causes the local OAM client to exchange configuration information with its remote peer. Remote loopback is disabled by default.

Remote loopback has these limitations:

•Only data packets are looped back.

•You cannot configure Ethernet OAM remote loopback on ISL ports or ports that belong to an EtherChannel.

•Remote loopback can be supported on a max of 16 ports.

To enable Ethernet OAM remote loopback on an interface, follow these steps:

configure terminal

interface interface-id

ethernet oam remote-loopback {supported | 
timeout seconds} 

end

ethernet oam remote-loopback {start | stop} 
{interface interface-id} 

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

Use the no ethernet oam remote-loopback {supported | timeout} interface configuration command to disable remote loopback support or remove the timeout setting.

This example shows how to enable OAM Remote Loopback:

Switch(config)# int gi1/3

Switch(config-if)# ethernet oam

Switch(config-if)# ethernet oam remote-loopback supported

Switch(config-if)# end

Switch# show running int gi1/1

Building configuration...

Current configuration : 209 bytes

!

interface GigabitEthernet1/1

 switchport trunk encapsulation dot1q

 switchport trunk allowed vlan 1,19

 switchport mode trunk

 ethernet oam remote-loopback supported

 ethernet oam

end

Switch# ethernet oam remote-loopback start int gi1/1

This is a intrusive loopback.

Therefore, while you test Ethernet OAM MAC connectivity,

you will be unable to pass traffic across that link.

Proceed with Remote Loopback? [confirm]

Switch# ethernet oam remote-loopback stop int gi1/1

Switch#

*Apr  9 12:52:39.793: %ETHERNET_OAM-6-LOOPBACK: Interface Gi1/1 has exited the master 
loopback mode.

Configuring Ethernet OAM Link Monitoring

You can configure high and low thresholds for link-monitoring features. If no high threshold is configured, the default is none—no high threshold is set. If you do not set a low threshold, the default is a value lower than the high threshold.

Cisco does not generate link event PDUs for rxcrc and trxcrc errors because these are nonstandard.

To configure Ethernet OAM link monitoring on an interface, follow these steps:

configure terminal

interface interface-id

ethernet oam link-monitor supported

ethernet oam link-monitor symbol-period 
{threshold {high {high symbols | none} | low 
{low-symbols}} | window symbols}

Repeat this step to configure both high and 
low thresholds.

ethernet oam link-monitor frame {threshold 
{high {high-frames | none} | low {low-frames}} | 
window milliseconds}

Repeat this step to configure both high and 
low thresholds.

ethernet oam link-monitor frame-period 
{threshold {high {high-frames | none} | low 
{low-frames}} | window frames}

Repeat this step to configure both high and 
low thresholds.

ethernet oam link-monitor frame-seconds 
{threshold {high {high-frames | none} | low 
{low-frames}} | window milliseconds}

Repeat this step to configure both high and 
low thresholds.

ethernet oam link-monitor receive-crc 
{threshold {high {high-frames | none} | low 
{low-frames}} | window milliseconds}

Repeat this step to configure both high and 
low thresholds.

[no] ethernet link-monitor on

end

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

The ethernet oam link-monitor transmit-crc {threshold {high {high-frames | none} | low {low-frames}} | window milliseconds} command is visible on the switch and you can enter it, but it is not supported. Enter the no form of the command to disable the configuration. Use the no form of each command to disable the threshold setting.

Symbol error counters are supported on the following line cards and supervisor cards:

•Supervisor cards: WS-X4515, WS-X4516, WS-X4013+, WS-X4013+TS, WS-X4516-10GE, WS-X4013+10GE

•Line cards: WS-X4148-RJ, WS-X4124-RJ, WS-X4232, WS-X4232-RJ-XX, WS-X4148-RJ21, WS-X4504-FX-MT, WS-X4224-RJ21-XX, WS-X4124-FX-MT, WS-X4232-L3

The rest of the cards do not support symbol error counters.

Configuring Ethernet OAM Remote Failure Indications

You can configure an error-disable action to occur on an interface when the following occur:

•crossing the high thresholds configured on the interface for link monitoring

•on reception of Dying Gasp, executing shut on the interface

•on reception of Dying Gasp, executing reload command

•on reception of Dying Gasp, executing no ethernet oam command on the interface

To enable Ethernet OAM remote-failure indication actions on an interface, follow these steps:

configure terminal

interface interface-id

ethernet oam remote-failure [dying-gasp] 
action error-disable-interface 

end

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

This example shows how to configure Ethernet OAM remote-failure action on the switch interface:

Switch# conf t

Enter configuration commands, one per line.  End with CNTL/Z.

Switch(config)# int gi1/1

Switch(config-if)# ethernet oam remote-failure dying-gasp action error

Switch(config-if)# ethernet oam link-monitor high-threshold action error

Switch(config-if)# end

Switch# show running-config int gi1/1

Building configuration...

Current configuration : 353 bytes

!

interface GigabitEthernet1/1

 switchport trunk encapsulation dot1q

 switchport trunk allowed vlan 1,19

 switchport mode trunk

 ethernet oam remote-loopback supported

 ethernet oam link-monitor high-threshold action error-disable-interface

 ethernet oam remote-failure dying-gasp action error-disable-interface

 ethernet oam

end

Switch# show ethernet oam status int gi1/1

GigabitEthernet1/1

General

-------

  Admin state:           enabled

  Mode:                  active

  PDU max rate:          10 packets per second

  PDU min rate:          1 packet per 1 second

  Link timeout:          5 seconds

  High threshold action: error disable interface

  Link fault action:     no action

  Dying gasp action:     error disable interface

  Critical event action: no action

Link Monitoring

---------------

  Status: supported (on)

  Symbol Period Error

    Window:              100 x 1048576 symbols

    Low threshold:       1 error symbol(s)

    High threshold:      none

  Frame Error

    Window:              10 x 100 milliseconds

    Low threshold:       1 error frame(s)

    High threshold:      none

  Frame Period Error

    Window:              1000 x 10000 frames

    Low threshold:       1 error frame(s)

    High threshold:      none

  Frame Seconds Error

    Window:              100 x 100 milliseconds

    Low threshold:       1 error second(s)

    High threshold:      none

  Receive-Frame CRC Error

    Window:              10 x 100 milliseconds

    Low threshold:       10 error frame(s)

    High threshold:      none

  Transmit-Frame CRC Error: Not Supported

To enable Ethernet OAM failover action on an EtherChannel interface, follow these steps:

configure terminal

interface port-channel interface-id

switchport mode mode

ethernet oam link-monitor high-threshold 
action failover

end

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

The switch does not generate Link Fault or Critical Event OAM PDUs. However, if these PDUs are received from a link partner, they are processed. The switch supports sending and receiving Dying Gasp OAM PDUs when Ethernet OAM is disabled, the interface is shut down, the interface enters the error-disabled state, or the switch is reloading. It can respond to but not generate Dying Gasp PDUs based on loss of power. Enter the no ethernet remote-failure {critical-event | dying-gasp | link-fault} action command to disable the remote failure indication action.

Configuring Ethernet OAM Templates

You can create a template for configuring a common set of options on multiple Ethernet OAM interfaces. The template can be configured to monitor frame errors, frame-period errors, frame-second errors, received CRS errors, and symbol-period errors and thresholds. You can also set the template to put the interface in error-disabled state if any high thresholds are exceeded. These steps are optional and can be performed in any sequence or repeated to configure different options.

To configure an Ethernet OAM template and to associate it with an interface, follow these steps:

configure terminal

template template-name

ethernet oam link-monitor receive-crc 
{threshold {high {high-frames | none} | low 
{low-frames}} | window milliseconds}

ethernet oam link-monitor symbol-period 
{threshold {high {high symbols | none} | low 
{low-symbols}} | window symbols}

ethernet oam link-monitor frame {threshold 
{high {high-frames | none} | low {low-frames}} | 
window milliseconds}

ethernet oam link-monitor frame-period 
{threshold {high {high-frames | none} | low 
{low-frames}} | window frames}

ethernet oam link-monitor frame-seconds 
{threshold {high {high-seconds | none} | low 
{low-seconds}} | window milliseconds}

ethernet oam link-monitor high threshold action 
error-disable-interface 

exit

interface interface-id

source-template template-name

end

show ethernet oam status [interface 
interface-id]

copy running-config startup-config

The switch does not support monitoring egress frames with CRC errors. The ethernet oam link-monitor transmit-crc {threshold {high {high-frames | none} | low {low-frames}} | window milliseconds} command is visible on the switch and you can enter it, but it is not supported. Use the
no form of each command to remove the option from the template. Use the
no source-template template-name command to remove the source template association.

The following example illustrates how to configure an Ethernet OAM template and to associate it with an interface:

Switch# conf t

Switch(config)# template oam

Switch(config-template)# ethernet oam link-monitor receive-crc threshold high 1000

Switch(config-template)# ethernet oam link-monitor receive-crc threshold low 10

Switch(config-template)# ethernet oam link-monitor symbol-period threshold high 5000

Switch(config-template)# ethernet oam link-monitor symbol-period threshold low 5

Switch(config-template)# ethernet oam link-monitor frame threshold high 8000

Switch(config-template)# ethernet oam link-monitor frame threshold low 8

Switch(config-template)# ethernet oam link-monitor frame-period threshold hig 9000

Switch(config-template)# ethernet oam link-monitor frame-period threshold low 9

Switch(config-template)# ethernet oam link-monitor high action error-disable-interface

Switch(config-template)# exit

Switch(config)# int gi1/2

Switch(config-if)# source template oam

Switch(config-if)# end

Switch# show ethernet oam status int gi1/2

GigabitEthernet1/2

General

-------

  Admin state:           enabled

  Mode:                  active

  PDU max rate:          10 packets per second

  PDU min rate:          1 packet per 1 second

  Link timeout:          5 seconds

  High threshold action: error disable interface

  Link fault action:     no action

  Dying gasp action:     no action

  Critical event action: no action

Link Monitoring

---------------

  Status: supported (on)

  Symbol Period Error

    Window:              100 x 1048576 symbols

    Low threshold:       5 error symbol(s)

    High threshold:      5000 error symbol(s)

  Frame Error

    Window:              10 x 100 milliseconds

    Low threshold:       8 error frame(s)

    High threshold:      8000 error frame(s)

  Frame Period Error

    Window:              1000 x 10000 frames

    Low threshold:       9 error frame(s)

    High threshold:      9000 error frame(s)

  Frame Seconds Error

    Window:              100 x 100 milliseconds

    Low threshold:       1 error second(s)

    High threshold:      none

  Receive-Frame CRC Error

    Window:              10 x 100 milliseconds

    Low threshold:       10 error frame(s)

    High threshold:      1000 error frame(s)

  Transmit-Frame CRC Error: Not Supported

Displaying Ethernet OAM Protocol Information

To display Ethernet OAM protocol information, you can use the privileged EXEC commands in Table 55-3.

show ethernet oam discovery [interface 
interface-id]

show ethernet oam statistics [interface 
interface-id]

show ethernet oam status [interface 
interface-id]

show ethernet oam summary 

These examples show how to apply these commands:

Switch# show ethernet oam discovery

GigabitEthernet1/1

Local client

------------

  Administrative configurations:

    Mode:              active

    Unidirection:      not supported

    Link monitor:      supported (on)

    Remote loopback:   supported

    MIB retrieval:     not supported

    Mtu size:          1500

  Operational status:

    Port status:       operational

    Loopback status:   no loopback

    PDU revision:      10

Remote client

-------------

  MAC address: 000f.8f03.3591

  Vendor(oui): 00000C(cisco)

  Administrative configurations:

    PDU revision:      2

    Mode:              active

    Unidirection:      not supported

    Link monitor:      supported

    Remote loopback:   supported

    MIB retrieval:     not supported

    Mtu size:          1500

Switch# show ethernet oam statistics

GigabitEthernet1/1

Counters:

---------

  Information OAMPDU Tx                   : 101163

  Information OAMPDU Rx                   : 51296

  Unique Event Notification OAMPDU Tx     : 0

  Unique Event Notification OAMPDU Rx     : 0

  Duplicate Event Notification OAMPDU TX  : 0

  Duplicate Event Notification OAMPDU RX  : 0

  Loopback Control OAMPDU Tx              : 12

  Loopback Control OAMPDU Rx              : 0

  Variable Request OAMPDU Tx              : 0

  Variable Request OAMPDU Rx              : 0

  Variable Response OAMPDU Tx             : 0

  Variable Response OAMPDU Rx             : 0

  Cisco OAMPDU Tx                         : 7

  Cisco OAMPDU Rx                         : 8

  Unsupported OAMPDU Tx                   : 0

  Unsupported OAMPDU Rx                   : 0

  Frames Lost due to OAM                  : 0

Local Faults:

-------------

  0 Link Fault records

  2 Dying Gasp records

    Total dying gasps       : 7

    Time stamp              : 1d01h

    Total dying gasps       : 6

    Time stamp              : 1d01h

  0 Critical Event records

Remote Faults:

--------------

  0 Link Fault records

  2 Dying Gasp records

    Total dying gasps       : 8

    Time stamp              : 1d01h

    Total dying gasps       : 7

    Time stamp              : 1d01h

  0 Critical Event records

Local event logs:

-----------------

  0 Errored Symbol Period records

  0 Errored Frame records

  0 Errored Frame Period records

  0 Errored Frame Second records

Remote event logs:

------------------

  0 Errored Symbol Period records

  0 Errored Frame records

  0 Errored Frame Period records

  0 Errored Frame Second records

Switch# show ethernet oam summary

Symbols:          * - Master Loopback State,  # - Slave Loopback State

                  & - Error Block State

Capability codes: L - Link Monitor,  R - Remote Loopback

                  U - Unidirection,  V - Variable Retrieval

  Local                       Remote

Interface       MAC Address    OUI    Mode    Capability

  Gi1/1         000f.8f03.3591 00000C active   L R

Ethernet CFM and Ethernet OAM Interaction

You can also configure the OAM Manager infrastructure to interact between CFM and Ethernet OAM. When the Ethernet OAM protocol is running on an interface that has CFM MEPs configured, Ethernet OAM informs CFM of the state of the interface. Interaction is unidirectional from the Ethernet OAM to the CFM protocol, and the only information exchanged is the user network interface port status.

The Ethernet OAM protocol notifies CFM when these conditions occur:

•Error thresholds are crossed at the local interface.

CFM responds to the notification by sending a port status of Local_Excessive_Errors in the Port StatusType Length Value (TLV).

•Ethernet OAM receives an OAM PDU from the remote side showing that an error threshold is exceeded on the remote endpoint.

CFM responds to the notification by sending a port status of Remote_Excessive_Errors in the Port Status TLV.

•The local port is set into loopback mode.

CFM responds by sending a port status of Test in the Port Status TLV.

•The remote port is set into loopback mode.

CFM responds by sending a port status of Test in the Port Status TLV.

For more information about CFM and interaction with Ethernet OAM, see the Ethernet Connectivity Fault Management feature module at this URL:

http://www.cisco.com/en/US/docs/ios/12_2sr/12_2sra/feature/guide/srethcfm.html

Ethernet OAM and CFM Configuration Example

These are configuration example of the interworking between Ethernet OAM and CFM in a sample service provider network. Such a hypothetical network would contain a provider-edge switch connected to a customer edge switch at each endpoint. You must configure CFM, E-LMI, and Ethernet OAM between the customer edge and the provider edge switch.

Customer-edge switch 1 (CE1) configuration:

Switch# config t

Switch(config)# interface GigabitEthernet1/1

Switch(config-if)# switchport trunk allowed vlan 10

Switch(config-if)# switchport mode trunk

Switch(config-if)# ethernet oam remote-loopback supported

Switch(config-if)# ethernet oam

Switch(config-if)# exit

Provider-edge switch 1 (PE1) configuration:

Switch# config t

Switch(config)# interface FastEthernet1/20

Switch(config-if)# switchport trunk encapsulation dot1q

Switch(config-if)# switchport mode trunk

Switch(config-if)# ethernet cfm mip level 7

Switch(config-if)# ethernet cfm mep level 4 mpid 100 vlan 100

Switch(config-if)# ethernet oam remote-loopback supported

Switch(config-if)# ethernet oamt

Provider-edge switch 2 (PE2) configuration:

Switch# config t

Switch(config)# interface GigabitEthernet1/20

Switch(config-if)# switchport mode trunk

Switch(config-if)# ethernet cfm mip level 7

Switch(config-if)# ethernet cfm mep level 4 mpid 101 vlan 10

Switch(config-if)# ethernet oam remote-loopback supported

Switch(config-if)# ethernet oam

Customer-edge switch 2 (CE2) configuration:

Switch# config t

Switch(config)# interface GigabitEthernet1/1

Switch(config-if)# switchport trunk allowed vlan 10

Switch(config-if)# switchport mode trunk

Switch(config-if)# ethernet oam remote-loopback supported

Switch(config-if)# ethernet oam

Switch(config-if)# exit

These output examples show provider-edge switch port status of the configuration. Port status shows as UP at both switches.

Switch PE1:

Switch# show ethernet cfm maintenance points remote

MPID  Level Mac Address    Vlan PortState InGressPort     Age(sec) Service ID

101 * 4     0015.633f.6900 10   UP        Gi1/1         27       blue

Switch PE2:

Switch# show ethernet cfm maintenance points remote

MPID  Level Mac Address    Vlan PortState InGressPort     Age(sec) Service ID

100 * 4     0012.00a3.3780 10   UP        Gi1/1         8        blue

Total Remote MEPs: 1

This example shows the output when you start remote loopback on CE1 (or PE1). The port state on the remote PE switch shows as Test and the remote CE switch enters into error-disable mode.

Switch# ethernet oam remote-loopback start interface gigabitethernet 1/1

This is a intrusive loopback.

Therefore, while you test Ethernet OAM MAC connectivity,

you will be unable to pass traffic across that link.

Proceed with Remote Loopback? [confirm]

Switch PE1:

Switch# show ethernet cfm maintenance points remote

MPID  Level Mac Address    Vlan PortState InGressPort     Age(sec) Service ID

101 * 4     0015.633f.6900 10   UP        Gi1/1         27       blue

Switch PE2:

Switch# show ethernet cfm maintenance points remote

MPID  Level Mac Address    Vlan PortState InGressPort     Age(sec) Service ID

100 * 4     0012.00a3.3780 10   TEST      Gi1/1         8        blue

Total Remote MEPs: 1

In addition, if you shut down the CE1 interface that connects to PE1, the remote PE2 port shows a PortState of Down.