Implementing BFD

Bidirectional forwarding detection (BFD) provides low-overhead, short-duration detection of failures in the path between adjacent forwarding engines. BFD allows a single mechanism to be used for failure detection over any media and at any protocol layer, with a wide range of detection times and overhead. The fast detection of failures provides immediate reaction to failure in the event of a failed link or neighbor.

Cisco NCS 5500 Router supports BFD with VRF context.


Note


On the Cisco NCS 5500 Router, BFD supports up to 4 different minimum interval timers if BFD over Bundle (BoB) is configured, and up to 5 different minimum interval timers if BoB is not configured.



Tip


You can programmatically configure BFD and retrieve operational data using openconfig-bfd.yang OpenConfig data model. To get started with using data models, see the Programmability Configuration Guide.


BFD Overview

Table 1. Feature History Table

Feature Name

Release Information

Feature Description

BFD Support for VRRP

Release 7.5.1

This feature is now supported on routers that have the Cisco NC57 line cards installed and operate in native and compatible modes.

This feature introduces BFD support over VRRP interfaces.

Bidirectional forwarding detection (BFD) provides low-overhead, short-duration detection of failures in the path between adjacent routers. BFD allows a single mechanism to be used for failure detection over any media and at any protocol layer, with a wide range of detection times and overhead. The fast detection of failures provides immediate reaction to failure in the event of a failed link or neighbor.

The router supports BFD with VRF context.

Restrictions

These restrictions apply to BFD:

  • Demand mode is not supported in Cisco IOS XR software.

  • BFD echo mode and encryption are not supported.

  • BFD hardware offload for IPv4 is supported.

  • Only the static, OSPF, BGP and IS-IS applications are supported on BFD.

  • BFD dampening for IPv4 is supported starting from Cisco IOS XR Release 6.3.2.

  • Starting IOS XR Release 6.3.2, BFD multihop over an IP core is supported.

  • BFD supports BFDv6 on bundle-ether for VRF BGP IPv6 single-hop.

  • BFD multihop over non-IP core (Label Distribution Protocol or Segment Routing) is supported starting from IOS XR Release 7.1.1.

  • BFD over bundle feature is supported only in IETF mode.

  • Dampening extensions for BFD are not supported.

  • Starting from Cisco IOS XR Release 6.6.1, BFD over VRF is supported.

  • Starting from Cisco IOS XR Release 7.1.1, BFD over BVI is supported only on fixed NCS5500 platforms.

  • Starting from Cisco IOS XR Release 7.2.1, BFD support over VRRP interface is supported.

  • Egress IPv4 ACLs block all traffic, including router-generated traffic for the following routers and line cards:

    • NC57-24DD

    • NC57-18DD-SE

    • NC57-36H-SE

    • NC57-36H6D-S

    • NC57-MOD-S

    • NCS-57B1-6D24-SYS

    • NCS-57B1-5DSE-SYS

    For all other routers and line cards, egress IPv4 ACLs do not block certain router-generated traffic, such as ICMP messages.

BFD Timers


Note


The router uses six unique timer profiles without BFD over Bundle (BoB) configuration. Up to five timer profiles are available, when you configure BoB.

The following example shows configuration of four profiles with minimum-interval timer.

  • 150ms *3

  • 450ms *3

  • 200ms *3

  • 750ms *3


Table 2. IPv4 BFD Timers

Type of BFD Session

Minimum Timer Supported

Default/Minimum-interval timer (Multipliers)

Supported Timer Profiles

(Up to 6 unique timers profiles)

Single Hop

4ms

3

Any

BFD over Bundle Members (BoB)

4ms

3

Any

BFD over Logical bundle (BLB)

50ms

3

Any

BGP Multi Hop

50ms

3

Any

Table 3. IPv6 BFD Timers

Type of BFD Session

Minimum Timer Supported

Default/Minimum- interval timer (Multipliers)

Supported Timer Profile

(Up to 6 unique timer profiles)

Maximum Scale depending on Minimum Interval

Single Hop

4ms

3

Any

150 (with 8ms and above, all 256 sessions are configurable)

BFD over Bundle Members (BoB)

4ms

3

Any

150ms (with 8ms and above, all 256 sessions are configurable)

BFD over Logical bundle (BLB)

50ms

3

Any

256

BGP Multi Hop

50ms

3

Any

256

Enabling BFD on a Static Route

The following procedure describes how to enable BFD on a static route.

RP/0/RP0/CPU0:router(config)# configure

/* Enter static route configuration mode, and configure static routing. */
Router(config)# router static

/* Enter address family configuration mode. */
Router(config-static)# address-family ipv4 unicast 

/* Specify an unicast destination address and next-hop IPv4 address. 
Enable BFD fast-detection on the specified IPv4 unicast destination address */
Router(config-static)# 192.168.2.2/32 HundredGigE0/0/0/2 192.168.6.2 bfd fast-detect minimum-interval 4 multiplier 3 


Configuration example of a BFD single-hop scenario:

!
router  
 address-family ipv4 unicast
  10.2.150.193/32 BVI1252 10.2.153.1 bfd fast-detect

Note


The next-hop IPv4 address (10.2.153.1) is determined from the IP address of the directly connected interface.


Configuration example of a BFD multi-hop scenario:

!
router static
address-family ipv4 unicast
 10.10.10.0/24 10.20.20.20 bfd fast-detect multihop 10.30.30.30

Note


In a BFD multi-hop scenario:

  • The next-hop (10.20.20.20) must be reachable.

  • The next-hop must be resolved through the connected interface if the interface is provided as part of the configuration. If the interface is provided, the 'multihop' option will not be available.


Configuration example of BFD on directly-connected host routes (/32 or /128):

!
router static
address-family ipv4 unicast
 10.102.134.140/32 TenGigE0/0/0/1 10.102.134.140 bfd fast-detect

Note


In this BFD single-hop specific scenario:

  • The prefix and next-hop are the same (this makes nexthop a static nexthop). The prefix length is either 32 or 128.

  • Next-hop (10.102.134.140) is resolved using static routing.


Running Configuration

router static
 address-family ipv4 unicast
  192.168.2.2/32 HundredGigE0/0/0/2 192.168.6.2 bfd fast-detect minimum-interval 4 multiplier 3
 ! 
!  

Enabling BFD for OSPF on an Interface

The following procedures describe how to configure BFD for Open Shortest Path First (OSPF) on an interface. The steps in the procedure are common to the steps for configuring BFD on IS-IS ; only the command mode differs.

SUMMARY STEPS

  1. configure
  2. router ospf process-name
  3. area area-id
  4. interface type interface-path-id
  5. bfd fast-detect
  6. bfd minimum-interval milliseconds
  7. bfd multiplier multiplier
  8. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

router ospf process-name

Example:


RP/0/RP0/CPU0:router(config)# router ospf 0

Enters OSPF configuration mode, allowing you to configure the OSPF routing process.

Note

 
To configure BFD for IS-IS, enter the corresponding configuration mode.

Step 3

area area-id

Example:


RP/0/RP0/CPU0:router(config-ospf)# area 0 

Configures an Open Shortest Path First (OSPF) area.

Replace area-id with the OSPF area identifier.

Step 4

interface type interface-path-id

Example:


RP/0/RP0/CPU0:router(config-ospf-ar)# interface TengigabitEthernet 0/3/0/1 

Enters interface configuration mode and specifies the interface name.

Step 5

bfd fast-detect

Example:


RP/0/RP0/CPU0:router(config-ospf-ar-if)# bfd fast-detect 

Enables BFD to detect failures in the path between adjacent routers.

Step 6

bfd minimum-interval milliseconds

Example:


RP/0/RP0/CPU0:router(config-ospf-ar-if)# bfd minimum-interval 6500

Sets the BFD minimum interval. Range is 4-30000 milliseconds.

This example sets the BFD minimum interval to 6500 milliseconds.

Step 7

bfd multiplier multiplier

Example:


RP/0/RP0/CPU0:router(config-ospf-ar-if)# bfd multiplier 7

Sets the BFD multiplier. This is optional, the minimum is 3 and by default the multiplier will be 3 for all protocols.

This example sets the BFD multiplier to 7.

Step 8

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Enable BFD for IS-IS on an Interface

Perform the following steps to configure BFD for Integrated Intermediate System-to-Intermediate System (IS-IS) on an interface.


Note


BFD per interface configuration is supported for OSPF and IS-IS only.



Router# configure

/* Enter IS-IS configuration mode to configure the IS-IS routing process. */
Router(config)# router isis 65444

/* Set the system type (area or backbone router). Each IS-IS instance can support either a single Level 1 or Level 2, or one of each.*/
Router(config-isis)#is-type level-2-only

/* Specify a NET for each routing instance if you are configuring multi-instance IS-IS.*/
Router(config-isis)# net 49.0001.0840.3803.4088.00

/* Enter interface configuration mode. */
Router(config-isis)# interface gigabitEthernet 0/3/0/1

/* Set the BFD minimum interval. */
Router(config-isis-if)# bfd minimum-interval 6500

/* Set the BFD multiplier.  */
Router(config-isis-if)# bfd multiplier 7

/* Enable BFD to detect failures in the path between adjacent forwarding engines. Only IPv4 is supported.*/
Router(config-isis-if)# bfd fast-detect ipv4

/* Specify the IPv4 address family and enters router address family configuration mode. */
Router(config-isis-if)# address-family ipv4 unicast

!

Running Configuration


configure 
  router isis 65444
   is-type level-2-only
   net 49.0001.0840.3803.4088.00
    interface gigabitEthernet 0/3/0/1
     bfd minimum-interval 6500
     bfd multiplier 7
     bfd fast-detect ipv4
     address-family ipv4 unicast
  !
!

Enabling BFD on a BGP Neighbor

BFD can be enabled per neighbor, or per interface. This task describes how to enable BFD for BGP on a neighbor router.

SUMMARY STEPS

  1. configure
  2. router bgp autonomous-system-number
  3. neighbor ip-address
  4. remote-as autonomous-system-number
  5. bfd fast-detect
  6. bfd minimum-interval milliseconds
  7. bfd multiplier multiplier
  8. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

router bgp autonomous-system-number

Example:


RP/0/RP0/CPU0:router(config)# router bgp 120

Enters BGP configuration mode, allowing you to configure the BGP routing process.

Step 3

neighbor ip-address

Example:


RP/0/RP0/CPU0:router(config-bgp)# neighbor 172.168.40.24

Places the router in neighbor configuration mode for BGP routing and configures the neighbor IP address as a BGP peer.

This example configures the IP address 172.168.40.24 as a BGP peer.

Step 4

remote-as autonomous-system-number

Example:


RP/0/RP0/CPU0:router(config-bgp-nbr)# remote-as 2002

Creates a neighbor and assigns it a remote autonomous system.

This example configures the remote autonomous system to be 2002.

Step 5

bfd fast-detect

Example:


RP/0/RP0/CPU0:router(config-bgp-nbr)# bfd fast-detect

Enables BFD between the local networking devices and the neighbor whose IP address you configured to be a BGP peer in Step 3.

In the example in Step 3, the IP address 172.168.40.24 was set up as the BGP peer. In this example, BFD is enabled between the local networking devices and the neighbor 172.168.40.24.

Step 6

bfd minimum-interval milliseconds

Example:


RP/0/RP0/CPU0:router(config-bgp-nbr)#bfd minimum-interval 6500

Sets the BFD minimum interval. Range is 4-30000 milliseconds.

Step 7

bfd multiplier multiplier

Example:


RP/0/RP0/CPU0:router(config-bgp-nbr)#bfd multiplier 7

Sets the BFD multiplier. This is optional, the minimum is 3 and by default the multiplier will be 3 for all protocols

Step 8

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Clear and Display BFD Counters

The following procedure describes how to display and clear BFD packet counters. You can clear packet counters for BFD sessions that are hosted on a specific node or on a specific interface.


RP/0/RP0/CPU0:router# show bfd counters all packet location 0/3/cpu0
RP/0/RP0/CPU0:router# clear bfd counters all packet location 0/3/cpu0
RP/0/RP0/CPU0:router# show bfd counters all packet location 0/3/cpu0

BFD over Bundle and BFD over Logical Bundle

Link Aggregation Control Protocol (LACP) allows a network device to negotiate an automatic bundling of links by sending LACP packets to their directly connected peer. LACP provides a keep-alive mechanism for the link members. While the default keep-alive is 30s, it is configurable to up to 1s. LACP can detect failures on a per-physical-member link. However, the LACP timers do not fulfill the criteria of current fast convergence requirements.

Differences between BFD over Bundle and BFD over Logical Bundle

BFD over Bundle (BoB) (RFC 7130) has a BFD session on each bundle member. The client is the bundle manager. If a BFD session goes down on a specific member link, the whole bundle interface goes down. That is, when the member link goes down, the number of available links falls below the required minimum. Hence the routing session is brought down.

BFD over Logical Bundle (BLB) (RFC 5880) treats a bundle interface with all its members as a single interface. BLB is a multipath (MP) single-hop session. If BLB is configured on a bundle there is only one single BFD session that is active. This implies that only one bundle member is being monitored by BFD at any given time. The client is one of the routing protocols. When BFD detects a failure, the client brings down the routing session.

The mode (BoB or BLB) is determined by how you configure BFD:

  • You can enable BoB by configuring BFD under a Bundle-Ether interface.

  • You can enable BLB by configuring BFD under a Bundle-Ether interface on a routing client.

Link Aggregation Control Protocol (LACP) allows a network device to negotiate an automatic bundling of links by sending LACP packets to their directly connected peer. LACP provides a keep-alive mechanism for the link members. While the default keep-alive is 30s, it is configurable to up to 1s. LACP can detect failures on a per-physical-member link. However, the LACP timers do not fulfill the criteria of current fast convergence requirements.

BFD over Bundle

BFD over Bundle

BFD Over Bundle (BoB) (RFC 7130) has a BFD session on each bundle member. BOB verifies the ability for each member link to be able to forward Layer 3 packets.

For BFD over Bundle, the BFD client is bundlemgr. When BFD detects a failure on a bundle member, bundlemgr removes that member from the bundle. If there are not enough members to keep the bundle up, then the main Bundle-Ether interface will go down so that all routing protocols running on the main bundle interface or a subinterface will detect an interface down.

BoB does not provide a true Layer 3 check and is not supported on subinterfaces. However, subinterfaces will go down at the same time as the main interface.

BoB implementation is a standard-based fast failure detection of link aggregation (LAG) member links that is interoperable between different platforms. NCS 5500 platforms only support the IETF mode.

Configure BFD Over Bundle

Perform the following tasks to configure the BOB feature:

  • Enable BFD sessions on bundle members

  • Specify the BFD destination address on a bundle

  • Configure the minimum thresholds for maintaining an active bundle

  • Configure BFD packet transmission intervals and failure detection times on a bundle

    Configure BFD over bundles IETF mode support on a per-bundle basis


/* Enable BFD sessions on bundle members */
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd address-family ipv4 fast-detect
Router(config-if)# bfd mode ietf

/* Specify the BFD destination address on a bundle */
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd address-family ipv4 destination 10.20.20.1
	 
/* Configure the minimum thresholds for maintaining an active bundle */
Router(config)# interface Bundle-Ether 1
Router(config-if)# bundle minimum-active bandwidth 580000
Router(config-if)# bundle minimum-active links 2

/* Configure BFD packet transmission intervals and failure detection times on a bundle */
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd address-family ipv4 minimum-interval 2000
Router(config-if)# bfd address-family ipv4 multiplier 30

/* Configure BFD over bundles IETF mode support on a per-bundle basis */
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd mode ietf
Router(config-if)# bfd address-family ipv4 fast-detect

Bidirectional Forwarding Detection over Logical Bundle

Table 4. Feature History Table

Feature Name

Release Information

Feature Description

Coexistence of BFD over bundle and BFD over logical bundle

Release 7.4.1

This feature provides the benefits of both BFD over bundle (BOB) and BFD over logical bundle (BLB). This feature enables you to configure both BOB and BLB over physical bundle interfaces and subinterfaces. BOB functionality allows you to detect failures in physical bundle interfaces. BLB functionality allows you to detect failures in the client protocols configured on the subinterfaces.

Table 5. Feature History Table

Feature Name

Release Information

Feature Description

BFD over Logical Bundle

Release 7.4.1

This feature is now supported on routers that have Cisco NC57 line cards installed and operate in compatibility mode.

The BLB feature implements and deploys BFD over bundle VLAN interfaces. The advantage of BFD, low-overhead and short-duration detection of path failures between routers, is extended to bundle VLAN interfaces.

BFD over Logical Bundle

The BLB feature implements and deploys BFD over bundle interfaces based on RFC 5880. In the BLB, the bundle interface is a single interface, whereas, in BOB, BFD is implemented per member link. BLB is a multipath (MP) single-hop session so at least one line card must be configured under the bfd multipath command before a BLB session can come up. Because BFD treats the bundle as a single big interface, BLB requires limited knowledge of the bundle interfaces on which the sessions run. BLB requires information about IP addresses, interface types, and caps on bundle interfaces only. Information such as a list of bundle members, member states, and configured minimum or maximum bundle links are not required. In the case of BLB, the BFD client is not the bundle link but protocols running over the bundle link. In BLB, the BFD client is not bundlemgr but the protocols running over bundle link. BLB is supported on IPv4 address, IPv6 global address, and IPv6 link-local address. The current version of the software supports a total of 200 sessions (which includes BFD Single hop for physical and logical sub-interfaces; BFD over Bundle (BoB) and BLB) per line card. The maximum processing capability of BFD control packets, per line card, has also increased to 7000 pps.

Configuration Example

  • Create VLAN subinterface under bundle interface

  • Enable BFD on a static route

  • Enable BFD on IS-IS

  • Enable BFD for OSPF on an interface

  • Enable BFD on a BGP neighbor

  • Configure multipath capability under BFD

/* Create VLAN subinterface under bundle interface */
Router# configure 
Router(config)# interface Bundle-Ether 2.1 
Router(config-if)# ipv4 address 10.1.1.1 255.255.255.0 
Router(config-if)# encapsulation dot1q 1 
Router(config-if)# end 
 
/* Enable BFD on a static route. */
Router# configure 
Router(config)# router static 
Router(config-static)# address-family ipv4 unicast 
Router(config-static)# 10.158.3.13/32 10.1.1.2 bfd fast-detect minimum-interval 300 multiplier 3 

/* Enable BFD on IS-IS. */
Router# configure 
Router(config)# router isis cybi 
Router(config-isis)# interface Bundle-Ether 2.1 
Router(config-isis-if)# bfd minimum-interval 300 
Router(config-isis-if)# bfd multiplier 3 
Router(config-isis-if)# bfd fast-detect ipv4 
Router(config-isis-if)# address-family ipv4 unicast 
Router(config-isis-if-af)# end 

/* Enable BFD for OSPF on an interface. */
Router# configure 
Router(config)# router ospf cybi 
Router(config-ospf)# area 0 
Router(config-ospf)# interface Bundle-Ether 2.1 
Router(config-ospf-if)# bfd fast-detect 
Router(config-ospf-if)# bfd minimum-interval 300 
Router(config-ospf-if)# bfd multiplier 3 
Router(config-ospf-if)# end 

/* Enable BFD on a BGP neighbor.*/
Router# configure
Router(config)# router bgp 4787 
Router(config-bgp)# neighbor 10.158.1.1 
Router(config-bgp-nbr)# remote-as 4787 
Router(config-bgp-nbr)# update-source Bundle-Ether 2.1 
Router(config-bgp-nbr)# bfd fast-detect 
Router(config-bgp-nbr)# bfd minimum-interval 300 
Router(config-bgp-nbr)# bfd multiplier 3 
Router(config-bgp-nbr)# address-family ipv4 unicast 
Router(config-bgp-nbr-af)# route-policy PASS-ALL in 
Router(config-bgp-nbr-af)# route-policy PASS-ALL out 
Router(config-bgp-nbr-af)# exit 
Router(config-bgp-nbr)# commit 

/* Configure a specific LC (or LCs) to host BLB sessions. The BLB sessions and bundle member links need not be configured on the same LC. For example, you can configure the bundle member links on LC slot 2 and slot 3 while you configure BLB sessions to be hosted on LC slot 5. */
Router(config)# bfd 
Router(config-bfd)# multipath include location 0/6/CPU0 
Router(config-bfd)# multipath include location 0/2/CPU0 

Enabling BFD Sessions on Bundle Members

To enable BFD sessions on bundle member links, complete these steps:

SUMMARY STEPS

  1. configure
  2. interface Bundle-Ether bundle-id
  3. bfd address-family ipv4 fast-detect
  4. bfd mode ietf
  5. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

interface Bundle-Ether bundle-id

Example:


RP/0/RP0/CPU0:router(config)# interface Bundle-Ether 1

Enters interface configuration mode for the specified bundle ID.

Step 3

bfd address-family ipv4 fast-detect

Example:


RP/0/RP0/CPU0:router(config-if)# bfd address-family ipv4 fast-detect

Enables IPv4 BFD sessions on bundle member links.

Step 4

bfd mode ietf

Example:


RP/0/RP0/CPU0:router(config-if)# bfd mode ietf

Enables IETF mode for BFD over bundle for the specified bundle.

Step 5

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Specifying the BFD Destination Address on a Bundle

To specify the BFD destination address on a bundle, complete these steps:

SUMMARY STEPS

  1. configure
  2. interface Bundle-Ether bundle-id
  3. bfd address-family ipv4 destination ip-address
  4. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

interface Bundle-Ether bundle-id

Example:


RP/0/RP0/CPU0:router(config)# interface Bundle-Ether 1

Enters interface configuration mode for the specified bundle ID.

Step 3

bfd address-family ipv4 destination ip-address

Example:


RP/0/RP0/CPU0:router(config-if)# bfd address-family ipv4 destination 10.20.20.1

Specifies the primary IPv4 address assigned to the bundle interface on a connected remote system, where ip-address is the 32-bit IP address in dotted-decimal format (A.B.C.D).

Step 4

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Configuring the Minimum Thresholds for Maintaining an Active Bundle

The bundle manager uses two configurable minimum thresholds to determine whether a bundle can be brought up or remain up, or is down, based on the state of its member links.

  • Minimum active number of links

  • Minimum active bandwidth available

Whenever the state of a member changes, the bundle manager determines whether the number of active members or available bandwidth is less than the minimum. If so, then the bundle is placed, or remains, in DOWN state. Once the number of active links or available bandwidth reaches one of the minimum thresholds, then the bundle returns to the UP state.

To configure minimum bundle thresholds, complete these steps:

SUMMARY STEPS

  1. configure
  2. interface Bundle-Ether bundle-id
  3. bundle minimum-active bandwidth kbps
  4. bundle minimum-active links links
  5. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

interface Bundle-Ether bundle-id

Example:


RP/0/RP0/CPU0:router(config)# interface Bundle-Ether 1

Enters interface configuration mode for the specified bundle ID.

Step 3

bundle minimum-active bandwidth kbps

Example:


RP/0/RP0/CPU0:router(config-if)# bundle minimum-active bandwidth 580000

Sets the minimum amount of bandwidth required before a bundle can be brought up or remain up. The range is from 1 through a number that varies depending on the platform and the bundle type.

Step 4

bundle minimum-active links links

Example:


RP/0/RP0/CPU0:router(config-if)# bundle minimum-active links 2

Sets the number of active links required before a bundle can be brought up or remain up. The range is from 1 to 32.

Note

 
When BFD is started on a bundle that is already active, the BFD state of the bundle is declared when the BFD state of all the existing active members is known.

Step 5

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Configuring BFD Packet Transmission Intervals and Failure Detection Times on a Bundle

BFD asynchronous packet intervals and failure detection times for BFD sessions on bundle member links are configured using a combination of the bfd address-family ipv4 minimum-interval and bfd address-family ipv4 multiplier interface configuration commands on a bundle.

The BFD control packet interval is configured directly using the bfd address-family ipv4 minimum-interval command. The failure detection times are determined by a combination of the interval and multiplier values in these commands.

To configure the minimum transmission interval and failure detection times for BFD asynchronous mode control packets on bundle member links, complete these steps:

SUMMARY STEPS

  1. configure
  2. interface Bundle-Ether bundle-id
  3. bfd address-family ipv4 minimum-interval milliseconds
  4. bfd address-family ipv4 multiplier multiplier
  5. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

interface Bundle-Ether bundle-id

Example:


RP/0/RP0/CPU0:router(config)# interface Bundle-Ether 1

Enters interface configuration mode for the specified bundle ID.

Step 3

bfd address-family ipv4 minimum-interval milliseconds

Example:


RP/0/RP0/CPU0:router(config-if)#bfd address-family ipv4 minimum-interval 2000

Note

 
Specifies the minimum interval, in milliseconds, for asynchronous mode control packets on IPv4 BFD sessions on bundle member links. The range is from 4 to 30000.

Step 4

bfd address-family ipv4 multiplier multiplier

Example:


RP/0/RP0/CPU0:router(config-if)#bfd address-family ipv4 multiplier 30

Specifies a number that is used as a multiplier with the minimum interval to determine BFD control packet failure detection times and transmission intervals for IPv4 BFD sessions on bundle member links. The range is from 2 to 50. The default is 3.

Note

 
Although the command allows you to configure a minimum of 2, the supported minimum is 3.

Step 5

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Configure BFD over Bundles IETF Mode Support on a Per Bundle Basis

To configure BFD over Bundles IETF mode support on a per bundle basis use these steps:

SUMMARY STEPS

  1. configure
  2. interface Bundle-Ether bundle-id
  3. bfd mode ietf
  4. bfd address-family ipv4 fast-detect
  5. Use the commit or end command.
  6. show bundle bundle-ether bundle-id

DETAILED STEPS

  Command or Action Purpose

Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters global configuration mode.

Step 2

interface Bundle-Ether bundle-id

Example:


RP/0/RP0/CPU0:router(config)# interface Bundle-Ether 1

Enters interface configuration mode for the specified bundle ID.

Step 3

bfd mode ietf

Example:


RP/0/RP0/CPU0:router(config-if)# bfd mode ietf

Enables IETF mode for BFD over bundle for the specified bundle.

Step 4

bfd address-family ipv4 fast-detect

Example:


RP/0/RP0/CPU0:router(config-if)# bfd address-family ipv4 
fast-detect

Enables IPv4 BFD sessions on the specified bundle.

Step 5

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Step 6

show bundle bundle-ether bundle-id

Displays the selected bundle mode.

BFD over Bundle with IPv4 Unnumbered Interfaces

Table 6. Feature History Table

Feature Name

Release Information

Feature Description

BFD for BoB with IPv4 Unnumbered

Release 7.3.1

This feature is now supported on routers that have Cisco NC57 line cards installed and operate in the native mode.

BFD over Bundle with IPv4 Unnumbered Interfaces feature enables BFD to run on IP unnumbered interfaces, which take the IP address from the loopback address. The same loopback address is used on multiple interfaces. This saves IP addresses space or range.

BFD creates a session on the unnumbered interface for which the BFD clients provide the source and destination IP address along with the interface index. BFD establishes the session on the Layer 3 unnumbered link to which the interface index corresponds. The source address is derived from the Loopback interface at the source. The destination node also uses IP unnumbered interface with loopback address and that is used as destination IP address.

BFD sends control packets to the unnumbered interfaces. These control packets are the regular IP BFD packets. Address Resolution Protocol (ARP) resolves the destination loopback IP address to the destination node’s router MAC address.

Restriction

Only Asynchronous mode is supported.

Configure BFD over Bundle with IPv4 Unnumbered Interface

  • Configure loopback address

  • Add physical interface to bundle

  • Configure BOB session on an unnumbered interface

Configure Loopback Address

Router(config)# interface loopback 1
Router(config-if)# ipv4 address 10.1.1.1 255.255.255.0

Add Physical Interface to Bundle

Router(config)# interface HundredGigE0/0/1/0
Router(config-if)# bundle id 1 mode on

Configure a BFD over Bundle Session on an Unnumbered Interface

Router(config)# interface Bundle-Ether1
Router(config-if)# bfd address-family ipv4 destination 10.2.2.2
Router(config-if)# bfd address-family ipv4 fast-detect
Router(config-if)# ipv4 point-to-point
Router(config-if)# ipv4 unnumbered Loopback1
Router(config-if)# lldp
Router(config-if)# enable

Running Configuration


interface Loopback1
ipv4 address 10.1.1.1 255.255.255.0
!
interface HundredGigE0/0/1/0
bundle id 1 mode on
!
interface Bundle-Ether1
bfd address-family ipv4 destination 10.2.2.2
bfd address-family ipv4 fast-detect
ipv4 point-to-point
ipv4 unnumbered Loopback1
lldp
 enable

BFD Transparency

Bidirectional Forwarding Detection(BFD) protocol is a simple hello mechanism that detects failures in a network in less than one second, depending on the timer value configured.

Both endpoints of a BFD Session periodically send Hello packets to each other. If a number of those packets are not received, the session is considered down. BFD provides fast BFD peer failure detection times independently of all media types, encapsulations, topologies, and routing protocols BGP, IS-IS, and OSPF.

BFD Transparency feature enables you to configure BFD Sessions between customer edge devices connected over an L2VPN network. These BFD sessions are transparent to the core. For example, BFD packets being exchanged between CEs are neither dropped on any router in the core, nor punted on any core device.

In this section, you will learn how to configure BFD Transparency in Ethernet VPN (EVPN) Virtual Private Wire Service (VPWS).

Ethernet VPN Virtual Private Wire Service

EVPN VPWS (Ethernet VPN Virtual Private Wire Service) is a BGP control plane solution for point-to-point services. It implements signaling and encapsulation techniques for establishing an EVPN instance between a pair of provider edge devices.

EVPN VPWS supports both single-homing and multi-homing.

Configuration

The following sections describes the procedure for configuring IP Fast Reroute with Remote LFA.

  • Configure L2VPN on the provide edge router

  • Configure BFD on the customer edge router

Configure L2VPN on the Provide Edge Router

Configure L2VPN on the provider edge router.


/* Enable IS-IS and configure routing level for an area. */
RP/0//CPU0:router# configure
RP/0//CPU0:router(config)# interface tengige 0/1/0/8/2.1
RP/0//CPU0:router(config-subif)# exit
RP/0//CPU0:router(config)# router isis
RP/0//CPU0:router(config-isis)# is-type level-2-only
RP/0//CPU0:router(config-isis)# net 49.1234.2222.2222.2222.00
RP/0//CPU0:router(config-isis)# nsr
RP/0//CPU0:router(config-isis)# nsf cisco
RP/0//CPU0:router(config-isis)# address-family ipv4 unicast
RP/0//CPU0:router(config-isis-af)# metric style wide
RP/0//CPU0:router(config-isis)# end
RP/0//CPU0:router(config)# interface Bundle-Ether 199
RP/0//CPU0:router(config-if)# address-family ipv4 unicast
RP/0//CPU0:router(config-if)# end
RP/0//CPU0:router(config)# interface Loopback 0
RP/0//CPU0:router(config-if)# end
RP/0//CPU0:router(config-if)# address-family ipv4 unicast
RP/0//CPU0:router(config-if)# exit

/* Configure L2VPN EVPN address family. */
RP/0//CPU0:router(config)# router bgp 100
RP/0//CPU0:router(config-bgp)# bgp router-id 10.10.10.1
RP/0//CPU0:router(config-bgp)# address-family l2vpn evpn
RP/0//CPU0:router(config-bgp)# neighbor 192.0.2.1
RP/0//CPU0:router(config-bgp-nbr)# remote-as 100
RP/0//CPU0:router(config-bgp-nbr)# update-source Loopback 0
RP/0//CPU0:router(config-bgp-nbr)# address-family l2vpn evpn

/* Configure MPLS LDP for the physical core interface. */
RP/0//CPU0:router(config-bgp-nbr-af)# mpls ldp
RP/0//CPU0:router(config-bgp-nbr-af)# exit
RP/0//CPU0:router(config-bgp-nbr)# exit
RP/0//CPU0:router(config-bgp)# exit
RP/0//CPU0:router(config)# interface Bundle-Ether 199
RP/0//CPU0:router(config-if)# exit

/* Configure L2VPN Xconnect. */ 
RP/0//CPU0:router(config)# l2vpn 
RP/0//CPU0:router(config-l2vpn)# router-id 10.10.10.1 
RP/0//CPU0:router(config-l2vpn)# xconnect group bfdtr
RP/0//CPU0:router(config-l2vpn-xc)# p2p vpws-ce
RP/0//CPU0:router(config-l2vpn-xc-p2p)# interface TenGigE 0/4/0/2/0.1
RP/0//CPU0:ios(config-l2vpn-xc-p2p)# neighbor evpn evi 100 target 3 source 4

Configure BFD on the Customer Edge Router

Configure BFD on the customer edge router.


RP/0//CPU0:router# configure
RP/0//CPU0:router(config)#  router bgp 100
RP/0//CPU0:router(config-bgp)# bgp router-id 10.10.10.1
RP/0//CPU0:router(config-bgp)# address-family ipv4 unicast
RP/0//CPU0:router(config-bgp-af)# exit
RP/0//CPU0:router(config-bgp)# neighbor 172.16.0.1
RP/0//CPU0:router(config-bgp)# address-family ipv4 unicast
RP/0//CPU0:router(config-bgp-nbr)# remote-as 100
RP/0//CPU0:router(config-bgp-nbr)# bfd fast-detect
RP/0//CPU0:router(config-bgp-nbr)# bfd multiplier 2
RP/0//CPU0:router(config-bgp-nbr)# bfd minimum-interval 100
RP/0//CPU0:router(config-bgp-nbr)# update-source TenGigE 0/0/0/16.1
RP/0//CPU0:router(config-bgp-nbr)# address-family ipv4 unicast
RP/0//CPU0:router(config-bgp-nbr-af)# 

Running Configuration

This section shows the BFD Transparency configuration.


!
interface TenGigE 0/4/0/2/0.1
  l2transport
router isis 1
   is-type level-2-only
   net 49.0000.1000.0000.0001.00
   nsr
   nsf cisco
   address-family ipv4 unicast
    metric-style wide
! 
  interface Bundle-Ether199
     address-family ipv4 unicast
  interface Loopback0
      address-family ipv4 unicast
router bgp 100
  bgp router-id 10.10.10.1
   address-family l2vpn evpn
    neighbor 192.0.2.1
     remote-as 100
     update-source Loopback 0
    address-family l2vpn evpn
 !   
   mpls ldp
   interface Bundle-Ether199
 !
  l2vpn
   router-id 10.10.10.1 
    xconnect group bfdtr
   p2p vpws-ce
    interface TenGigE 0/4/0/2/0.1
     neighbor evpn evi 100 target 3 source 4
 
         router bgp 100
          bgp router-id 10.10.10.1
           address-family ipv4 unicast
           !
           neighbor 172.16.0.1
            address-family ipv4 unicast
            remote-as 100
            bfd fast-detect
            bfd multiplier 2
            bfd minimum-interval 100
            update-source TenGigE0/0/0/16.1
            address-family ipv4 unicast

Verification

The show outputs given in the following section display the details of the configuration of the BFD transparency, and the status of their configuration.


/* Verify if the BFD session is up, and the timers are configured. */ 
RP/0//CPU0:router# show bfd session    

Thu Jan  4 03:07:15.529 UTC
Interface      Dest Addr  Local det time(int*mult)  State        Echo  Async   H/W      NPU     
-------------- ---------- ------------------------  --------     ----  -----   ---      ----
Te0/0/0/4.1    10.1.1.1   0s(0s*0)                  20ms(10ms*2) UP    Yes     0/RP0/CPU0                                                      Yes   0/RP0/CPU0


/* Verify if the BFD session is up and check the timer value, numbers of hellos exchanged, and information 
about last packet. */

RP/0//CPU0:router# show bfd session destination 10.1.1.1 detail
Thu Jan  4 03:09:48.573 UTC
I/f: TenGigE0/0/0/4.1, Location: 0/RP0/CPU0
Dest: 10.1.1.1
Src:  10.1.1.2
 State: UP for 0d:0h:9m:27s, number of times UP: 1
 Session type: PR/V4/SH
Received parameters:
 Version: 1, desired tx interval: 10 ms, required rx interval: 10 ms
Required echo rx interval: 0 ms, multiplier: 2, diag: None
 My discr: 2147483898, your discr: 2147483899, state UP, D/F/P/C/A: 0/0/0/1/0
Transmitted parameters:
 Version: 1, desired tx interval: 10 ms, required rx interval: 10 ms
 Required echo rx interval: 0 ms, multiplier: 2, diag: None
 My discr: 2147483899, your discr: 2147483898, state UP, D/F/P/C/A: 0/1/0/1/0
Timer Values:
 Local negotiated async tx interval: 10 ms
 Remote negotiated async tx interval: 10 ms
 Desired echo tx interval: 0 s, local negotiated echo tx interval: 0 ms
 Echo detection time: 0 ms(0 ms*2), async detection time: 20 ms(10 ms*2)
Local Stats:
 Intervals between async packets:
   Tx: Number of intervals=100, min=6 ms, max=6573 ms, avg=1506 ms
       Last packet transmitted 186 s ago
   Rx: Number of intervals=100, min=4 ms, max=5 s, avg=575 ms
       Last packet received 184 s ago
 Intervals between echo packets:
   Tx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet transmitted 0 s ago
   Rx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet received 0 s ago
 Latency of echo packets (time between tx and rx):
   Number of packets: 0, min=0 ms, max=0 ms, avg=0 ms
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  bgp-default          10 ms      2          10 ms      2         

H/W Offload Info:
 H/W Offload capability : Y, Hosted NPU     : 0//CPU0
 Async Offloaded        : Y, Echo Offloaded : N
 Async rx/tx            : 344/209 

Platform Info:
NPU ID: 0 
Async RTC ID        : 1          Echo RTC ID        : 0
Async Feature Mask  : 0x0        Echo Feature Mask  : 0x0
Async Session ID    : 0xfb       Echo Session ID    : 0x0
Async Tx Key        : 0x800000fb  Echo Tx Key        : 0x0
Async Tx Stats addr : 0x0   Echo Tx Stats addr : 0x0
Async Rx Stats addr : 0x0   Echo Rx Stats addr : 0x0

/* Verify the complete history including session state, type, transitions, offload history, last down reason if any,
 received and transmitted packets, rx/tx intervals, location, timestamp, and local and remote descriptors. */

RP/0/RP0/CPU0:router# show bfd session status history destination 10.1.10.1 location 0/RP0/CPU0  

Thu Jan  4 03:45:18.768 UTC
I/f: TenGigE0/0/0/4.10, Location: 0//CPU0 table_id:0xe0000000
State: UP, flags:0x80040
Iftype: 0x19, basecaps: 107
Async dest addr: 10.1.10.1
Async src addr: 10.1.10.2
Echo dest addr: 10.1.10.2
Echo src addr: 192.0.2.1
Additional info from Flags: 
 FIB is READY
 Session Active on 0/RP0/CPU0
Platform Info: 0x0, Mac Length: 18
Redundancy session info:
 Created from active BFD server
Last Down Diag: None
Last UP Time: Jan  4 03:00:19.272 

Received parameters:
 Version: 1, desired tx interval: 10 ms, required rx interval: 10 ms
 Required echo rx interval: 0 ms, multiplier: 2, diag: None
 My discr: 2147483747, your discr: 2147483751, state UP, D/F/P/C/A: 0/0/0/1/0

Transmitted parameters:
 Version: 1, desired tx interval: 10 ms, required rx interval: 10 ms
 Required echo rx interval: 0 ms, multiplier: 2, diag: None
 My discr: 2147483751, your discr: 2147483747, state UP, D/F/P/C/A: 0/1/0/1/0

Tx Echo pkt :
 Version: 0, Local Discr: 2147483751, Sequence No: 0

History:
[Jan  4 03:00:19.272] Session (v1) state change, triggered by event 'Remote
    state init', from INIT to UP with current diag being None
[Jan  4 03:00:16.851] Session (v1) state change, triggered by event 'Remote
    state down', from DOWN to INIT with current diag being None
[Jan  4 03:00:16.509] Session (v1) state change, triggered by event 'Session
    create', from Unknown to DOWN with current diag being None
[Jan  4 03:00:16.509] Flag cleared: session creation is in-progress, currently
    set flags (0x80040)

Offload history:
[Jan  4 03:06:42.013] Packet punted to sw: Packet word0 : (0x20c80218),
desired_min_tx_interval 10000, required_min_rx_interval 10000, Last punted pkt
    required_min_rx_interval 10000
[Jan  4 03:06:42.003] Packet punted to sw: Packet word0 : (0x20d80218),
desired_min_tx_interval 10000, required_min_rx_interval 10000, Last punted pkt
    required_min_rx_interval 10000
[Jan  4 03:06:41.989] Packet punted to sw: Packet word0 : (0x20c80218),
desired_min_tx_interval 10000, required_min_rx_interval 10000, Last punted pkt
    required_min_rx_interval 10000
[Jan  4 03:06:41.980] Packet punted to sw: Packet word0 : (0x20d80218),
desired_min_tx_interval 10000, required_min_rx_interval 10000, Last punted pkt
    required_min_rx_interval 10000
	
Rx Counters and Timestamps :
Async valid packets received: count 5280
  [Jan  4 03:06:42.013]  [Jan  4 03:06:42.003]  [Jan  4 03:06:41.989]
Async valid packets while session is not in Up state: count 3
  [Jan  4 03:00:19.272]  [Jan  4 03:00:18.030]  [Jan  4 03:00:16.851]

Coexistence Between BFD over Bundle and BFD over Logical Bundle

The coexistence between BFD over Bundle (BoB) and BFD over Logical Bundle (BLB) feature allows you to monitor either physical bundle member for BOB, or logical interface for BLB, or both. This feature enables BFD to converge fast.

Configuration Example

Configure one or more linecards to allow hosting of MP BFD sessions. If no linecards are included, linecards groups are not formed, and consequently no BFD MP sessions are created. For default settings of group size and number, you must add at least two lines with the bfd multiple-paths include location node-id command and valid line cards to the configuration for the algorithm to start forming groups and BFD MP sessions to be established.

Router(config)# bfd multipath include location 0/0/CPU0
Router(config)# bfd multipath include location 0/1/CPU0
/* Configure inherited coexistence mode */
Router(config)# bfd
Router(config-bfd)# bundle coexistence bob-blb inherit

/* Configure logical coexistence mode */
Router(config)# bfd
Router(config-bfd)# bundle coexistence bob-blb logical

Running Configuration

Running confiiguration for inherited coexistence:

bfd
bundle coexistence bob-blb inherit 

Running confiiguration for logical mode:

bfd
bundle coexistence bob-blb logical

Verification

Verify BOB and BLB coexistence inherited mode.

Router# show bfd session                                                
Mon May 31 02:55:44.584 UTC
Interface           Dest Addr           Local det time(int*mult)      State     
                                    Echo             Async   H/W   NPU     
------------------- --------------- ---------------- ---------------- ----------
Te0/0/0/7           33.33.33.2      0s(0s*0)         450ms(150ms*3)   UP    
                                                             Yes   0/0/CPU0       
BE123               33.33.33.2      n/a              n/a              UP        
                                                             No    n/a            
BE123.1             34.34.34.2      n/a              n/a              UP        
                                                             No    n/a            
Router# show bfd session interface bundle-ether 123 detail   
Fri May 28 13:49:35.268 UTC
I/f: Bundle-Ether123, Location: 0/RP0/CPU0
Dest: 33.33.33.2
Src: 33.33.33.1
 State: UP for 0d:0h:29m:50s, number of times UP: 1
 Session type: PR/V4/SH/BI/IB
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  bundlemgr_distrib    150 ms     3          150 ms     3         
Session association information:
  Interface            Dest Addr / Type                   
  -------------------- -----------------------------------
  Te0/0/0/7            33.33.33.2                              
                       BFD_SESSION_SUBTYPE_RTR_BUNDLE_MEMBER   
  BE123.1              34.34.34.2                              
                       BFD_SESSION_SUBTYPE_STATE_INHERIT       
Router# show bfd session interface bundle-ether 123.1 detail 
Fri May 28 13:49:59.316 UTC
I/f: Bundle-Ether123.1, Location: 0/RP0/CPU0
Dest: 34.34.34.2
Src: 34.34.34.1
 State: UP for 0d:0h:12m:54s, number of times UP: 1
 Session type: PR/V4/SH/IH
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  ipv4_static          100 ms     3          100 ms     3         
Session association information:
  Interface            Dest Addr / Type                   
  -------------------- -----------------------------------
  BE123                33.33.33.2                              
                       BFD_SESSION_SUBTYPE_RTR_BUNDLE_INTERFACE
Router# show bfd session interface tenGigE 0/0/0/7 detail 
Mon May 31 03:00:04.635 UTC
I/f: TenGigE0/0/0/7, Location: 0/0/CPU0
Dest: 33.33.33.2
Src: 33.33.33.1
 State: UP for 2d:13h:40m:19s, number of times UP: 1
 Session type: PR/V4/SH/BM/IB
Received parameters:
 Version: 1, desired tx interval: 150 ms, required rx interval: 150 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 2147493276, your discr: 2147492184, state UP, D/F/P/C/A: 0/0/0/1/0
Transmitted parameters:
 Version: 1, desired tx interval: 150 ms, required rx interval: 150 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 2147492184, your discr: 2147493276, state UP, D/F/P/C/A: 0/0/0/1/0
Timer Values:
 Local negotiated async tx interval: 150 ms
 Remote negotiated async tx interval: 150 ms
 Desired echo tx interval: 0 s, local negotiated echo tx interval: 0 ms
 Echo detection time: 0 ms(0 ms*3), async detection time: 450 ms(150 ms*3)
Local Stats:
 Intervals between async packets:
   Tx: Number of intervals=4, min=5 ms, max=15 s, avg=6927 ms
       Last packet transmitted 222007 s ago
   Rx: Number of intervals=15, min=3 ms, max=1700 ms, avg=1133 ms
       Last packet received 222018 s ago
 Intervals between echo packets:
   Tx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet transmitted 0 s ago
   Rx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet received 0 s ago
 Latency of echo packets (time between tx and rx):
   Number of packets: 0, min=0 ms, max=0 ms, avg=0 ms
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  bundlemgr_distrib    150 ms     3          150 ms     3         
Session association information:
  Interface            Dest Addr / Type                   
  -------------------- -----------------------------------
  BE123                33.33.33.2                              
                       BFD_SESSION_SUBTYPE_RTR_BUNDLE_INTERFACE

H/W Offload Info:
 H/W Offload capability : Y, Hosted NPU     : 0/0/CPU0
 Async Offloaded        : Y, Echo Offloaded : N
 Async rx/tx            : 122/51 

Platform Info:
NPU ID: 0 
Async RTC ID        : 1          Echo RTC ID        : 0
Async Feature Mask  : 0x0        Echo Feature Mask  : 0x0
Async Session ID    : 0x2158     Echo Session ID    : 0x0
Async Tx Key        : 0x80002158  Echo Tx Key        : 0x0
Async Tx Stats addr : 0x0   Echo Tx Stats addr : 0x0
Async Rx Stats addr : 0x0   Echo Rx Stats addr : 0x0

Verify BOB and BLB coexistence logical mode.

show bfd session
Mon May 31 02:54:07.442 UTC
Interface           Dest Addr           Local det time(int*mult)      State     
                                    Echo             Async   H/W   NPU     
------------------- --------------- ---------------- ---------------- ----------
Te0/0/0/7           33.33.33.2      0s(0s*0)         450ms(150ms*3)   UP        
                                                             Yes   0/0/CPU0
       
BE123.1             34.34.34.2      0s(0s*0)         300ms(100ms*3)   UP        
                                                             Yes   0/0/CPU0
       
BE123               33.33.33.2      n/a              n/a              UP        
                                                             No    n/a            
Router# show bfd session interface bundle-ether 123 detail                        
Fri May 28 14:04:41.331 UTC
I/f: Bundle-Ether123, Location: 0/RP0/CPU0
Dest: 33.33.33.2
Src: 33.33.33.1
 State: UP for 0d:0h:44m:56s, number of times UP: 1
 Session type: PR/V4/SH/BI/IB
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  bundlemgr_distrib    150 ms     3          150 ms     3         
Session association information:
  Interface            Dest Addr / Type                   
  -------------------- -----------------------------------
  Te0/0/0/7            33.33.33.2                              
                       BFD_SESSION_SUBTYPE_RTR_BUNDLE_MEMBER   
Router# show bfd session interface tenGigE 0/0/0/7 detail 
Mon May 31 03:04:25.714 UTC
I/f: TenGigE0/0/0/7, Location: 0/0/CPU0
Dest: 33.33.33.2
Src: 33.33.33.1
 State: UP for 2d:13h:44m:40s, number of times UP: 1
 Session type: PR/V4/SH/BM/IB
Received parameters:
 Version: 1, desired tx interval: 150 ms, required rx interval: 150 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 2147493276, your discr: 2147492184, state UP, D/F/P/C/A: 0/0/0/1/0
Transmitted parameters:
 Version: 1, desired tx interval: 150 ms, required rx interval: 150 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 2147492184, your discr: 2147493276, state UP, D/F/P/C/A: 0/0/0/1/0
Timer Values:
 Local negotiated async tx interval: 150 ms
 Remote negotiated async tx interval: 150 ms
 Desired echo tx interval: 0 s, local negotiated echo tx interval: 0 ms
 Echo detection time: 0 ms(0 ms*3), async detection time: 450 ms(150 ms*3)
Local Stats:
 Intervals between async packets:
   Tx: Number of intervals=4, min=5 ms, max=15 s, avg=6927 ms
       Last packet transmitted 222268 s ago
   Rx: Number of intervals=15, min=3 ms, max=1700 ms, avg=1133 ms
       Last packet received 222279 s ago
 Intervals between echo packets:
   Tx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet transmitted 0 s ago
   Rx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet received 0 s ago
 Latency of echo packets (time between tx and rx):
   Number of packets: 0, min=0 ms, max=0 ms, avg=0 ms
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  bundlemgr_distrib    150 ms     3          150 ms     3         
Session association information:
  Interface            Dest Addr / Type                   
  -------------------- -----------------------------------
  BE123                33.33.33.2                              
                       BFD_SESSION_SUBTYPE_RTR_BUNDLE_INTERFACE

H/W Offload Info:
 H/W Offload capability : Y, Hosted NPU     : 0/0/CPU0

 Async Offloaded        : Y, Echo Offloaded : N
 Async rx/tx            : 122/51 

Platform Info:
NPU ID: 0 
Async RTC ID        : 1          Echo RTC ID        : 0
Async Feature Mask  : 0x0        Echo Feature Mask  : 0x0
Async Session ID    : 0x2158     Echo Session ID    : 0x0
Async Tx Key        : 0x80002158  Echo Tx Key        : 0x0
Async Tx Stats addr : 0x0   Echo Tx Stats addr : 0x0
Async Rx Stats addr : 0x0   Echo Rx Stats addr : 0x0
Router# show bfd session interface bundle-ether 123.1 detail                      
Fri May 28 14:04:46.893 UTC
I/f: Bundle-Ether123.1, Location: 0/0/CPU0

Dest: 34.34.34.2
Src: 34.34.34.1
 State: UP for 0d:0h:5m:18s, number of times UP: 1
 Session type: SW/V4/SH/BL
Received parameters:
 Version: 1, desired tx interval: 100 ms, required rx interval: 100 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 984, your discr: 124, state UP, D/F/P/C/A: 0/0/0/1/0
Transmitted parameters:
 Version: 1, desired tx interval: 100 ms, required rx interval: 100 ms
 Required echo rx interval: 0 ms, multiplier: 3, diag: None
 My discr: 124, your discr: 984, state UP, D/F/P/C/A: 0/1/0/1/0
Timer Values:
 Local negotiated async tx interval: 100 ms
 Remote negotiated async tx interval: 100 ms
 Desired echo tx interval: 0 s, local negotiated echo tx interval: 0 ms
 Echo detection time: 0 ms(0 ms*3), async detection time: 300 ms(100 ms*3)
Label:
 Internal label: 24000/0x5dc0
Local Stats:
 Intervals between async packets:
   Tx: Number of intervals=3, min=103 ms, max=19 s, avg=7023 ms
       Last packet transmitted 318 s ago
   Rx: Number of intervals=15, min=1 ms, max=1704 ms, avg=1315 ms
       Last packet received 318 s ago
 Intervals between echo packets:
   Tx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet transmitted 0 s ago
   Rx: Number of intervals=0, min=0 s, max=0 s, avg=0 s
       Last packet received 0 s ago
 Latency of echo packets (time between tx and rx):
   Number of packets: 0, min=0 ms, max=0 ms, avg=0 ms
MP download state: BFD_MP_DOWNLOAD_ACK
State change time: May 28 13:59:07.124
Session owner information:
                            Desired               Adjusted
  Client               Interval   Multiplier Interval   Multiplier
  -------------------- --------------------- ---------------------
  ipv4_static          100 ms     3          100 ms     3         

H/W Offload Info:
 H/W Offload capability : Y, Hosted NPU     : 0/0/CPU0

 Async Offloaded        : Y, Echo Offloaded : N
 Async rx/tx            : 16/4 

Platform Info:
NPU ID: 0 
Async RTC ID        : 1          Echo RTC ID        : 0
Async Feature Mask  : 0x0        Echo Feature Mask  : 0x0
Async Session ID    : 0x7c       Echo Session ID    : 0x0
Async Tx Key        : 0x7c  Echo Tx Key        : 0x0
Async Tx Stats addr : 0x0   Echo Tx Stats addr : 0x0
Async Rx Stats addr : 0x0   Echo Rx Stats addr : 0x0

BFD Hardware Offload Support for IPv4

The Bidirectional Forwarding detection (BFD) Hardware Offload feature enables the offload of a BFD session to the network processing units of the line cards, in an IPv4 network. BFD hardware offload improves scale and reduces the overall network convergence time by sending rapid failure detection packets to the routing protocols for recalculating the routing table.

Restrictions

  • This feature is not supported over MPLS LDP interface and VRRP interface.

  • This feature is not supported over MPLS TE or RSVP tunnel.

  • BFD multihop will flap if underlay paths that consist of multiple bundle VLANs flap.

Confiugration Example

/* Configure BFD over Bundle(BOB) for hardware offload. */
Router# config
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd mode ietf
Router(config-if)# bfd address-family ipv4 multiplier 3
Router (config-if)# bfd address-family ipv4 destination 10.20.20.1
Router (config-if)# bfd address-family ipv4 fast-detect 
Router(config-if)# bfd address-family ipv4 minimum-interval 2000
Router(config-if)# ipv4 address 10.20.20.2/30
/* To define the line card to host BLB and BFD multihop sessions. */
Router(config)# bfd
Router(config-bfd)# multipath include location 0/0/CPU0
/* Configure BFD with a static route. */
Router(config)# router static
Router(config-static)# address-family ipv4 unicast 10.1.1.0/24 10.6.0.2 bfd fast-detect minimum-interval 350 multiplier 4
/* Configure BFD with IS-IS. */
Router(config)# router isis 65444
Router(config-isis)# address-family ipv4 unicast
Router(config-isis)# exit
Router(config-isis)# interface gigabitEthernet 0/3/0/1
Router(config-isis-if)#  bfd minimum-interval 6500
Router(config-isis-if)# bfd multiplier 7
Router(config-isis-if)# bfd fast-detect ipv4
Router(config-isis-if)# address-family ipv4 unicast
/* Configure BFDv4 with OSPF. */
Router(config)# router ospf main
Router(config-ospfv3)# area 0
Router(config-ospfv3-ar)#  interface gigabitEthernet 1/0/0/1
Router(config-ospfv3-ar-if)# bfd multiplier 7
Router(config-ospfv3-ar-if)# bfd fast-detect
Router(config-ospfv3-ar-if)# bfd minimum-interval 6500
/* Configuring BFD over BGP. */
Router(config)# router bgp 120
Router(config-bgp)# neighbor 10.6.6.1
Router(config-bgp-nbr)# bfd fast-detect
Router(config-bgp-nbr)# bfd multiplier 7
Router(config-bgp-nbr)# bfd minimum-interval 6500

Verification

Use the show bfd ipv4 session command to verify the configuration:

Router# show bfd ipv4 session
Interface           Dest Addr           Local det time(int*mult)       State     
                                    Echo             Async   H/W       NPU     
------------------- --------------- ---------------- ---------------- ----------
Hu0/0/0/22.93       10.20.20.1       0s(0s*0)         12ms(4ms*3)      UP        
                                                             Yes   0/0/CPU0 

BFD Hardware Offload Support for IPv6

The Bidirectional Forwarding detection (BFD) Hardware Offload feature enables the offload of a BFD session to the network processing units of the line cards, in an IPv6 network. BFD hardware offload feature improves scale and reduces the overall network convergence time by sending rapid failure detection packets to the routing protocols for recalculating the routing table.

Restrictions

  • This feature is not supported over MPLS LDP interface and VRRP interface.

  • This feature is not supported over MPLS TE or RSVP tunnel.

  • BFD Dampening is not supported for BFD over IPv6.

  • BFD over Bundle (BOB) over IPv6 is not supported with dynamically configured link-local address. It must be statically configured.

  • BFD multihop will flap if underlay paths that consist of multiple bundle VLANs flap.

Configuration Example

/* Configure BFD over Bundle(BOB) for hardware offload. */
Router# config
Router(config)# interface Bundle-Ether 1
Router(config-if)# bfd mode ietf
Router(config-if)# bfd address-family ipv6 multiplier 3
Router (config-if)# bfd address-family ipv6 destination 10.20:20::1
Router (config-if)# bfd address-family ipv6 fast-detect 
Router(config-if)# bfd address-family ipv6 minimum-interval 2000
Router(config-if)# ipv6 address 10:20:20::2/64
/* To define the line card to host BLB and BFD multihop sessions. */
Router(config)# bfd
Router(config-bfd)# multipath include location 0/0/CPU0
/* Configure BFD with a static route. */
Router(config)# router static
Router(config-static)# address-family ipv6 unicast 1011:17e4::1/128 ab11:15d2::2 bfd fast-detect minimum-interval 50 multiplier 3
/* Configure BFD with IS-IS. */
Router(config)# router isis 65444
Router(config-isis)# address-family ipv6 unicast
Router(config-isis)# exit
Router(config-isis)# interface gigabitEthernet 0/3/0/1
Router(config-isis-if)#  bfd minimum-interval 6500
Router(config-isis-if)# bfd multiplier 7
Router(config-isis-if)# bfd fast-detect ipv6
Router(config-isis-if)# address-family ipv6 unicast
/* Configure BFDv6 with OSPFv3. */
Router(config)# router ospfv3 main
Router(config-ospfv3)# area 0
Router(config-ospfv3-ar)#  interface gigabitEthernet 1/0/0/1
Router(config-ospfv3-ar-if)# bfd multiplier 7
Router(config-ospfv3-ar-if)# bfd fast-detect
Router(config-ospfv3-ar-if)# bfd minimum-interval 6500
/* Configuring BFD over BGP. */
Router(config)# router bgp 120
Router(config-bgp)# neighbor 2001:DB8:1::1
Router(config-bgp-nbr)# bfd fast-detect
Router(config-bgp-nbr)# bfd multiplier 7
Router(config-bgp-nbr)# bfd minimum-interval 6500

Verification

Use the show bfd ipv6 session command to verify the configuration:

Router# show bfd ipv6 session
Interface           Dest Addr      
                                        Local det time(int*mult)      State     
H/W                 NPU             Echo             Async           
------------------- --------------- ---------------- ---------------- ----------
BE7.2               fe80::28a:96ff:fed6:9cdb                      
Yes                 0/0/CPU0        0s(0s*0)         900ms(300ms*3)   UP        
BE7.4               fe80::28a:96ff:fed6:9cdb                      
Yes                 0/0/CPU0        0s(0s*0)         900ms(300ms*3)   UP        

IPv4 Multihop BFD

Table 7. Feature History Table

Feature Name

Release Information

Feature Description

Multihop BFD for IPv4 nondefault VRF Release 7.7.1 BFD provides fast forwarding path failure detection between two routing devices that are connected by a network link. BFD Multihop enables you to detect connectivity between routers that span multiple network hops and follow unpredictable paths. Prior to this release, BFD Multihop was supported on default VRFs only. This feature provides you the flexibility to extend BFD Multihop for IPv4 non-default VRFs.
Table 8. Feature History Table

Feature Name

Release Information

Feature Description

IPv4 Multihop BFD

Release 7.4.1

This feature is now supported on routers that have Cisco NC57 line cards that are installed and operate in native and compatibility modes.

IPv4 Multihop BFD is a BFD session between two nodes, such as a PE and CE node, or between routers that are several TTL hops away. This feature provides subsecond forwarding failure detection for a destination more than one hop, and up to 255 hops away.

Thus, the advantage of BFD, low-overhead, and short-duration detection of path failures between routers, is extended to a multihop scenario.

Table 9. Feature History Table

Feature Name

Release Information

Feature Description

IPv4 BFD Multihop over MPLS Core and Segment Routing

Release 7.5.1

This feature is now supported on routers that have the Cisco NC57 line cards that are installed and operate in native and compatible modes.

This feature enables you to configure IPv4 Multihop BFD on MPLS LDP and Segment Routing.

IPv4 Multihop BFD is a BFD session between two addresses between two nodes. An example of this feature is a BFD session between PE and CE loopback addresses or BFD sessions between routers that are several TTL hops away. The applications that support IPv4 Multihop BFD are external and internal BGP. IPv4 Multihop BFD feature supports BFD on arbitrary paths, which can span multiple networks hops.

A Virtual Routing and Forwarding (VRF) instance is a logical separation of a router's routing table. VRF allows you to have multiple routing tables on a single router, each with its own set of routes.

The default VRF is the first VRF that is created on a router. It is the VRF that is used by default for all routing protocols and interfaces.

Non-default VRFs must be explicitly configured.

The IPv4 Multihop BFD feature provides subsecond forwarding failure detection for a destination more than one hop, and up to 255 hops, away. IPv4 Multihop BFD feature is supported on all currently supported media-type for BFD single hop.

You can set up a BFD multihop session between a unique source-destination address pair that is provided by the client. You can set up a session two endpoints that have IP connectivity.

Multihop BFD over nondefault VRF feature runs on both default and non-default VRF.

Multihop BFD over nondefault VRF feature runs on IPv4 only.

The IPv4 BFD Multihop feature enables you to configure IPv4 Multihop BFD on MPLS LDP and Segment Routing.

For information on configuring MPLS LDP, see the MPLS Configuration Guide for Cisco NCS 5500 Series Routers.

For information on configuring Segment Routing, see the Segment Routing Configuration Guide for Cisco NCS 5500 Series Routers.

Configure IPv4 Multihop BFD

This section describes how you can configure IPv4 Multihop BFD feature.


Router# configure
Router(config)# bfd
Router(config)# multipath include location 0/7/CPU0
Router(config)# router bgp 100
Router(config-bgp)# neighbor 209.165.200.225
Router(config-bgp-nbr)# remote-as 2000
Router(config-bgp-nbr)# update-source loopback 1
Router(config-bgp-nbr)# bfd fast-detect
Router(config-bgp-nbr)# bfd multiplier 3
Router(config-bgp-nbr)# bfd minimum-interval 300
Router(config-bgp-nbr)#address-family ipv4 unicast
Router(config-bgp-nbr-af)# route-policy pass-all in
Router(config-bgp-nbr-af)# route-policy pass-all out
Router(config-bgp-nbr-af)# commit

Running Configuration


bfd
 multipath include location 0/7/CPU0
router bgp 100
 neighbor 209.165.200.225
  remote-as 2000
  update-source loopback 1
  bfd fast-detect
  bfd multiplier 3
  bfd minimum-interval 300
address-family ipv4 unicast
      route-policy PASS-ALL in
    route-policy PASS-ALL out
  !
!

Verification

The show outputs given in the following section display the details of the configuration of the IPv4 Multihop BFD feature, and the status of their configuration.


Router# show tech-support bfdhwoff location 0/7/CPU0 file
harddisk: 
Tue Mar 20 11:20:29.214 PDT
++ Show tech start time: 2018-Mar-20.112029.PDT ++
Tue Mar 20 11:20:30 PDT 2018 Waiting for gathering to complete .............................
Tue Mar 20 11:22:37 PDT 2018 Compressing show tech output Show tech output available at 0/RP0/CPU0 :
/harddisk:/showtech-bfd-hwoff-platform-2018-Mar-20.112029.PDT.tgz
++ Show tech end time: 2018-Mar-20.112237.PDT ++

Configure Multihop BFD on IPv4 nondefault VRFs

Configure the following steps to configure Multihop BFD on IPv4 nondefault VRF:

Router(config)# router bgp 100
Router(config-bgp)# neighbor 209.165.200.225
Router(config-bgp-nbr)#vrf vrf1
Router(config-bgp-nbr-vrf)# exit
Router(config-bgp-nbr)# rd auto
Router(config-bgp-nbr)#address-family ipv4 unicast
Router(config-bgp-nbr-af)#redistribute connected
Router(config-bgp-nbr-af)# exit
Router(config-bgp)# neighbor 209.165.200.225 
Router(config-bgp-nbr)# remote-as 2000
Router(config-bgp-nbr)# bfd fast-detect
Router(config-bgp-nbr)# bfd multiplier 3
Router(config-bgp-nbr)# bfd minimum-interval 50
Router(config-bgp-nbr)# ebgp-multihop 255
Router(config-bgp-nbr)# update-source loopback 1
/* You can configure any interface here, including loopback or bvi */
Router(config-bgp-nbr)#address-family ipv4 unicast
Router(config-bgp-nbr-af)# route-policy pass-all in
Router(config-bgp-nbr-af)# route-policy pass-all out
Router(config-bgp-nbr-af)# commit

Running Configuration


router bgp 100
  neighbor 209.165.200.225
   vrf vrf1  
    exit
     rd auto
      address-family ipv4 unicast
       redistribute connected
        exit
         neighbor 209.165.200.225   
         remote-as 2000
         bfd fast-detect
         bfd multiplier 3
         bfd minimum-interval 50
         ebgp-multihop 255
         update-source loopback 1
          address-family ipv4 unicast
           route-policy pass-all in
           route-policy pass-all out

Verification


Router# show bfd session source 209.165.200.225 
Thu Mar 10 10:13:43.305 IST
Src Addr         Dest Addr    VRF Name         H/W NPU          
                               Local det time(int*mult)   State     
                              Echo    Async           
--------------   ---------   ------   ------ --------- ---------
209.165.200.225  192.0.2.254   vrf_1    Yes   0/0/CPU0     
                              n/a  150ms(50ms*3)          UP        
Router# show cef vrf vrf_1 209.165.200.225 location 0/0/CPU0
Thu Mar 10 10:24:13.372 IST
209.165.200.0/24, version 40, internal 0x5000001 0x30 (ptr 0x8ae26458) [1], 0x0 (0x0), 0xa08 (0x8dc144a8)
 Updated Mar  9 15:09:43.398
 Prefix Len 24, traffic index 0, precedence n/a, priority 3
LDI Update time Mar  9 14:59:28.284
   via 1.1.1.1/32, 605 dependencies, recursive [flags 0x6000]
    path-idx 0 NHID 0x0 [0x8dd35988 0x0]
    recursion-via-/32
    next hop VRF - 'default', table - 0xe0000000
    next hop 10.1.1.1/32 via 24015/0/21
    next hop 192.0.2.255/32 Te0/0/0/3.1  labels imposed {ImplNull 24162}

IPv6 Multihop BFD

Table 10. Feature History Table

Feature Name

Release Information

Feature Description

IPv6 Multihop BFD support

Release 7.4.1

BFD IPv6 Multihop feature enables IPv6 Multihop BFD sessions where BFD neighbors can be multiple hops away, either physically or logically.

It removes the restriction of a single path IPv6 BFD session, where the BFD neighbor is always one hop away, and the BFD Agent in the line card always receives or transmits BFD packets over a local interface on the same line card.

Thus, the advantage of BFD, low-overhead and short-duration detection of path failures between (IPv6) routers, is extended to a multihop scenario.

This feature is also supported on routers that have Cisco NC57 line cards installed and operate in native and compatibility modes.

In a Multihop Bidirectional Forwarding Detection (BFD) scenario, more than one path is available for a node to reach its IPv6 BFD neighbor. BFD packets are received on a line card that may or may not host the respective BFD session. The BFD Agent in one line card may need to transmit BFD packets out of an egress interface on a different line card. The BFD switching mechanism for IPv6 Multihop link is used when the BFD packets are transmitted from one end point node to the other. The BFD punting mechanism is employed when BFD packets are received at the remote end point node.

Configuration Example

/* BFD IPv6 Multihop for an eBGP Neighbor */

Router# configure

Specify a line card to host a BFD multihop session.

Router(config)# bfd multipath include location 0/7/CPU0

Enable multihop peering with the eBGP neighbor, and enable BFD fast detection.

Router(config)# router bgp 65001
Router(config-bgp)#neighbor 21:1:1:1:1:1:1:2 ebgp-multihop 255
Router(config-bgp)#neighbor 21:1:1:1:1:1:1:2 bfd fast-detect
Router(config-bgp)#commit

/* BFD IPv6 Multihop for an iBGP Neighbor */

Router# configure

Specify a line card to host a BFD multihop session.

Router(config)# bfd multipath include location 0/0/CPU0
Router(config)# router bgp 65001

Specifies an iBGP neighbor and enable BFD fast detection.

Router(config-bgp)#neighbor 21:1:1:1:1:1:1:2
Router(config-bgp-nbr)#commit

Running Configuration

/* BFD IPv6 Multihop for an eBGP Neighbor */

bfd
 multipath include location 0/7/CPU0
!
router bgp 65001
 neighbor 21:1:1:1:1:1:1:2
  bfd fast-detect
  ebgp-multiphop 255

/* BFD IPv6 Multihop for an iBGP Neighbor */

bfd
 multipath include location 0/0/CPU0
!
router bgp 65001
 neighbor 21:1:1:1:1:1:1:2
  bfd fast-detect

BFD-Triggered FRR

The BFD-triggered Fast Reroute (FRR) feature allows you to obtain link and node protection using Bidirectional Forwarding Detection (BFD) protocol. This feature provides fast forwarding path failure detection for the following:

  • All media types

  • Encapsulations

  • Topologies

  • Routing protocols

In addition to fast forwarding path failure detection, BFD provides a consistent failure detection method for network administrators.

When you enable FRR on Interior Gateway Protocol (IGP) in an IP network, BFD triggers FRR. BFD switches to the backup path when either the primary or the secondary link fails.

When you enable FRR on IGP in an MPLS TE network, you can enable BFD on single-hop IGP links. MPLS TE uses these links to define the protected TE tunnel paths. During traffic disruption on either the primary or the secondary links, BFD triggers a link-down event that triggers FRR. 

When you enable FRR on IGP in a segment routing network, BFD triggers FRR.

This feature complements the link-scan failure detection feature. Among these two features, whichever detects the link down event first, triggers the FRR. 

Restrictions

BFD-triggered FRR feature has the following restrictions:

  • You cannot enable BFD on TE tunnels.

  • You cannot enable BFD for multihop.

  • You cannot enable BFD on logical bundles.

Configuration Example


Router# config
Router(config)# router ospf 100
Router(config-ospf)# router-id 10.32.32.32
Router(config-ospf)# area 0
Router(config-ospf)# exit
Router(config)# interface TenGigE 0/4/0/0.1
Router(config-subif)# bfd minimum-interval 3 
Router(config-subif)# bfd fast-detect
Router(config-if)# exit
Router(config)# interface TenGigE 0/4/0/3
Router(config-if)# bfd minimum-interval 3 
Router(config-if)# bfd fast-detect
Router(config-if)# exit
Router(config)# mpls traffic-eng
Router(config-mpls)# interface TenGigE 0/4/0/0.1
Router(config-mpls-if)# backup-path tunnel-te 5000
Router(config-mpls-if)# exit
Router(config-mpls)# interface TenGigE0/4/0/3
Router(config-mpls-if)# interface TenGigE0/4/0/3
Router(config-if)# exit
Router(config)# mpls ldp
Router(config-ldp)# router-id 10.32.32.32
Router(config-ldp)# interface Bundle-Ether1
Router(config-ldp-if)# exit
Router(config-ldp)# interface TenGigE0/4/0/0.1
Router(config-ldp-if)# exit
Router(config-ldp-if)# interface TenGigE0/4/0/3
Router(config-ldp-if)# exit
Router(config-ldp)# exit
Router(config)# rsvp
Router(config-rsvp)# interface TenGigE0/4/0/0.1
Router(config-rsvp-if)# bandwidth 100000000
Router(config-rsvp-if)# exit
Router(config-rsvp)# interface TenGigE0/4/0/3
Router(config-rsvp-if)# bandwidth 100000000
Router(config-rsvp-if)# exit
Router(config-rsvp)# exit
Router(config)# exit
Router(config)# interface tunnel-te5000
Router(config-if)# ipv4 unnumbered Loopback0
Router(config-if)# destination 10.31.31.31
Router(config-if)# path-option 1 explicit name ind1
Router(config-if)# exit
Router(config)# exit
Router(config)# explicit-path name direct1-sub
Router(config-expl-path)# index 1 next-address strict ipv4 unicast 10.2.36.2
Router(config-expl-path)# destination 10.31.31.31
Router(config-expl-path)# exit
Router(config)# explicit-path name ind1
Router(config-expl-path)# index 1 next-address strict ipv4 unicast 10.1.33.2
Router(config-expl-path)# index 1 next-address strict ipv4 unicast 10.1.30.1
Router(config-expl-path)# exit

Running Configuration

router ospf 100
 router-id 32.32.32.32
 area 0
interface TenGigE0/4/0/0.1
   bfd minimum-interval 3
   bfd fast-detect
interface TenGigE0/4/0/3
   bfd minimum-interval 3
   bfd fast-detect
!
mpls traffic-eng
interface TenGigE0/4/0/0.1
  backup-path tunnel-te 5000
interface TenGigE0/4/0/3
 !
mpls ldp
 router-id 32.32.32.32
 interface Bundle-Ether1
 !
 interface TenGigE0/4/0/0.1
 interface TenGigE0/4/0/3
 !
rsvp
interface TenGigE0/4/0/0.1
  bandwidth 100000000
interface TenGigE0/4/0/3
  bandwidth 100000000
! 
interface tunnel-te1
 ipv4 unnumbered Loopback0
 destination 10.31.31.31
 fast-reroute
 path-option 1 explicit name direct1-sub
 !
interface tunnel-te5000
 ipv4 unnumbered Loopback0
 destination 10.31.31.31
 path-option 1 explicit name ind1
!
 !
explicit-path name direct1-sub
 index 1 next-address strict ipv4 unicast 10.2.36.2
 !
explicit-path name ind1
 index 1 next-address strict ipv4 unicast 10.1.33.2
 index 2 next-address strict ipv4 unicast 10.1.30.1

BFD over BVI

Table 11. Feature History Table

Feature Name

Release Information

Feature Description

BFD Over BVI

Release 7.4.1

This feature is now supported on routers that have Cisco NC57 line cards installed and operate in native and compatibility modes.

This feature is also supported on routers that have Cisco NCS550x and Cisco NCS55Ax line cards installed and operate in native and compatibility modes.

BFD over IRB, using a BVI, is a multipath single-hop session. In a BFD multipath session, BFD can be applied over virtual interfaces or between interfaces that are multihops away. The advantage of BFD, low-overhead and short-duration detection of path failures between routers, is extended to an IRB deployment scenario.

In order for a VLAN to span a router, the router must be capable of forwarding frames from one interface to another, while maintaining the VLAN header. If the router is configured for routing a Layer 3 (network layer) protocol, it will terminate the VLAN and MAC layers at the interface on which a frame arrives. The MAC layer header can be maintained if the router bridges the network layer protocol. However, even regular bridging terminates the VLAN header.

Using the Integrated Routing Bridging (IRB) feature, a router can be configured for routing and bridging the same network layer protocol, on the same interface. This allows the VLAN header to be maintained on a frame while it transits a router from one interface to another. IRB provides the ability to route between a bridged domain and a routed domain with the Bridge Group Virtual Interface (BVI). The BVI is a virtual interface within the router that acts like a normal routed interface that does not support bridging, but represents the comparable bridge group to routed interfaces within the router. The interface number of the BVI is the number of the bridge group that the virtual interface represents. This number is the link between the BVI and the bridge group.

Because the BVI represents a bridge group as a routed interface, it must be configured only with Layer 3 (L3) characteristics, such as network layer addresses. Similarly, the interfaces configured for bridging a protocol must not be configured with any L3 characteristics.

BFD over IRB is a multipath single-hop session. In a BFD multipath session, BFD can be applied over virtual interfaces or between interfaces that are multihops away. The Cisco IOS XR Software BFD multihop is based on the RFC 5883—Bidirectional Forwarding Detection (BFD) for Multihop Paths. BFD over IRB is supported on IPv4 address, IPv6 global address, and IPv6 link-local address. The BFD over IRB is supported only in asynchronous mode and does not support echo mode.

For more information on IRB feature, see Integrated Routing and Bridging Chapter in Interface and Hardware Component Configuration Guide for Cisco NCS 5500 Series Routers.

Configuration Example


/* Configure a BVI interface and assign an IP address */
Router(config)# interface bvi 1
Router(config-if)# ipv4 address 192.168.1.1 255.255.255.0
Router(config-if)# exit

/* Configure the Layer 2 AC interface */
Router(config)# interface HundredGigE 0/0/1/3
Router(config-if)# l2transport
Router(config-if)# exit 

/* Configure OSPF as the routing protocol */
Router(config)# router ospf 100
Router(config-ospf)# router-id 192.168.1.1  
Router(config-ospf)# area 0
Router(config)# interface Loopback 100
Router(config)# exit

/* Configure BFD on BVI */
Router(config)# interface bvi1
Router(config-if)# bfd minimum-interval 100  
Router(config-if)# bfd fast-detect
Router(config-if)# bfd multiplier 3

/* Configure the line cards to allow hosting of Multipath BFD sessions. */
Router# configure
Router(config)# bfd multipath include location 0/5/CPU0

/* Configure BVI on the peer nodes */
Router(config-if)# l2vpn  
Router(config-l2vpn)# bridge group 1
Router(config-l2vpn-bg)# bridge domain 1
Router(config-l2vpn-bg-bd)# interface HundredGigE 0/0/1/3
Router(config-l2vpn-bg-bd-ac)# routed interface bvi 1

Running Configuration

interface BVI1
 ipv4 address 192.168.1.1 255.255.255.0
!
interface HundredGigE0/0/1/3
 l2transport
 !
!
router ospf 100
 router-id 192.168.1.1
 area 0
  interface Loopback100
  !
  interface BVI1
   bfd minimum-interval 100
   bfd fast-detect
   bfd multiplier 3
  !
 ! 
bfd multipath include location 0/5/CPU0        
!
l2vpn
 bridge group 1
  bridge-domain 1
   interface HundredGigE0/0/1/3
   !
   routed interface BVI1
   !

Multihop BFD over BVI

Table 12. Feature History Table

Feature Name

Release Information

Feature Description

Multihop BFD over Bridge Group Virtual Interface (BVI)

Release 7.4.1

The multihop BFD over Bridge Group Virtual Interface (BVI) feature introduces support for multihop BFD over (BVI). You can set up a multihop BFD session between two endpoints that have IP connectivity. This session is between a unique source-destination address pair that the client provides.

This feature allows you to extend BFD on arbitrary paths. These arbitrary paths can span multiple network hops, hence detecting link failures.

Multihop BFD over BVI feature allows you to configure both routing and bridging on the same interface using Integrated Routing Bridging (IRB). IRB enables you to route between a bridged domain and a routed domain with the Bridge Group Virtual Interface (BVI).

The BVI is a virtual interface within the router that acts like a normal, routed interface that does not support bridging, but represents the comparable bridge group to routed interfaces within the router.

Restrictions

  • The minimum Multihop BFD timer for the BVI interface is 50 msec.

  • The multihop ttl-drop-threshold command is not supported.

  • The Multihop BFD over BVI or IRB functionality is supported only in asynchronous mode and does not support echo mode.

  • The Multihop BFD over BVI feature is not supported over MPLS and SR core.

Supported Functionality

  • This feature is supported in both IPv4 and IPv6.

  • BFD Multihop over BVI feature supports on client BGP.

  • BFD Multihop supports only over IP core.

  • BFD Multihop supports on all currently supported media-type for BFD single-hop.

Configuration


/* Configure a BVI interface and assign an IP address */
Router(config)# interface BVI1
Router(config-if)# host-routing
Router(config-if)# mtu 8986
Router(config-if)# ipv4 address 10.1.1.1 255.255.255.0
Router(config-if)# ipv6 address 10:1:1::1/120

/* Configure the Layer 2 AC interface */ 
Router(config-if)# interface TenGigE0/5/0/6/0.1 l2transport
Router(config-subif)# encapsulation dot1q 1
Router(config-subif)# rewrite ingress tag pop 1 symmetric

/* Configure L2VPN Bridge Domain */
Router(config-subif)# l2vpn 
Router(config-subif)# bridge group 1 
Router(config-subif)# bridge-domain 1 
Router(config-l2vpn-bg-bd)# interface TenGigE0/5/0/6/0.1
Router(config-l2vpn-bg-bd)# routed interface BVI1

Running Configuration

interface BVI1
 host-routing
 mtu 8986
 ipv4 address 10.1.1.1 255.255.255.0
 ipv6 address 10:1:1::1/120
!
interface TenGigE0/5/0/6/0.1 l2transport
 encapsulation dot1q 1
 rewrite ingress tag pop 1 symmetric
!
l2vpn
 bridge group 1
  bridge-domain 1
   interface TenGigE0/5/0/6/0.1
   !
   routed interface BVI1
   !

Repeat the configuration on the peer router.


/* Configure BGP as the routing protocol  */
Router(config)# router bgp 1
Router(config-bgp)# neighbor 2.2.1.1
Router(config-bgp-nbr)# remote-as 1
Router(config-bgp-nbr)# bfd fast-detect
Router(config-bgp-nbr)# bfd minimum-interval 300
Router(config-bgp-nbr)# update-source Loopback1
Router(config-bgp-nbr)# address-family ipv4 unicast

/* Configure reachability to the BGP neighbour IP either via static or IGP*/
Router(config-bgp-nbr-af)# router static
Router(config-static)# address-family ipv4 unicast
Router(config-static-afi)# 2.2.1.1/32 10.1.1.2 

/* Configure the line cards to allow hosting of Multipath BFD sessions. */
Router(config-static-afi)# bfd
Router(config-bfd)# multipath include location 0/5/CPU0
 

router bgp 1 
neighbor 2.2.1.1
  remote-as 1
  bfd fast-detect
  bfd minimum-interval 300
  update-source Loopback1
  address-family ipv4 unicast
  !
router static
 address-family ipv4 unicast
  2.2.1.1/32 10.1.1.2
 !
bfd
 multipath include location 0/5/CPU0
!
 

Note


To avoid the unsupported three-level recursion on BVI interfaces on the first and second generation of line cards, you must not configure the BVI interface as the next-hop in the static route configuration.


Verification

Router# show bfd session destination 2.2.1.1
Fri May 28 14:35:52.566 IST

Src Addr            Dest Addr       VRF Name                      H/W NPU          
                                        Local det time(int*mult)      State     
                                    Echo             Async           
------------------- --------------- ---------------- ---------------- ----------
1.1.1.1             2.2.1.1         default                       Yes 0/5/CPU0     
                                    n/a              900ms(300ms*3)   UP