LDP

The ldpd daemon is a standardised protocol that permits exchanging MPLS label information between MPLS devices. The LDP protocol creates peering between devices, so as to exchange that label information. This information is stored in MPLS table of zebra, and it injects that MPLS information in the underlying system (Linux kernel or OpenBSD system for instance). ldpd provides necessary options to create a Layer 2 VPN across MPLS network. For instance, it is possible to interconnect several sites that share the same broadcast domain.

FRR implements LDP as described in RFC 5036; other LDP standard are the following ones: RFC 6720, RFC 6667, RFC 5919, RFC 5561, RFC 7552, RFC 4447. Because MPLS is already available, FRR also supports RFC 3031.

Running Ldpd

The ldpd daemon can be invoked with any of the common options (Common Invocation Options).

--ctl_socket: This option allows you to override the path to the ldpd.sock file used to control this daemon. If specified this option overrides the -N option path addition.

The zebra daemon must be running before ldpd is invoked.

Configuration for the daemon should be saved in the FRR integrated configuration file located in /etc/frr/frr.conf, see Integrated Config File for more information on system configuration.

Prior versions of FRR supported reading and writing per-daemon config files; however, with the introduction of the centralized management daemon mgmtd this could no longer be supported.

In order to allow for an orderly transition from per-daemon config files to the integrated config file, FRR daemons will continue to try and read their specific per-daemon configuration file as before. Additionally the config can still be loaded directly using the -f or --config-file CLI options; however, these files will not be updated when the configuration is written (e.g., with the write mem command).

Warning

Per-daemon files will no longer be updated when the user issues a write memory command. Therefore these per-daemon config files should only be used as a mechanism for transitioning to the integrated config, and then removed.

Understanding LDP principles

Let’s first introduce some definitions that permit understand better the LDP protocol:

LSR : Labeled Switch Router. Networking devices handling labels used to forward traffic between and through them.
LERLabeled Edge Router. A Labeled edge router is located at the edge of
an MPLS network, generally between an IP network and an MPLS network.

LDP aims at sharing label information across devices. It tries to establish peering with remote LDP capable devices, first by discovering using UDP port 646 , then by peering using TCP port 646. Once the TCP session is established, the label information is shared, through label advertisements.

There are different methods to send label advertisement modes. The implementation actually supports the following : Liberal Label Retention + Downstream Unsolicited + Independent Control. The other advertising modes are depicted below, and compared with the current implementation.

Liberal label retention versus conservative mode In liberal mode, every label sent by every LSR is stored in the MPLS table. In conservative mode, only the label that was sent by the best next hop (determined by the IGP metric) for that particular FEC is stored in the MPLS table.
Independent LSP Control versus ordered LSP Control MPLS has two ways of binding labels to FEC’s; either through ordered LSP control, or independent LSP control. Ordered LSP control only binds a label to a FEC if it is the egress LSR, or the router received a label binding for a FEC from the next hop router. In this mode, an MPLS router will create a label binding for each FEC and distribute it to its neighbors so long as he has a entry in the RIB for the destination. In the other mode, label bindings are made without any dependencies on another router advertising a label for a particular FEC. Each router makes it own independent decision to create a label for each FEC. By default IOS uses Independent LSP Control, while Juniper implements the Ordered Control. Both modes are interoperable, the difference is that Ordered Control prevent blackholing during the LDP convergence process, at cost of slowing down the convergence itself
unsolicited downstream versus downstream on demand Downstream on demand label distribution is where an LSR must explicitly request that a label be sent from its downstream router for a particular FEC. Unsolicited label distribution is where a label is sent from the downstream router without the original router requesting it.

LDP Configuration

mpls ldp: Enable or disable LDP daemon

router-id A.B.C.D: The following command located under MPLS router node configures the MPLS router-id of the local device.

ordered-control: Configure LDP Ordered Label Distribution Control.

address-family [ipv4 | ipv6]: Configure LDP for IPv4 or IPv6 address-family. Located under MPLS route node, this subnode permits configuring the LDP neighbors.

interface IFACE: Located under MPLS address-family node, use this command to enable or disable LDP discovery per interface. IFACE stands for the interface name where LDP is enabled. By default it is disabled. Once this command executed, the address-family interface node is configured.

discovery transport-address A.B.C.D | A:B::C:D: Located under mpls address-family interface node, use this command to set the IPv4 or IPv6 transport-address used by the LDP protocol to talk on this interface.

ttl-security disable

Located under the LDP address-family node, use this command to disable the GTSM procedures described in RFC 6720 (for the IPv4 address-family) and RFC 7552 (for the IPv6 address-family).

Since GTSM is mandatory for LDPv6, the only effect of disabling GTSM for the IPv6 address-family is that ldpd will not discard packets with a hop limit below 255. This may be necessary to interoperate with older implementations. Outgoing packets will still be sent using a hop limit of 255 for maximum compatibility.

If GTSM is enabled, multi-hop neighbors should have either GTSM disabled individually or configured with an appropriate ttl-security hops distance.

neighbor A.B.C.D password PASSWORD: The following command located under MPLS router node configures the router of a LDP device. This device, if found, will have to comply with the configured password. PASSWORD is a clear text password wit its digest sent through the network.

neighbor A.B.C.D holdtime HOLDTIME: The following command located under MPLS router node configures the holdtime value in seconds of the LDP neighbor ID. Configuring it triggers a keepalive mechanism. That value can be configured between 15 and 65535 seconds. After this time of non response, the LDP established session will be considered as set to down. By default, no holdtime is configured for the LDP devices.

neighbor A.B.C.D ttl-security disable: Located under the MPLS LDP node, use this command to override the global configuration and enable/disable GTSM for the specified neighbor.

neighbor A.B.C.D ttl-security hops (1-254): Located under the MPLS LDP node, use this command to set the maximum number of hops the specified neighbor may be away. When GTSM is enabled for this neighbor, incoming packets are required to have a TTL/hop limit of 256 minus this value, ensuring they have not passed through more than the expected number of hops. The default value is 1.

discovery hello holdtime HOLDTIME

discovery hello interval INTERVAL: INTERVAL value ranges from 1 to 65535 seconds. Default value is 5 seconds. This is the value between each hello timer message sent. HOLDTIME value ranges from 1 to 65535 seconds. Default value is 15 seconds. That value is added as a TLV in the LDP messages.

dual-stack transport-connection prefer ipv4: When ldpd is configured for dual-stack operation, the transport connection preference is IPv6 by default (as specified by RFC 7552). On such circumstances, ldpd will refuse to establish TCP connections over IPv4. You can use above command to change the transport connection preference to IPv4. In this case, it will be possible to distribute label mappings for IPv6 FECs over TCPv4 connections.

Show LDP Information

These commands dump various parts of ldpd.

show mpls ldp neighbor [A.B.C.D]

This command dumps the various neighbors discovered. Below example shows that local machine has an operation neighbor with ID set to 1.1.1.1.

west-vm# show mpls ldp neighbor
AF   ID              State       Remote Address    Uptime
ipv4 1.1.1.1         OPERATIONAL 1.1.1.1         00:01:37
west-vm#

show mpls ldp neighbor [A.B.C.D] capabilities

show mpls ldp neighbor [A.B.C.D] detail: Above commands dump other neighbor information.

show mpls ldp discovery [detail]

show mpls ldp ipv4 discovery [detail]

show mpls ldp ipv6 discovery [detail]: Above commands dump discovery information.

show mpls ldp ipv4 interface

show mpls ldp ipv6 interface

Above command dumps the IPv4 or IPv6 interface per where LDP is enabled. Below output illustrates what is dumped for IPv4.

west-vm# show mpls ldp ipv4 interface
AF   Interface   State  Uptime   Hello Timers  ac
ipv4 eth1       ACTIVE 00:08:35 5/15           0
ipv4 eth3       ACTIVE 00:08:35 5/15           1

show mpls ldp ipv4|ipv6 binding

Above command dumps the binding obtained through MPLS exchanges with LDP.

west-vm# show mpls ldp ipv4 binding
AF   Destination          Nexthop         Local Label Remote Label  In Use
ipv4 1.1.1.1/32           1.1.1.1         16          imp-null         yes
ipv4 2.2.2.2/32           1.1.1.1         imp-null    16                no
ipv4 10.0.2.0/24          1.1.1.1         imp-null    imp-null          no
ipv4 10.115.0.0/24        1.1.1.1         imp-null    17                no
ipv4 10.135.0.0/24        1.1.1.1         imp-null    imp-null          no
ipv4 10.200.0.0/24        1.1.1.1         17          imp-null         yes
west-vm#

LDP debugging commands

debug mpls ldp KIND

Enable or disable debugging messages of a given kind. KIND can be one of:

discovery
errors
event
labels
messages
zebra

Sample configuration

Below configuration gives a typical MPLS configuration of a device located in a MPLS backbone. LDP is enabled on two interfaces and will attempt to peer with two neighbors with router-id set to either 1.1.1.1 or 3.3.3.3.

mpls ldp
 router-id 2.2.2.2
 neighbor 1.1.1.1 password test
 neighbor 3.3.3.3 password test
 !
 address-family ipv4
  discovery transport-address 2.2.2.2
  !
  interface eth1
  !
  interface eth3
  !
 exit-address-family
 !

Deploying LDP across a backbone generally is done in a full mesh configuration topology. LDP is typically deployed with an IGP like OSPF, that helps discover the remote IPs. Below example is an OSPF configuration extract that goes with LDP configuration

router ospf
 ospf router-id 2.2.2.2
  network 0.0.0.0/0 area 0
 !

Below output shows the routing entry on the LER side. The OSPF routing entry (10.200.0.0) is associated with Label entry (17), and shows that MPLS push action that traffic to that destination will be applied.

north-vm# show ip route
Codes: K - kernel route, C - connected, S - static, R - RIP,
       O - OSPF, I - IS-IS, B - BGP, E - EIGRP, N - NHRP,
       T - Table, v - VNC, V - VNC-Direct, A - Babel, D - SHARP,
       F - PBR,
       > - selected route, * - FIB route

O>* 1.1.1.1/32 [110/120] via 10.115.0.1, eth2, label 16, 00:00:15
O>* 2.2.2.2/32 [110/20] via 10.115.0.1, eth2, label implicit-null, 00:00:15
O   3.3.3.3/32 [110/10] via 0.0.0.0, loopback1 onlink, 00:01:19
C>* 3.3.3.3/32 is directly connected, loopback1, 00:01:29
O>* 10.0.2.0/24 [110/11] via 10.115.0.1, eth2, label implicit-null, 00:00:15
O   10.100.0.0/24 [110/10] is directly connected, eth1, 00:00:32
C>* 10.100.0.0/24 is directly connected, eth1, 00:00:32
O   10.115.0.0/24 [110/10] is directly connected, eth2, 00:00:25
C>* 10.115.0.0/24 is directly connected, eth2, 00:00:32
O>* 10.135.0.0/24 [110/110] via 10.115.0.1, eth2, label implicit-null, 00:00:15
O>* 10.200.0.0/24 [110/210] via 10.115.0.1, eth2, label 17, 00:00:15
north-vm#

Additional example demonstrating use of some miscellaneous config options:

interface eth0
!
interface eth1
!
interface lo
!
mpls ldp
 dual-stack cisco-interop
 neighbor 10.0.1.5 password opensourcerouting
 neighbor 172.16.0.1 password opensourcerouting
 !
 address-family ipv4
  discovery transport-address 10.0.1.1
  label local advertise explicit-null
  !
  interface eth0
  !
  interface eth1
  !
 !
 address-family ipv6
  discovery transport-address 2001:db8::1
  !
  interface eth1
  !
 !
!
l2vpn ENG type vpls
 bridge br0
 member interface eth2
 !
 member pseudowire mpw0
  neighbor lsr-id 1.1.1.1
  pw-id 100
 !
!