Introduction to MPLS

Luc De Ghein
ldeghein@cisco.com

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

Agenda
History of MPLS
Benefits of MPLS MPLS Technology Introduction

MPLS Applications

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

Before MPLS
Several WAN protocols existed
ATM, Frame-Relay

They were cost-effective

Lacked ease of deployment, provisioning, and management

IP was winning the battle Ethernet was cheaper and easier than ATM People began to look for
a good integration of IP over ATM an easy way to deploy virtual private networks over an IP backbone
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

History of MPLS

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009+

Time

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

Benefits of MPLS
One unified network infrastructure
Better IP over ATM integration BGP-free core
economics

Peer-to-peer Virtual Private Network (VPN) model

Optimal traffic flow across provider network Flexible way to do traffic engineering

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

One Unified Network Infrastructure

MPLS = Multi Protocol Label Switching An MPLS backbone is an enabler for multiservice Carry all kinds of traffic across one MPLS enabled network :
IPv6 IP VPN ATM

IPv4 IPv6 Layer 2 frames (Ethernet, ATM, FrameRelay, HDLC, PPP) TDM

PPP

IP/MPLS

Frame Relay

Internet VoIP

Ethernet

Adding labels to the packet enables the possibility to carry other protocols than just IP over an MPLS-enabled Layer 3 IP backbone, similarly to what was previously only possible with Frame Relay or ATM Layer 2 networks

PSTN

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

Better IP over ATM Integration

Before MPLS, there were three ways to implement MPLS:
(RFC) 1483, Multiprotocol Encapsulation over ATM Adaptation Layer 5,
all ATM circuits had to be manually established and all mappings between IP next hops and ATM endpoints had to be manually configured on every ATM-attached router in the network

LANE (LAN Emulation)

this technology never achieved the scalability or reliability requirements of large service provider networks

MPOA (Multiprotocol over ATM )

the tightest integration of IP over ATM, but also the most complex solution

All these methods were cumbersome to implement and troubleshoot. A better solution for integrating IP over ATM was one of the driving reasons for the invention of MPLS. The prerequisites for MPLS on ATM switches were that the ATM switches had to become more intelligent. The ATM switches had to run an IP routing protocol and implement a label distribution protocol.

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

BGP-Free Core
MPLS network
BGP Route Reflector (RR)

BGP sessions

edge MPLS router edge MPLS router

BGP-free core
edge MPLS router edge MPLS router

MPLS labeling is done on edge routers the label assigned is the one associated with the BGP next-hop address The BGP next-hop address is known in the network via the IGP

Forwarding on core routers is done by looking at MPLS label there is no IP lookup

Core routers do not need to run BGP
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

The Peer-to-Peer Model for MPLS VPN

A Virtual Private Network (VPN) is a network that emulates a private network over a common infrastructure.
In the overlay model:
Examples are ATM and Frame-Relay The service provider provides a service of point-to-point links or virtual circuits across his network between the routers of the customer The customer routers form routing peering between them directly across the links or virtual circuits from the service provider

In the peer-to-peer VPN model:

The service providers routers carry the customers data across the network, but they also participate in the customers routing Easier provisioning Adding one customer site means that on the PE router only the peering with the CE router must be added There is no hassle with creating many virtual circuits as with the overlay model or with configuring packet filters or route filters with the peer-to-peer VPN model over an IP network
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

The Overlay Model for VPN

service providers Frame Relay or ATM network virtual circuit

VPN red customer router

Frame Relay or ATM switch

VPN green customer router

VPN red customer router

VPN green customer router

VPN green customer router VPN red customer router

In the overlay model:

The service provider provides a service of point-to-point links or virtual circuits across his network between the routers of the customer The customer routers form routing peering between them directly across the links or virtual circuits from the service provider
Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

10

Peer-to-Peer Model for MPLS VPN

Customer Edge router

routing peering
Provider Edge router

routing peering routing peering = iBGP

Customer Edge router Provider Edge router

routing peering

Provider Edge router

Customer Edge router

In the peer-to-peer model:

One peering: between the CE router and the PE router Internal BGP takes care of the VPN routing in SP network
Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

11

Optimal Traffic Flow

Frame Relay or ATM switch Frame Relay or ATM switch VPN red CE VPN red CE

VPN red CE

VPN red CE

VPN red CE VPN red CE VPN red CE

VPN red CE

VPN red CE

VPN red CE

VPN red CE

VPN red CE

overlay model

peer-to-peer model

Layer 2 devices in the core: customer routers interconnect through them by means of virtual circuits (VC) created In order for any router to send traffic directly to any other router at the edge, a virtual circuit must be created between them directly For optimal traffic flow in all cases: full mesh needed For n customer routers: (n-1) * n / 2 number of VCs needed

Peer-to-peer MPLS VPN model has optimal traffic flow in all cases No VCs per customer

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

12

Traffic Engineering MPLS TE

MPLS TE = optimally use the network
TE looks how much BW is free on any link Use underutilized links TE must provide the possibility to steer traffic through the network on paths different than the preferred path (least cost path)

MPLS network with TE enabled

IP

least cost path

C
IP

B E F

traffic engineered path

If this network is an IP only network, you could never have router B send the traffic along the bottom path by configuring anything on router A. Router Bs decision to send traffic on the top or bottom path is solely its own decision. If you enable MPLS traffic engineering in this network, you can have router A send the traffic towards router D along the bottom path. The MPLS TE forces router B to toward the traffic A-D onto the bottom path. This can be done in MPLS becauserights the label forwarding mechanism. 2009 Cisco Systems, Inc. All of reserved. Cisco Public

Intro to MPLS Luc De Ghein

13

MPLS Technology Introduction

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

14

MPLS Label and Label Encapsulation

MPLS Label
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1

Label # 20bits

EXP S

TTL-8bits

COS/EXP = Class of Service: 3 Bits; S = Bottom of Stack; TTL = Time to Live

MPLS Label Encapsulation

PPP Header (Packet over SONET/SDH) LAN MAC Label Header
Intro to MPLS Luc De Ghein

PPP Header

Label

Layer 2/L3 Packet

MAC Header

Label

Layer 2/L3 Packet

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

15

Label Stacking
There may be more than one label in an MPLS packet
Each label at a different place in the label stack has its meaning
Examplethere can be one label for routing the packet to an egress point and another that separates a customer A packet from customer B Inner labels can be used to designate services

e.g. L3VPNs, L2VPN

Outer label used to route/switch the MPLS packets in the network Last label in the stack is marked with EOS bit Allows building services such as
MPLS VPNs
Traffic engineering and fast reroute VPNs over traffic engineered core Any transport over MPLS Inner Label
Intro to MPLS Luc De Ghein

Outer Label TE Label LDP Label

VPN Label
IP Header

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

16

MPLS Label
MPLS Label has local significance
One router assigns the MPLS label independently There is no global assignment for the whole network
No global authority

20 bits for the label gives label range of 0-1048575

Default label range might be lower Label range is limited on some platforms

Normal MPLS labels are: 16-1048575 Reserved label range is: 0-15
See later slides for some examples
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

17

LDP Label Advertisment

IGP
LDP label advertisement
10.0.0.0/8 label L1

LDP
LDP label advertisement
10.0.0.0/8 label L2

LDP label advertisement

10.0.0.0/8 label L3 10.0.0.0/8

PE

PE

Local/In label
L1

Intro to MPLS Luc De Ghein

Prefix
10.0.0.0/8

Out Intf
POS0/0/0

Remote/Out Label
L2

LFIB

LDP = Label Distribution Protocol, defined in RFC 3035 and 3036 LDP advertises label bindings Label binding = IP prefix + MPLS label LDP is a superset of Tag Distribution Protocol
2009 Cisco Systems, Inc. All rights reserved. Cisco Public

18

Label Packet Forwarding

IGP
IP IP Label L1 IP Label L2

LDP
IP Label L3 IP

PE

PE

Local/In label
L1

Prefix
10.0.0.0/8

Out Intf
POS0/0/0

Remote/Out Label
L2

swapping incoming label with outgoing label

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

19

Control and Forward Plane Separation

control plane control plane used to distribute labels and build label-switched paths (LSPs)
RIB Routing Process

Route Updates/ Adjacency Label Binding Updates/ Adjacency

LIB

MPLS Process

forwarding plane forwarding plane used to forward IP or labeled packets

LFIB

FIB

MPLS Traffic

IP Traffic

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

20

LIB
Label Information Base holds the label bindings One local label (in label) allocated by the router per prefix One or more remote labels (out label) per prefix LIB stores all received label bindings
PE1#show mpls ldp binding tib entry: 10.1.1.0/24, rev 3 local binding: tag: imp-null remote binding: tsr: 10.100.1.3:0, tag: imp-null remote binding: tsr: 10.100.1.6:0, tag: 16 tib entry: 10.1.5.0/24, rev 20 local binding: tag: imp-null remote binding: tsr: 10.100.1.3:0, tag: 18 remote binding: tsr: 10.100.1.6:0, tag: imp-null tib entry: 10.100.1.2/32, rev 4 local binding: tag: imp-null remote binding: tsr: 10.100.1.3:0, tag: 16 remote binding: tsr: 10.100.1.6:0, tag: 19 tib entry: 10.100.1.4/32, rev 18 local binding: tag: 22 remote binding: tsr: 10.100.1.3:0, tag: 19 remote binding: tsr: 10.100.1.6:0, tag: 21
Intro to MPLS Luc De Ghein

one LIB entry

...
2009 Cisco Systems, Inc. All rights reserved. Cisco Public

21

LFIB
Label Forwarding Information Base LFIB stores one local and one remote label per prefix LFIB is used to forward labeled incoming packet
Outgoing packet can be labeled Outgoing packet can be unlabeled

PE1#show mpls forwarding-table

Local tag 16 17 18 19 Outgoing tag or VC Pop tag Pop tag Pop tag Pop tag Prefix or Tunnel Id 10.1.2.0/24 10.1.3.0/24 10.1.4.0/24 10.100.1.3/32 Bytes tag switched 0 0 0 0 Outgoing interface Et0/0 Se3/0 Se3/0 Se3/0 10.1.5.6 point2point point2point point2point Next Hop

20
21 22 23

Pop tag
Aggregate 19 Untagged

10.100.1.6/32
11.1.1.0/24[V] 10.100.1.4/32 11.100.1.1/32[V]

0
0 0 0

Et0/0
Se3/0 Se2/0

10.1.5.6
point2point point2point

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

22

Label Forwarding Operations

SWAP PUSH
55 16 34 16 34

IP

IP

IP

IP

POP
23 16 16

UNTAG/ NO LABEL
23 16

IP

IP

IP

IP

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

23

Implicit Null Label -> PHP

Penultimate Hop Popping (PHP) removing the MPLS label at the one-but-last MPLS router

IP

IP

Label L1

IP

Label L2

IP

10.100.1.0/24

PE

P
LDP label advertisement
10.0.0.0/8 label 3

PE

penultimate hop router

Implicit null label (label 3) is used in a few cases

for example for connected routes

Purpose: avoid double lookup: first MPLS lookup, followed by IP lookup

P#show Local tag 17 mpls forwarding-table 10.100.1.0 255.255.255.0 Outgoing Prefix Bytes tag Outgoing tag or VC or Tunnel Id switched interface Pop tag 10.100.1.0/24 5948 Se3/0 Next Hop point2point

POP is outgoing label in LFIB

(no label is added in the label stack)

P#show mpls ldp bindings 10.100.1.0 24 tib entry: 10.100.1.0/24, rev 14 local binding: tag: 17 remote binding: tsr: 10.100.1.6:0, tag: 19 remote binding: tsr: 10.100.1.1:0, tag: imp-null remote binding: tsr: 10.100.1.4:0, tag: 19 Intro to MPLS
Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

24

Explicit Null Label

Label 0 is advertised Label 0 is put in the label stack Double lookup needed Label 0 does not entail a forwarding vector, but QOS information (EXP bits) is used of explicit null label
The QOS information is retained till the last hop, whereas with PHP, the QOS information in the top label is lost, because the top label was popped

replacing label with explicitnull label

! mpls label protocol ldp mpls ldp explicit-null !

IP

IP

Label L1

IP

Label L2

IP

Label 0

10.100.1.0/24

PE

P
LDP label advertisement
10.0.0.0/8 label 0

PE
P#show mpls ldp bindings 10.100.1.0 24 tib entry: 10.100.1.0/24, rev 14 local binding: tag: 17 remote binding: tsr: 10.100.1.6:0, tag: 19 remote binding: tsr: 10.100.1.4:0, tag: 19 remote binding: tsr: 10.100.1.1:0, tag: exp-null P#show Local tag 17 mpls forwarding-table 10.100.1.0 Outgoing Prefix Bytes tag tag or VC or Tunnel Id switched 0 10.100.1.0/24 0

penultimate hop router

Outgoing interface Se3/0

Next Hop point2point

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

25

Overview of MPLS Applications

MPLS Layer 3 VPNs
MPLS Layer 2 VPNs
Point-to-point Point-to-multipoint

MPLS Traffic Engineering

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

26

MPLS VPN Control Plane

iBGP exchanging vpnv4 prefixes + MPLS label

red VPN

eBGP IGP

eBGP IGP red VPN

CE

PE

IGP LDP

IGP LDP

IGP LDP

PE

CE

VRF interface

VRF interface

Route Distinguisher (RD): 8-byte fieldunique value assigned by a provider to each VPN to make different VPN routes unique VPNv4 address: RD+VPN IP prefix Route Target (RT): 8-byte field, unique value assigned by a provider to define the import/export rules for the routes from/to each VPN MP-iBGP: facilitates advertisement of VPNv4* prefixes + labels between BGP peers Virtual Routing Forwarding Instance (VRF): contains VPN site routes; only on PE routers Intro MPLS to Multi-VRF CE (VRF-Lite): CE device supporting multiple VRFs w/o MP-iBGP & VPN labels 27
Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

MPLS VPN Data Forwarding

Penultimate Hop Popping (PHP)

IGP red VPN

IP IP VPN Label Label L2 IP VPN Label Label L1

LDP
IP VPN Label IP

red VPN

CE

PE

PE

CE

Ingress PE router: lookup in VRF RIB, adds vpn label, add LDP label P routers: label swapping (top label only) Egress PE router: looks up vpn label in LFIB, forwards IP packet onto VRF interface

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

28

L2VPN Options
L2VPN Models VPWS
Virtual Private Wire Service
Point to Point

VPLS
Virtual Private LAN Service
Point to Multipoint
MPLS Core

L2TPv3
IP Core

AToM
MPLS Core

Ethernet

Ethernet
Frame Relay ATM (AAL5 and Cell) PPP and HDLC
Intro to MPLS Luc De Ghein

Ethernet
Frame Relay ATM (AAL5 and Cell) PPP and HDLC
AToM = Any Transport over MPLS

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

29

Any Transport over MPLS Architecture IETFs Pseudo-Wire Reference Model

IETF working group PWE3
Pseudo wire emulation edge to edge requirements detailed in draft-ietf-pwe3-requirements now RFC3916 draft-ietf-pwe3-architecture(framework) now RFC3985

AC

PSN Tunnel
pseudo wires

AC

AC

PE

PE

AC

emulated services

The pseudowire (PW) is a connection between 2 PE routers emulating an end-to-end service and connecting 2 Attachments Circuits (AC)
2009 Cisco Systems, Inc. All rights reserved. Cisco Public

Intro to MPLS Luc De Ghein

30

IETFs L2VPN Logical Context

Provider Edge SP Network SP Interconnection

Provider Edge

Pseudo Wire Many Subscriber Encapsulations Supportable

ATM HDLC Ethernet

FR

PPP

An L2VPN is comprised of switched connections between subscriber endpoints over a shared network Non-subscribers do not have access to those same endpoints
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

31

AToM Building Blocks

Targeted LDP (Label Distribution Protocol) Session
Used for VC-label negotiation, withdrawal, error notification

Control Connection

Transport Component

Tunnel header (tunnel label)

To get PDU from ingress to egress PE using MPLS LSP

Tunneling Component

Demultiplexer field (VC label)

To identify individual circuits within a tunnel (VC label is a MPLS Label)

Emulated L2 PDU

Emulated VC encapsulation (control word)

Information on enclosed Layer 2 PDU 4 bytes sitting in between the label stack and the MPLS payload

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

32

Any Transport over MPLS Control Plane How PWs Are Established
2. PE1 Starts Targeted LDP session with PE2 if one does not already exist

CE

PE1

PE2

CE

1. CE-PE: AC connection 3. PE1 Allocates VC Label for new interface and binds to configured VC ID 5. PE2 Receives VC Type and VC Label that matches Local VCID Note: PE2 Repeats Steps 15 so that Bidirectional Label/VCID Mappings Are Established
Cisco Public

4. PE1 sends label mapping message to PE2 over LDP session

Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved.

33

Any Transport over MPLS How Traffic Is Forwarded on an Emulated Circuit

L2 frame

L2 frame VC Label Label L2

L2 frame VC Label Label L1

L2 frame VC Label

L2 frame

CE

PE1

PE2

CE

VC label LDP label

VC label is only looked up at egress PE Tunnel label (LDP/IGP label) is changed at every hop

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

34

Virtual Private LAN Service Overview

Site1 CE

PE1

PE2

Site2 CE

MPLS WAN

Site3 CE VPLS defines an architecture that delivers Ethernet Multipoint Services (EMS) over an MPLS network VPLS operation emulates an IEEE Ethernet bridge. VPLS network acts like a virtual switch that emulates conventional L2 bridge.
Forwarding of ethernet frames Forwarding of unicast frames with unknown destination MAC address Forwarding of multicast and broadcast frames Dynamic learning of MAC addresses MAC address aging

It supports communication between fully meshed L2 sites without the spanning tree complexities Intro to MPLS
Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

35

VPLS Components
full mesh of Targeted-LDP sessions exchange VC labels Attachment Circuit
n-PE PW CE PW CE Red VSI Blue VSI Green VSI Directed LDP session between participating PEs Tunnel LSP PW CE CE Red VSI Blue VSI Green VSI
CE

CE

n-PE

CE

Full mesh of PWs between VSIs

Virtual Switching Instance: VSI or VFI (Virtual Forwarding Instance) VPN ID: Unique value for each VPLS VPN Attachment VCs are port mode or VLAN ID
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

36

MPLS TE Deployment Models

Bandwidth Optimization
Strategic
R1 IP/MPLS R1

Tactical
IP/MPLS

R8 R2 R2

R8

deploy TE everywhere; optimise BW everywhere

Protection
R1 IP/MPLS

move BW away from congested points Point-to-Point SLA

R1 IP/MPLS

R8 R2 R2

R8

Fast ReRouting (FRR)

Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

Virtual Leased Lines; QOS; BW guarantees

37

How MPLS TE Works

Link information Distribution
Head end

ISIS-TE
IP/MPLS

OSPF-TE

Path Calculation (CSPF)

Path Setup (RSVP-TE) Forwarding Traffic down Tunnel
Auto-route Static PBR
Mid-point Tail end

CBTS

Forwarding Adjacency
Tunnel select
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

38

TE Fundamentals
1. Information Distribution
need Link State protocol IS-IS or OSPF

2. 3. 4. 5.

Path selection/calculation Path setup Trunk admission control Forwarding traffic on to tunnel 6. Path maintenance headend

IGP (OSPF or ISIS) used to flood BW information

tailend midpoints

Path Calculation (PCALC/CSPF) uses IGP advertisements to compute constrained paths TE tunnel

RSVP/TE used to distribute labels, provide CAC, failure notification, etc.

Upstream
Intro to MPLS Luc De Ghein 2009 Cisco Systems, Inc. All rights reserved. Cisco Public

Unidirectional Tunnel

Downstream
39

Path Setup Example

PCALC calculates path on headend or explicit path configured PATH messages are sent with requested bandwidth RESV messages are sent with MPLS label for the TE tunnel There is admission control at each hop to see if the bandwidth requirement can be met
headend RESV RESV PATH PATH RESV tailend

PATH

TE tunnel

Headend router has view of complete network topology in TE database

Intro to MPLS Luc De Ghein

thanks to link state topology on headend router

2009 Cisco Systems, Inc. All rights reserved. Cisco Public

40

MPLS TE Fast Re-Route (FRR)

IP/MPLS
R1

Subsecond recovery against node/link failures

R8

Fast because backup tunnel is pre-signaled

Scalable 1:N protection Greater protection granularity Cost-effective alternative to optical protection Bandwidth protection

R2

Primary TE LSP Backup TE LSP

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

41

FRR Link Protection Operation

Requires next-hop (NHOP) backup tunnel Point of Local Repair (PLR) swaps label and pushes backup label
R1 R2
25 22 22

IP/MPLS
R3

R6

R7

Backup terminates on Merge Point (MP) where traffic rejoins primary Restoration time expected under ~50 ms

16

22

R5

Primary TE LSP
Backup TE LSP
Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

42

FRR Node Protection Operation

Requires next-next-hop (NNHOP) backup tunnel
IP/MPLS
R3
25 36 36

Point of Local Repair (PLR) swaps next-hop label and pushes backup label
Backup terminates on Merge Point (MP) where traffic rejoins primary Restoration time depends on failure detection time

R1

R2

R4

R5

R6

16

22

36

R5

Primary TE LSP
Backup TE LSP
Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

43

Path Protection Operation

No local repair Requires second end-toend signalled TE LSP
IP/MPLS
R5 R6 R7

Point of Repair (PLR) is headend router

Restoration time expected under ~200 ms

R1

R2

R3

R4

Primary TE LSP
Backup TE LSP
Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

44

Q&A

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

45

Terminology Reference
Terminology
AC
ECMP IGP Equal Cost Multipath Interior Gateway Protocol

Description
Attachment Circuit. An AC Is a Point-to-Point, Layer 2 Circuit Between a CE and a PE.

LAN
LDP LER LFIB LIB LSP LSR P Router PE Router PSN Tunnel

Local Area Network

Label Distribution Protocol, RFC 3036. Label Edge Router. An Edge LSR Interconnects MPLS and non-MPLS Domains. Labeled Forwarding Information Base Labeled Information Base Label Switched Path Label Switching Router An Interior LSR in the Service Provider's Autonomous System An LER in the Service Provider Administrative Domain that Interconnects the Customer Network and the Backbone Network. Packet Switching Tunnel

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

46

Terminology Reference
Terminology
Pseudo-Wire PWE3 RD RIB RR RT RSVP-TE VPN VFI

Description
A Pseudo-Wire Is a Bidirectional Tunnel" Between Two Features on a Switching Path. Pseudo-Wire End-to-End Emulation Route Distinguisher Routing Information Base (Routing Table) Route Reflector Route Target Resource Reservation Protocol based Traffic Engineering Virtual Private Network Virtual Forwarding Instance

VPLS
VPWS VRF VSI

Virtual Private LAN Service

Virtual Private WAN Service Virtual Route Forwarding Instance Virtual Switching Instance

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

47

Intro to MPLS Luc De Ghein

2009 Cisco Systems, Inc. All rights reserved.

Cisco Public

48