MPLS and LDP: Label Distribution and Label Switching Paths

MPLS and LDP: Label Distribution and Label Switching Paths

This guide covers RouterOS 7 MPLS fundamentals and the Label Distribution Protocol (LDP): enabling MPLS forwarding on interfaces, configuring LDP for automatic label exchange, verifying label-switched paths (LSPs), and reading the Label Forwarding Information Base (LFIB).

MPLS separates the IP routing decision from the per-hop forwarding action. Once a packet enters an MPLS domain, routers forward it by swapping short fixed-length labels rather than performing a full IP lookup at every hop. LDP automates the process of distributing those labels between routers so that end-to-end LSPs form automatically from the IGP routing table.

Overview

Label Operations

Every router in an MPLS domain performs one of three label operations on each packet:

Operation	Router Role	Action
Push	Ingress LER	Attach an MPLS label to an incoming IP packet, creating an MPLS packet.
Swap	Transit LSR	Replace the top label with a new label and forward to the next hop.
Pop	Egress LER	Remove the MPLS label and deliver the original IP packet.

Router Roles

• LER (Label Edge Router) — sits at the boundary between the IP and MPLS domains. The ingress LER pushes the first label; the egress LER pops the final label.
• LSR (Label Switch Router) — a core router that only swaps labels. It never examines the inner IP header.

Penultimate Hop Popping (PHP)

By default, RouterOS uses PHP: the router before the egress LER pops the label, forwarding a plain IP packet to the egress. This removes one lookup at the egress router. The label value 3 (implicit null) signals PHP to the penultimate hop.

Label Distribution Protocol (LDP)

LDP automates LSP creation from the IGP topology:

1. Hello discovery — LDP sends UDP hellos to the multicast address 224.0.0.202 on port 646 to discover directly connected LDP neighbors.
2. Session establishment — neighbors negotiate a TCP session on port 646 using their LSR-IDs (typically loopback addresses) as transport endpoints.
3. Label advertisement — each router assigns a local label to every IGP prefix it knows and advertises the binding to all LDP neighbors. Neighbors chain these bindings into end-to-end LSPs.

Prerequisites

• RouterOS 7.x (MPLS features are available by default; verify with /system/package/print)
• An IGP (OSPF or IS-IS) advertising loopback /32 addresses and transport subnets between all MPLS routers
• A stable loopback interface on each router to serve as the LDP LSR-ID and OSPF router-id
• IP connectivity between all LDP transport addresses before enabling LDP

Note

LDP builds LSPs for all IGP-known prefixes by default. Ensure your IGP is stable before enabling LDP to avoid excessive label churn.

Configuration

The examples below use a three-router linear topology matching the diagram above:

Router	Role	Loopback	Link to next-hop
R1	Ingress LER	1.1.1.1/32	ether1 — 10.0.0.0/30 → R2
R2	Transit LSR	2.2.2.2/32	ether1 — 10.0.0.0/30 → R1, ether2 — 10.0.0.4/30 → R3
R3	Egress LER	3.3.3.3/32	ether1 — 10.0.0.4/30 → R2

Note

The interface names above assume a dedicated lab topology. On the staging CHR routers used for validation, ether1 is the management interface on 192.168.86.0/23, so reproduce the core links on non-management ports instead (typically ether2 and ether3) while keeping the same MPLS/LDP configuration.

Step 1: Configure Loopback and IGP Underlay

Each router needs a loopback with a /32 address and an IGP advertising it. OSPF is used here; IS-IS works equally well. If your lab uses different physical port numbers, substitute your actual core-facing interfaces in the commands below.

R1:

# Create loopback bridge
/interface bridge add name=loopback

# Assign LSR-ID address
/ip address add address=1.1.1.1/32 interface=loopback

# Configure OSPF with router-id matching the loopback
/routing ospf instance add name=default router-id=1.1.1.1
/routing ospf area add name=backbone area-id=0.0.0.0 instance=default
/routing ospf interface-template add area=backbone interfaces=loopback passive
/routing ospf interface-template add area=backbone interfaces=ether1

R2:

/interface bridge add name=loopback
/ip address add address=2.2.2.2/32 interface=loopback
/routing ospf instance add name=default router-id=2.2.2.2
/routing ospf area add name=backbone area-id=0.0.0.0 instance=default
/routing ospf interface-template add area=backbone interfaces=loopback passive
/routing ospf interface-template add area=backbone interfaces=ether1,ether2

R3:

/interface bridge add name=loopback
/ip address add address=3.3.3.3/32 interface=loopback
/routing ospf instance add name=default router-id=3.3.3.3
/routing ospf area add name=backbone area-id=0.0.0.0 instance=default
/routing ospf interface-template add area=backbone interfaces=loopback passive
/routing ospf interface-template add area=backbone interfaces=ether1

Verify IGP reachability before proceeding:

# On R1 — should resolve to R3 loopback via OSPF
/ip route print where dst-address=3.3.3.3/32
/ping 3.3.3.3 count=3

Step 2: Enable MPLS on Interfaces

MPLS must be enabled on the physical interfaces that carry labeled traffic. Do not enable MPLS on the loopback itself. On staging CHR, that means using the non-management lab interfaces rather than ether1.

All routers — enable MPLS on core-facing interfaces:

# R1
/mpls interface add interface=ether1

# R2
/mpls interface add interface=ether1
/mpls interface add interface=ether2

# R3
/mpls interface add interface=ether1

Verify:

/mpls interface print

Expected output shows each interface with MPLS flag set.

Step 3: Configure LDP

LDP requires a global LSR-ID (the loopback address) and per-interface enabling. LDP uses the LSR-ID as the transport address for TCP sessions, so it must match an address reachable via IGP. As with MPLS, replace the example interface names with your actual core-facing ports when ether1 is reserved for management.

All routers:

# R1 — create LDP instance with LSR-ID and transport address
/mpls ldp add afi=ip lsr-id=1.1.1.1 transport-addresses=1.1.1.1
/mpls ldp interface add interface=ether1

# R2
/mpls ldp add afi=ip lsr-id=2.2.2.2 transport-addresses=2.2.2.2
/mpls ldp interface add interface=ether1
/mpls ldp interface add interface=ether2

# R3
/mpls ldp add afi=ip lsr-id=3.3.3.3 transport-addresses=3.3.3.3
/mpls ldp interface add interface=ether1

Step 4: Verify LDP Neighbor Sessions

LDP discovery starts within seconds of enabling LDP interfaces. Check for established sessions:

/mpls ldp neighbor print detail

A healthy neighbor entry shows the O (OPERATIONAL) flag and lists the remote transport address, local transport, peer, and addresses. If the O flag is absent (only C = active-connect or W = passive-wait), the TCP session has not formed — check IGP reachability to the remote LSR-ID.

Verification

Read the Label Forwarding Information Base (LFIB)

The LFIB is the active forwarding table used for labeled packet switching:

/mpls forwarding-table print detail

Key fields:

Field	Meaning
in-label	Label value matched on incoming packets
out-label	Label pushed on outgoing packets (3 = PHP/implicit-null, 0 = explicit-null)
interface	Outgoing interface
nexthop	Next-hop IP address
operation	swap, pop, or push

Verify LSP End-to-End

From the ingress LER, confirm a full LSP exists to the egress loopback:

# On R1 — confirm LFIB has a forwarding entry for R3 loopback
/mpls forwarding-table print where nexthop~"10.0.0"

If R2 is performing PHP, traffic from R1 will carry a label all the way to R2, where R2 pops it and sends a plain IP packet to R3. This is correct and expected behaviour.

Check Global MPLS Settings

/mpls settings print

Review propagate-ttl (whether IP TTL is decremented through the MPLS domain) and allow-fast-path (hardware/fast-path label switching).

Practical Examples

Example 1: Full LSP Verification on a Three-Router Chain

After completing the configuration above, run these checks on each router:

R1 (Ingress LER):

# LDP neighbor to R2 should be established
/mpls ldp neighbor print

# LFIB should push R2's advertised label for traffic to 3.3.3.3/32
/mpls forwarding-table print detail

R2 (Transit LSR):

# Two neighbors: R1 and R3
/mpls ldp neighbor print

# LFIB shows swap or pop operations for each prefix
/mpls forwarding-table print detail

R3 (Egress LER):

# LDP neighbor to R2
/mpls ldp neighbor print

# LFIB — 3.3.3.3/32 should show PHP (label 3/implicit-null) triggering pop at R2
/mpls forwarding-table print detail

Example 2: Targeted LDP Sessions

For applications like VPLS that require LDP sessions between non-adjacent routers, use targeted (unicast) hellos:

# On R1 — establish targeted LDP session to R3 loopback
/mpls ldp neighbor add transport=3.3.3.3

Targeted sessions use the same TCP port 646 but send unicast hellos instead of multicast. Verify with:

/mpls ldp neighbor print detail where transport=3.3.3.3

Troubleshooting

LDP Neighbors Not Forming

# 1. Verify MPLS is enabled on the interface
/mpls interface print detail

# 2. Verify LDP is enabled on the interface
/mpls ldp interface print detail

# 3. Check LDP global settings — lsr-id must be set
/mpls ldp print

# 4. Confirm IGP reachability to the remote LSR-ID
/ip route print where dst-address=<remote-lsr-id>/32
/ping <remote-lsr-id>

If /ping to the remote LSR-ID fails, fix the IGP underlay first. LDP TCP sessions cannot establish without routable transport addresses.

Labels Not Being Distributed

# Confirm session is established (not just discovered)
/mpls ldp neighbor print detail

# Verify OSPF is advertising all loopbacks and transport subnets
/routing ospf neighbor print
/ip route print where protocol=ospf

# Check LFIB — if empty, LDP session may be down or no IGP routes are present
/mpls forwarding-table print detail

Traffic Not Being Label-Switched

# Confirm LFIB has entries for destination prefixes
/mpls forwarding-table print detail

# Check if fast-path is enabled (hardware switching)
/mpls settings print

# Trace label path
/tool traceroute <destination> use-dns=no

Common Issues

Symptom	Likely Cause	Resolution
state=connecting in neighbor table	No IP route to remote LSR-ID	Fix IGP; ensure loopbacks are advertised
Empty LFIB after session up	No LDP instance configured	/mpls ldp add afi=ip lsr-id=<id> transport-addresses=<id>
LFIB missing entries	MPLS not enabled on interface	/mpls interface add interface=<iface>
PHP not occurring at penultimate hop	Remote router not advertising implicit-null	Check egress router LDP binding for label=3 on its own prefix
Labels assigned but traffic drops	MTU too small for MPLS overhead	Add 4 bytes per label to interface MTU; check /mpls settings

Command Reference

/mpls interface

Enables MPLS forwarding on a physical interface.

Parameter	Description
interface	Interface name

/mpls interface add interface=ether1
/mpls interface print detail
/mpls interface remove [find interface=ether1]

/mpls ldp

LDP instance configuration. LDP is enabled by creating an instance (one per address-family/VRF). There is no global on/off toggle — the instance existing is what enables LDP.

Parameter	Description
afi	Address family (ip for IPv4)
lsr-id	Router LSR-ID, used as LDP router identifier — typically loopback /32
transport-addresses	TCP session endpoint address — typically the same loopback /32

/mpls ldp add afi=ip lsr-id=1.1.1.1 transport-addresses=1.1.1.1
/mpls ldp print

/mpls ldp interface

Enables LDP on a specific interface for neighbor discovery.

/mpls ldp interface add interface=ether1
/mpls ldp interface print detail

/mpls ldp neighbor

View discovered and established LDP neighbors. Use add transport= to configure a static targeted neighbor.

/mpls ldp neighbor print detail
/mpls ldp neighbor add transport=3.3.3.3

/mpls forwarding-table

Active LFIB — the table used for label-based packet forwarding decisions.

/mpls forwarding-table print detail
/mpls forwarding-table print where in-label=100

/mpls settings

Global MPLS behaviour flags.

/mpls settings print
/mpls settings set propagate-ttl=yes

Related Information

Related Topics

• VPLS — Layer 2 VPN services built on top of MPLS LDP LSPs
• MPLS Traffic Engineering and RSVP-TE — TE tunnels, explicit paths, and bandwidth reservation with RSVP-TE
• MPLS-TE Explicit Paths — Detailed explicit path configuration and ERO notation
• MPLS-TE Hop Types — Strict and loose hop configuration for TE tunnels
• OSPF Configuration — IGP underlay required for LDP LSR-ID reachability

RFC References

• RFC 5036 — LDP Specification (core protocol)
• RFC 3031 — MPLS Architecture
• RFC 3032 — MPLS Label Stack Encoding
• RFC 3443 — TTL Processing in Multi-Protocol Label Switching (MPLS) Networks