Deploy Packet Transport with MPLS-TP - nanog.pdf

Deploy Packet
Deploy Packet
Transport with MPLS-
Transport Profile
Randy Zhang
Randy Zhang
Cisco Systems Advanced Services

Agenda
Agenda
Agenda
Agenda
 Transport Network Transformation
p
 Why MPLS TP?
 MPLS TP Technical Overview
MPLS TP Technical Overview
 Deployment Scenarios
 MPLS TP Deployment Lifecycle
 MPLS TP Deployment Lifecycle
© 2012 Cisco Systems, Inc. All rights reserved.
RZhang_MPLS_TP 2
2

Transport Network
Transport Network
Transport Network
Transport Network
Transformation
Transformation
RZhang_MPLS_TP 3
3

What is Transport Network?
What is Transport Network?
A network to provide a reliable aggregation and
A network to provide a reliable aggregation and
transport infrastructure for any client traffic type
Oh um, please do it at the lowest cost per bit…
RZhang_MPLS_TP 4

Specifically
Specifically
Cost effective
Multi-service
Quality of
service
Scalable
Transport Network
Transport Network
RZhang_MPLS_TP 5

Where Have We Been?
 Looking back in our memory
Where Have We Been?
 Looking back in our memory
for the past several decades
on network events that
ff t d th t t t k
affected the transport network
RZhang_MPLS_TP 6

Networks in the 1980s
1981: first IBM PC
1983: Novell, ISO OSI model, ARPANET runs on TCP/IP
1984: IBM PC AT system, Cisco Systems
y , y
1985: Standardization of Ethernet 10Base2, Sun Micro
NFS, IBM Token Ring
1987: Standardization of SONET, 10BaseT
1988: ATM cell format standardized
Top shows on TV were Dallas and The Cosby Show
RZhang_MPLS_TP 7
p y

1991: LAN switches
1992: Public frame relay service,100 Mbps Ethernet,
Windows 3.1
1994: GPS
1995: Cable modem, VOIP software
1997: Standardization of full duplex Ethernet
1998: Standardization of Gigabit Ethernet
Top shows on TV were Seinfeld, ER
RZhang_MPLS_TP 8
p ,

Transport Network Layers: 1980s 1990s
Transport Network Layers: 1980s – 1990s
O ti i d f i d TDM t ffi
Optimized for voice and TDM traffic
IP E h ATM F R l
SONET/SDH DSx
IP, Ethernet, ATM, Frame Relay
SONET/SDH, DSx
RZhang_MPLS_TP 9
9

Important Network Issues around 2000
 TDM was the primary transport
 TDM was the primary transport
 Business dominated bandwidth
Cli t th ti d l
 Client-server was the computing model
 Internet-based voice applications began to be widespread
RZhang_MPLS_TP 10

Significant Developments 2000 Today
 2001: standardization of G 709
Significant Developments 2000 - Today
 2001: standardization of G.709
 2002: standardization of 10 GigE
2009 t d di ti f MPLS TP
 2009: standardization of MPLS TP
 2010: standardization of 40 GigE and 100 GigE, OTU4
RZhang_MPLS_TP 11

Transport Network Layers: 2000s
Transport Network Layers: 2000s
R t fitti f D t
Retrofitting for Data
IP E h
SONET/SDH DWDM OTN
IP, Ethernet
SONET/SDH, DWDM, OTN
RZhang_MPLS_TP 12
12

 TDM transitioning to Ethernet
 TDM transitioning to Ethernet
 Consumer dominating bandwidth use
P t ti d l d ti th
 Peer-to-peer computing and cloud computing as the new
models
 Internet-based video applications putting demand on
 Internet-based video applications putting demand on
bandwidth
RZhang_MPLS_TP 13

The Explosion of Bandwidth Demand
The Explosion of Bandwidth Demand
The world of
Warcraft players Business video
HDTV
Cloud
computing
4G mobile and
Music and
video
download
4G mobile and
wireless users
RZhang_MPLS_TP 14
14

Bandwidth Availability
Moore’s Law: Computing
Bandwidth Availability
p g
power doubles every 18
months
Nielsen’s Law: Bandwidth
growth for home users doubles
every 21 months
User experience remains bandwidth-bound
RZhang_MPLS_TP 15
p

Why Traditional Transport Is Limited?
 Primary traffic type is now bursty data
Why Traditional Transport Is Limited?
 Primary traffic type is now bursty data
 SONET/SDH is capped at OC-768 (40 Gbps)
T diti l t k i b d TDM
 Traditional network is based on TDM
 TDM is expensive to operate
 Co-existing of multiple transport networks are costly
RZhang_MPLS_TP 16

Entering the Zettabyte Era
IP Traffic Growth
Internet
Video
Mobile
0.2 EB to 1.2 EB per Month
6x Increase from ‘10 to ’15 1 Exabyte = 109 Gigabyte
1 Zettabyte = 1000 Exabyte
RZhang_MPLS_TP 17
Source: Cisco Visual Networking Index (VNI) Global IP Traffic Forecast, 2010–2015

Circuit to Packet Migration
90+%
IP Traffic
Private Line
TDM Traffic
2011 2013 2016
Private Line
TDM Traffic
Private Line
TDM Traffic
Private/Public
~50-70%* 20-30% 0─10%
Private/Public
Private/Public
IP Traffic
~30-50%
IP Traffic
IP Traffic
70-80% 90+%
Legacy TDM
Traffic
 Dramatic shift in SP traffic make-up in next 5 years
 Network evolving
- Transformation: TDM to Packet
Traffic
- Transformation: TDM to Packet
- Convergence: Collapse Layers; IP + Optical Convergence
 SP revenue shifting from circuits to packet services
5 yrs  ~80% revenue derived from packet services
RZhang_MPLS_TP 18
5 yrs  ~80% revenue derived from packet services
Source: ACG Research 2011

Summary of Transport Transformation
 Explosion of data traffic
y p
Drivers
 Convergence of multiple networks into a single transport
network
network
 Reducing CAPEX and OPEX
 Provisioning agility and flexibility
RZhang_MPLS_TP 19

Next Generation Transport
 Packet will be the primary traffic type
Next Generation Transport
 Packet will be the primary traffic type
 Solutions to support packet will depend on cost
MPLS TP ill b th d i t t t
 MPLS-TP will be the predominant core transport
technology
 10/40/100 G DWDM and 10/40/100 G Ethernet on the
 10/40/100 G DWDM and 10/40/100 G Ethernet on the
core
 Circuit services will co-exist with packet services
p
RZhang_MPLS_TP 20

Packet Optical Transport Components
 OTN: a foundation technology for any service over WDM
Packet Optical Transport Components
 OTN: a foundation technology for any service over WDM
 Ethernet: a ubiquitous Layer 2 technology
MPLS TP i MPLS t h l th t id
 MPLS-TP: an emerging MPLS technology that provides
carrier grade transport
 MPLS Pseudowire: A circuit emulation technology based
 MPLS Pseudowire: A circuit emulation technology based
on MPLS
RZhang_MPLS_TP 21

Enabling Technologies
Enabling Technologies
SONET/SDH
SONET/SDH Ethernet
• Lower cost
• Designed for
data
Ethernet
• Lower cost
• Designed for
data
SONET/SDH
• Carrier class
• OAM&P
• QoS
SONET/SDH
• Carrier class
• OAM&P
• QoS data
data
The new
The new
OTN
OTN
MPLS
MPLS OTN
• CWDM and
DWDM
• G.709
OTN
• CWDM and
DWDM
• G.709
MPLS
• Virtual circuit
• Widely
deployed
MPLS
• Virtual circuit
• Widely
deployed
RZhang_MPLS_TP 22
22

A Converged Network
 A single transport network based on WDM
A Converged Network
 OTN provides the digital wrapper
MPLS T t P fil (TP) id SONET lik
 MPLS Transport Profile (TP) provides SONET like
services
 Ethernet technologies provide lower cost in CAPEX and
OPEX
Traditional TDM services and packet based services
p
carried over a single transport network
RZhang_MPLS_TP 23

Why MPLS
Why MPLS TP
TP?
?
RZhang_MPLS_TP 24
24

Motivation for MPLS TP
Motivation for MPLS TP
 Evolution of SONET/SDH transport networks to packet
switching driven by
• Growth in packet-based services (L2/L3 VPN, IPTV, VoIP, etc)
• Desire for bandwidth/QoS flexibility
Desire for bandwidth/QoS flexibility
 New packet transport networks need to retain same
operational model
 MPLS TP, defined jointly between IETF and ITU-T,
provides the next step
RZhang_MPLS_TP 25

Ethernet or MPLS Transport?
Ethernet or MPLS Transport?
 Ethernet
Lack of scalability, traffic engineering, fast protection, circuit
service support
 MPLS
 MPLS
Well accepted by carrier as core IP/MPLS network
More mature carrier-oriented packet technology.
RZhang_MPLS_TP 26

Transport Network Characteristics
Transport Network Characteristics
 Predetermined and long-lived connections
 Emphasis on manageability and deterministic behavior
 Fast fault detection and recovery (sub-50 ms)
y ( )
 In-band OAM
RZhang_MPLS_TP 27

MPLS Network Characteristics
MPLS Network Characteristics
 Dynamically routed label switched paths
 Traffic statistically multiplexed
 Data plane setup and torn down based on dynamic
p p y
control plane
 Optimized for a packet network
RZhang_MPLS_TP 28

Converging MPLS and Transport
Converging MPLS and Transport
MPLS Transport
Profile
IP/MPLS Transport
IP/MPLS
Widely deployed
Carrier grade
Transport
Transport operational model
Static and dynamic
provisioning
Multiservice
Connection oriented path
CAPEX and OPEX savings
Protection switching
triggered by data plane
IP-less transport OAM
functionality
g y
Bidirectional path
RZhang_MPLS_TP 29

Objectives of MPLS-TP
Objectives of MPLS TP
 To enable MPLS to be deployed in a transport network
and operated in a similar manner to existing transport
technologies (SDH/SONET/OTN)
 To enable MPLS to support packet transport services
 To enable MPLS to support packet transport services
with a similar degree of predictability, reliability, and
OAM to that found in existing transport networks
MPLS TP is a subset of MPLS to meet transport network
operational requirements plus additional functionality based on
transport requirements
RZhang_MPLS_TP 30

What is MPLS-TP?
What is MPLS TP?
 MPLS is bi-directional LSPs
 MPLS-TP
No LSP merging
No ECMP (Equal-cost multi-path routing)
Does not support connectionless mode
Simple in scope less complex in operation
Simple in scope, less complex in operation
 OAM/Data Fate sharing with congruent paths
Traffic and OAM must be congruent, achieved by MPLS-TP
g , y
GAL, and generic ACH to carry OAM packets and enable
processing at intermediate nodes when required.
RZhang_MPLS_TP 31

Summary of MPLS TP Characteristics
 Connection-oriented packet switching model
 No modifications to existing MPLS data plane
 IP or IP routing is not required for packet forwarding
 Interoperates/interworks with existing MPLS and
pseudowire control and data planes
RZhang_MPLS_TP 32

 Networks can be created and maintained using static
provisioning (management plane) or a dynamic control
provisioning (management plane) or a dynamic control
plane
 In-band OAM (congruent)
( g )
 Protection options: 1:1, 1+1 and 1:N
 Network operation similar to existing transport networks
Network operation similar to existing transport networks
RZhang_MPLS_TP 33

MPLS-TP: Transport like OAM
MPLS TP: Transport like OAM
 In-band OAM channels
 Performance monitoring for SLA verification
 Sub-path monitoring with multi-level operation
p g p
 Alarms and AIS
RZhang_MPLS_TP 34

MPLS-TP: Transport like Operation
MPLS TP: Transport like Operation
 Data plane / control plane independent
 Transport path fully operational without control plane
 Traffic engineered path control
g p
RZhang_MPLS_TP 35

MPLS-TP: Transport like Protection
MPLS TP: Transport like Protection
 Protection switching triggered by OAM
 Linear protection
 Ring protection
g p
 50 ms switchover
RZhang_MPLS_TP 36

Data Plane
MPLS-TP Summary
Data Plane
MPLS Bidirectional P2P and P2MP LSPs
 No LSP merging
PHP ti l
Control/Management Plane
NMS provisioning option
GMPLS t l l ti
 PHP optional
GACh: Generic Associate Channel
GAL: Generic Associate Label
PW (SS-PW, MS-PW)
MPLS
Forwarding
GMPLS control plane option
PW control plane option
PW (SS PW, MS PW)
MPLS Based
OAM
MPLS
P t ti
OAM Resiliency
OAM Protection
OAM
In-band OAM
Fault management:
 Proactive CC/CV: BFD based
Deterministic path protection
y
 Ping and trace: LSP ping based
 Alarm Suppression and Fault Indication
 AIS, RDI, LDI, and CFI
Sub-50ms switch over
 1:1, 1+1, 1:N protection
 Linear protection
RZhang_MPLS_TP 37
37
Performance monitoring: Loss and Delay  Ring protection

MPLS-TP Standards
MPLS TP Standards
 RFC 6423: Using the Generic Associated Channel Label for
Pseudowire in the MPLS Transport Profile (MPLS-TP)
 RFC 5654: Requirements of an MPLS Transport Profile
 RFC 5718: An In-Band Data Communication Network For the
MPLS Transport Profile
 RFC 5860: Requirements for Operations, Administration, and
Maintenance (OAM) in MPLS Transport Networks
( ) p
 RFC 5951: Network Management Requirements for MPLS-based
Transport Networks
 RFC 5960: MPLS Transport Profile Data Plane Architecture
RZhang_MPLS_TP 38

MPLS-TP Standards
MPLS TP Standards
 RFC 6370: MPLS Transport Profile (MPLS-TP) Identifiers
 RFC 6426: MPLS On Demand Connectivity Verification and Route
 RFC 6426: MPLS On-Demand Connectivity Verification and Route
Tracing
 RFC 6378: MPLS Transport Profile (MPLS-TP) Linear Protection
 RFC 6427: MPLS Fault Management Operations, Administration,
and Maintenance (OAM)
 RFC 6428; Proactive Connectivity Verification Continuity Check
 RFC 6428; Proactive Connectivity Verification, Continuity Check,
and Remote Defect Indication for the MPLS Transport Profile
 RFC 6435: MPLS Transport Profile Lock Instruct and Loopback
F i
Functions
RZhang_MPLS_TP 39

MPLS Transport
MPLS Transport
p
p
Profile (
Profile (TP
TP)
)
Technical Overview
Technical Overview
RZhang_MPLS_TP 40
40

MPLS Terminology Overview
Label Switch Router (LSR)
Label Edge Router (LER)
Label Switched Path (LSP)
 LSP defines the path through LSRs from ingress to egress LER
 LSP defines the path through LSRs from ingress to egress LER
A collection of label pushes, swaps and Pops
Can be defined in many different ways : statically, dynamically through LDP, BGP, RSVP
RZhang_MPLS_TP 41

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
MPLS Label
0 3 5 6 8 9 0 3 5 6 8 9 0 3 5 6 8 9 0
Label EXP S TTL
Label = 20 bits
EXP = Experimental bits or traffic class (TC), 3 bits
S = Bottom of Stack, 1 bit
TTL = Time to Live, 8 bits
,
 It can be used over a variety of L2 encapsulations.
 Labels can be stacked
Labels can be stacked
RZhang_MPLS_TP 42

LSP Example
LSP Example
In
Lab
Address
Prefix
Out
I/F
Out
Lab
21 172.68.2.2/32 Lo0 Pop
In
Lab
Out
I/F
Out
Lab
42 2 21
In
Lab
Address
Prefix
Out
I/F
Out
Lab
– 172.68.2.2/32 0 42
1
3 0
2
P
PE PE
LSP
P
PE
Lo0=172.68.2.2/32
PE
Lo0=172.68.1.2/32
Payload Data
42
Payload Data Payload Data
21
RZhang_MPLS_TP 43

MPLS Pseudowire Terminology Overview
Provider Router (P)
Provider Edge (PE)
Tunnel LSP
Pseudowire
Provider Edge (PE)
Attachment
 Pseudowire used to provide a service over MPLS
Attachment
Circuit (AC)
 Pseudowire used to provide a service over MPLS
 Two levels of label stacking
 Tunnel LSP: identifying the path from PE to PE
RZhang_MPLS_TP 44
 Pseudowire: identifying the pseudowire services

Pseudowire Example
In
Lab
Address
Prefix
Out
I/F
Out
Lab
Pseudowire Example
In
Lab
Out
I/F
Out
Lab
In
Lab
Address
Prefix
Out
I/F
Out
Lab
1
3
21 172.68.2.2/32 Lo0 Pop
0
2
42 2 21 – 172.68.2.2/32 0 42
1
3 0
2
P
PE
Lo0=172.68.2.2/32
PE
Lo0=172.68.1.2/32
LSP
Attachment Circuit Attachment Circuit
Attachment Circuit ID label = 1
Payload Data
42
Payload Data 1
Payload Data
21 1 Payload Data
RZhang_MPLS_TP 45

MPLS TP Architecture
MPLS TP Architecture
NMS for Network
Management Control
NMS for Network
Management Control
Working LSP
Client node Client node
Provider Edge
P t t LSP
Provider Edge
Protect LSP
MPLS-TP LSP (Static or Dynamic)
Pseudowire with e2e and
segment OAM
Section Section
Client Signal
Connection Oriented, pre-determined working path and protect path
RZhang_MPLS_TP 46
, p g p p p
Transport Tunnel 1:1 protection, switching triggered by in-band OAM

The Three Planes for MPLS
The Three Planes for MPLS
 Control plane
Routing and Signaling: label distribution and LSP setup
Traffic Engineering: constrain based path computation, fast
reroute
 Forwarding plane
Also called data plane: push, pop, swap
Responsible for actual data packet forwarding
 Management plane
Configuration, provisioning, maintenance
RZhang_MPLS_TP 47

MPLS TP Planes
MPLS TP Planes
 Data plane is based on MPLS label forwarding
• Push: adding an outgoing label
• Push: adding an outgoing label
• Pop: remove an incoming label
• Swap: replace the incoming label with an outgoing label
 Data plane bandwidth must be enforced with QoS
 Control plane is not required, with GMPLS optional
p q , p
 Interoperates/interworks with existing MPLS and
pseudowire control and data planes
 Labeled switched path (LSP) may be setup via the
management plan
RZhang_MPLS_TP 48

Management Plane for MPLS TP
Management Plane for MPLS TP
 NMS plays a central role in a transport network
space
space
 DCN provides the critical management
infrastructure
 Circuit provisioning and maintenance
Create and manage a LSP or PW across a
network
LSP establishment
LSP maintenance
LSP maintenance
 Fault, PM reporting
RZhang_MPLS_TP 49

MPLS TP Control Plane
MPLS TP Control Plane
 A control plane is defined but not mandatory
 GMPLS is an optional control plane for MPLS that can
dynamically set up LSPs in a transport network
 An end to end control plane is also supported
 An end to end control plane is also supported
 Management and control planes may co-exist in the
same MPLS TP domain
same MPLS TP domain
RZhang_MPLS_TP 50

MPLS TP LSP Characteristics
MPLS TP LSP Characteristics
 LSP is always bidirectional
 An LSP is contained within a tunnel
 Tunnel can be protected or unprotected
 In-band OAM on each LSP
MPLS-TP
LSP
MPLS TP
MPLS-TP
T l
Protect
LSP
Working
LSP
RZhang_MPLS_TP 51
OAM
Channel
MPLS-TP
Tunnel
Tunnel
Protected
LSP
OAM
Channel
OAM
Channel
LSP

OAM
OAM
 OAM packets co-routed with data packets (in-band) to
d t t d t l f lt
detect data plane faults
 OAM available at LSP and PW levels
MPLS-TP
Tunnel
Protected
Protect
LSP
OAM
OAM
Working
LSP
RZhang_MPLS_TP 52
Protected
OAM
Channel
OAM
Channel

Tunnel End Point
Tunnel End Point
 Tunnel holds a working LSP and optionally a protect LSP
g p y p
Working
Protect (optional)
 Tunnel may be configured with a bandwidth allocation
 Tunnel operationally up if at least one LSP operationally
UP (and not locked out)
UP (and not locked out)
 LSP operationally up if OAM (Continuity Check) session
operationally up
p y p
RZhang_MPLS_TP 53

Tunnel Mid Point
Tunnel Mid-Point
 LSP defined using LSP ID
 Semantics of
source/destination only
locally significant
y g
 Configuration of forward
(from tunnel source) and
(f l
MPLS-TP
reverse (from tunnel
destination) LSP directions
 Configuration of label
MPLS TP
LSP
OAM
Channel
MPLS-TP
Tunnel
LSP I t O t t O t t  Configuration of label
swapping (input label,
output label and output
i t f )
LSP
Direction
Input
Label
Output
Label
Output
Interface
Forward 323111 334111 Gi2/1
Reverse 343111 111 Gi2/4
RZhang_MPLS_TP 54
interface)

OAM Channel
OAM Channel
 MPLS TP OAM channel is called MPLS Generic
A i t d Ch l GACh
Associated Channel, or GACh
 GACh is identified by its header
Th t f h l i id tifi d b Ch l T
 The type of channel is identified by Channel Type
Reserved
0 0 0 1 Version Channel Type
RZhang_MPLS_TP 55

GACh for MPLS TP LSP
GACh for MPLS TP LSP
 A well-known label is assigned for GACh (13)
 A GACh Label (GAL) acts as an exception mechanism
to identify OAM packets
G-ACH
GAL
Label
Associated Channel
Generic Associated
Channel Label (GAL)
Reserved
13 TC 1
1
OAM
Payload
Associated Channel
Header
(ACH)
Reserved
RZhang_MPLS_TP 56

G-ACh Packet Structure for an MPLS-
TP LSP
TP LSP
1
13 TC 1
S
Label TC TTL LSP Shim Header
Generic Associated Channel Label (GAL)
Reserved
Length
Reserved
Length
TLV Type
Associated Channel Header (ACH)
( )
ACH TLV Header
ACH TLV (e g Source destination LSP Id PW Id)
Value
ACH TLV (e.g Source, destination, LSP Id, PW Id)
G-ACH Message G-ACh Message
 GAL as bottom of label stack
 GAL only processed if LSP label popped or LSP TTL expires
 Same ACH structure
RZhang_MPLS_TP 57

OAM Functions
OAM Functions
Function Description
Continuity Check Checks ability to receive traffic
Connectivity Verification Verifies that a packet reaches expected node
Diagnostic Tests General diagnostic tests (e.g. looping traffic)
Route Tracing Discovery of intermediate and end points
Lock Instruct
Instruct remote MEPs to lock path (only test/OAM
traffic allowed)
Lock Reporting Report a server-layer lock to a client-layer MEP
Lock Reporting Report a server-layer lock to a client-layer MEP
Alarm Reporting Report a server-layer fault to a client-layer MEP
Remote Defect Indication Report fault to remote MEP
Client Failure Indication Client failure notification between MEPs
Client Failure Indication Client failure notification between MEPs
Packet Loss Measurement Ratio of packets not received to packets sent
Packet Delay Measurement
One-way / two-way delay (first bit sent to last bit
received)
RZhang_MPLS_TP 58

LSP 1:1 Protection
Bidirectional LSP
P1
Working LSP
(Up, Active)
Working LSP
(Up, Active)
PE1
P1
( p )
PE1
PE2
Protect LSP
(Up, Standby) Protect LSP
(Up, Standby)
P2 P3 Bidirectional LSP
( p y)
RZhang_MPLS_TP 59

LSP Protection Switching with Fault
P1
Working LSP
(Down, Standby)
Working LSP
(Down, Standby)
PE1
P1
PE1
PE2
Protect LSP
(Up, Active)
Protect LSP
(Up Active)
P2 P3
( p )
(Up, Active)
RZhang_MPLS_TP 60

LSP Reversion
P1
Working LSP
(Up, Active)
Working LSP
(Up, Active)
PE1
P1
( p )
WTR timer
expired
WTR timer
expired
PE1
PE2
Protect LSP
(Up, Standby) Protect LSP
(Up, Standby)
P2 P3
( p y)
RZhang_MPLS_TP 61

LSP Connectivity Check with BFD
LSP Connectivity Check with BFD
 Bidirectional Forwarding Detection (BFD) is used to
actively detect LSP connectivity
actively detect LSP connectivity
 BFD relies on regularly receipt of Hello messages
 A loss of a certain (usually 3) consecutive Hello
 A loss of a certain (usually 3) consecutive Hello
messages will trigger BFD down. For example, a 3.3
ms Hello interval will allow 10 ms fault detection
 An LSP only becomes active when BFD is configured
and it is in the up state
RZhang_MPLS_TP 62

MPLS TP BFD Encapsulation
MPLS TP BFD Encapsulation
Tunnel label GAL
GACh
4 bytes 4 bytes 0001 | Ver | Resv | Channel Type BFD header
 BFD packet label
 GAL: 13
 GACh header with channel type 0x7
RZhang_MPLS_TP 63

LSP Protection Switching with BFD
P1
Working LSP
(Down, Standby)
Working LSP
(Down, Standby)
BFD Control
Detection Time
E i d
PE1
P1
Expired
BFD Control
Detection Time
Expired
Switching time < 50 ms
PE1
PE2
Protect LSP
(Up, Active)
Protect LSP
(Up Active)
P2 P3
( p )
(Up, Active)
RZhang_MPLS_TP 64

LSP Fault Detection with LDI
LSP Fault Detection with LDI
 LSP has fault detection built in
 A fault detected on any point of the LSP will cause the
immediate nodes to generate LDI (Link Down
Indication) messages and LOS
) g
 LSP end points will process LDI messages and trigger
LSP down action
 LSP end points will then generate RDI messages
 LSP is taken down on both directions
RZhang_MPLS_TP 65

MPLS TP Fault OAM
MPLS TP Fault OAM
Tunnel label GAL
GACh
Tunnel label
4 bytes
GAL
4 bytes 0001 | Ver | Resv | Channel Type Fault OAM header
 Fault OAM message types:
 AIS Alarm Indication Signal
 LDI Link Down Indication
 LKR Lockout
 Fault OAM packet label
 GAL: 13
RZhang_MPLS_TP 66
 GACh header with channel type 0x58

LSP Protection Switching with Fault
O
OAM
LDI
P1
Working LSP
(Down, Standby)
Working LSP
(Down, Standby)
LDI
LOS
PE1
P1
PE1
PE2
Protect LSP
(Up, Active)
Protect LSP
(Up Active)
P2 P3
( p )
(Up, Active)
RZhang_MPLS_TP 67

LSP Lockout
LSP Lockout
 An LSP can be administratively locked out
 A locked out LSP does not carry traffic
RZhang_MPLS_TP 68

LSP Protection Switching with Lockout
LKR
P1
Working LSP
(Up, Standby)
Working LSP
(Up, Standby)
LKR
LKR
PE1
P1
PE1
PE2
Protect LSP
(Up, Active)
Protect LSP
(Up Active)
P2 P3
( p )
(Up, Active)
RZhang_MPLS_TP 69

Mapping of Customer Traffic
Mapping of Customer Traffic
 Customer traffic connected via an Attachment Circuit
(AC)
(AC)
 An AC cross connected to an MPLS virtual circuit (VC)
or pseudowire
p
 A VC can be point to point or multipoint
RZhang_MPLS_TP 70

Pseudowire Reference Model
Pseudowire Reference Model
Emulated Service
Pseudowire
PSN
Tunnel
Native
Service
Native
Service
CE CE
PW2
PW1
AC
AC
An Attachment Circ it (AC) is the ph sical or irt al circ it
PE
PE
CE
CE
CE
CE
AC
AC
 An Attachment Circuit (AC) is the physical or virtual circuit
attaching a CE to a PE
 Customer Edge (CE) equipment perceives a PW as an
RZhang_MPLS_TP 71
unshared link or circuit

Virtual Private Wire Service (VPWS)
Virtual Private Wire Service (VPWS)
 A point to point circuit that emulates a line
 If Attachment Circuit (AC) is a physical port, Ethernet
Private Line
 If AC is sharing the port with other ACs Ethernet Virtual
 If AC is sharing the port with other ACs, Ethernet Virtual
Private Line
RZhang_MPLS_TP 72

Pseudowire Redundancy
Pseudowire Redundancy
 Second layer of redundancy in addition to MPLS-TP
LSP 1:1 Protection
LSP 1:1 Protection
 Protected pseudowires are in Active/Standby states
 Standby pseudowire is down pseudowire label is
 Standby pseudowire is down, pseudowire label is
released
RZhang_MPLS_TP 73

MPLS-TP Pseudowire Redundancy
MPLS TP Pseudowire Redundancy
LSP
Source
LSP
Destination
Working LSP 1
AC
AC
Active Pseudowire
TP Tunnel 1 TP Tunnel 1
Protect LSP 1
Working LSP 2
MPLS
Encap
MPLS
Encap
Protect LSP 2
T-PE T-PE
Standby Pseudowire
TP Tunnel 2 TP Tunnel 2
Pseudowire Protection
LSP 1:1 Protection
RZhang_MPLS_TP 74

Virtual Private LAN Service
Virtual Private LAN Service
 A multipoint circuit that emulates a LAN
 If AC is a physical port, Ethernet Private LAN
 If AC is sharing the port with other ACs, Ethernet Virtual
Private LAN
Private LAN
RZhang_MPLS_TP 75

VPLS Redundancy
VPLS Redundancy
 All PEs of the same private LAN are fully meshed
 Split horizon is enabled
 A protected MPLS TP LSP makes fiber fault
transparent to VPLS
transparent to VPLS
RZhang_MPLS_TP 76

Bandwidth Management
Bandwidth Management
 MPLS-TP LSPs can reserve bandwidth (for tunnel provisioning)
 LSP bandwidth reservation configured explicitly at each hop
 MPLS-TP LSPs have highest setup/hold priorities
 Data plane bandwidth enforcement requires QoS configuration
MPLS-TP LSP
Data link
RZhang_MPLS_TP 77

Data Plane QoS
Data Plane QoS
 Traffic type classification based on CoS, IP Prec/DSCP, VLAN etc
 End-to-end bandwidth provisioning and guarantee
 Low latency queuing for delay or jitter-sensitive traffic
 Prioritizing processing of control or management-plane traffic over
data-plane traffic
RZhang_MPLS_TP 78

MPLS
MPLS TP
TP
Deployment
Deployment
Scenarios
Scenarios
RZhang_MPLS_TP 79
79

Common Deployment Scenarios
Common Deployment Scenarios
 Migration of SONET/SDH to MPLS-TP
 Consolidation into a single transport network
 Greenfield deployment that requires SONET like
protection
protection
 Multipoint LAN services over transport
D l t E l
 Deployment Examples:
Metro aggregation/access
Mobile back-haul
Mobile back haul
RZhang_MPLS_TP 80

Common Deployment Practices
Common Deployment Practices
 LSPs are provisioned by NMS without a control plane
 BFD processed in hardware for 10 ms fault detection
 VPWS for point to point EPL or EVPL services
 Dual home pseudowires for site protection
 VPLS for multipoint services such as multicast video
di t ib ti
distribution
 Use of QoS for preferential services and
oversubscription
oversubscription
RZhang_MPLS_TP 81

Consolidation and Simplification
Consolidation and Simplification
 Currently multiple networks for TDM and packet
 Consolidate into a single transport network
 SONET like timing provided via Synchronous Ethernet
T1/E1, Ethernet, STM1
T1/E1 Ethernet OCn/STMn
Business
Packet Transport
with MPLS TP
T1/E1, Ethernet, OCn/STMn
GE
Corporate
Synchronous Ethernet timing
RZhang_MPLS_TP 82

FTTx Deployment
FTTx Deployment
 Aggregation of Ethernet services
 50 ms protection for mission critical services
 QoS for preferential delivery treatment
 Use of satellite boxes to increase density and reach
FTTB
CPE
Set-Top-Box
FTTB
FTTH
Packet Transport
with MPLS TP
Business
Tenant
CPE
Metro Ethernet
Protected or
unprotected TP
t l
RZhang_MPLS_TP 83
tunnels

FTTx Deployment Alternate
FTTx Deployment Alternate
 Use a ring of satellite boxes to reduce fiber usage
FTTB
CPE
Set-Top-Box
FTTH
Packet Transport
with MPLS TP
Business
Tenant
CPE
Metro Ethernet
Protected or
unprotected TP
t l
RZhang_MPLS_TP 84
tunnels

Mobile Backhaul Deployment
Mobile Backhaul Deployment
 Migration of TDM to packet transport
 50 ms protection
 SONET like timing provided via Synchronous Ethernet
GSM/GPRS/
EDGE BTS
GSM
BTS
UMTS
Node B
TDM
N d R
Packet Transport
with MPLS TP
Node Ra
Ethernet
Synchronous Ethernet timing
RZhang_MPLS_TP 85

Central Office Fiber Management
Central Office Fiber Management
 Use of satellite boxes to reduce fiber management at
CO
CO
PON FTTH Ethernet
PON
Access
FTTH
Access
Ethernet
Access
CO
Passive
Splitter
CO CO
Pro: Fiber Consolidation
Con: BW Constrained
Home-Run from
CO to each user
Pro: Bandwidth Scale
Con: Fiber Mgt. space
Home-Run from
CO to each user
Pro: Fiber Consolidation
Pro : Bandwidth Scale
RZhang_MPLS_TP 86
g p
overhead in CO

Multipoint LAN Services
Multipoint LAN Services
 Virtual LAN services over MPLS TP transport
Enterprise LAN
 Multicast video distribution services
Enterprise LAN
POP
R id ti l
p
Packet Transport
over DWDM
Residential
POP
Residential
RZhang_MPLS_TP 87
Residential

MPLS
MPLS-
-TP
TP
Deployment
Deployment
Lifecycle
Lifecycle
y
y
RZhang_MPLS_TP 88
88

From TDM Transport to Packet
Transport
Transport
 Know the differences
 Understand the new requirements
 Proof of concept testing
 Create designs
 Performance testing
 Turn-up and provisioning
Follow a lifecycle process to ensure deployment
success and timely delivery
RZhang_MPLS_TP 89

TDM Transport vs Packet Transport
TDM Transport vs Packet Transport
 Time division multiplexing vs statistical multiplexing
 New terminologies and technologies: LSP, pseudowire,
BFD, VPWS, VPLS, QoS, policing, queuing
 Provisioned bandwidth vs data plane QoS
 Provisioned bandwidth vs data plane QoS
 Staff training
RZhang_MPLS_TP 90

Creating Technical Requirements
Creating Technical Requirements
 Convert business requirements into technical
requirements
requirements
 Identify QoS requirements for circuit emulating traffic
 Generate topologies
 Generate topologies
 Document traffic flows
P i iti i t
 Prioritize requirements
RZhang_MPLS_TP 91

Proof of Concept Testing
Proof of Concept Testing
 Convert technical requirements into a basic design
 Convert topologies into a test lab
 Validate the concept
 Focus on general functionality and mandatory
requirements
RZhang_MPLS_TP 92

Design
Design
 Generate a high level design based on proof of concept
testing
testing
 Understand traffic flow patterns
 Identify MPLS TP parameters
 Identify MPLS TP parameters
 Identify MPLS virtual circuit characteristics
D t t k t
 Document network management
 Specify scalability limits
RZhang_MPLS_TP 93

QoS
QoS
 Identify circuits that require SONET-like protection
 The network can support both protected and
unprotected circuits
 Design QoS policies to support all types of circuits
 Design QoS policies to support all types of circuits
 Identify circuits that require dual-homing
RZhang_MPLS_TP 94

QoS Design Examples
QoS Design Examples
 Traffic are classified into TDM type circuits and packet
type circuits
type circuits
 For TDM type circuits:
No oversubscription
Priority queue (CoS 6)
Timing may be required
 For packet type circuits:
p yp
Oversubscription allowed
Weighted fair queues
Guaranteed Bandwidth for different queues
q
High (CoS 5)
Medium (CoS 3)
Low (Cos 0)
RZhang_MPLS_TP 95

Performance Testing
Performance Testing
 A detailed verification of each type of traffic in the
design
design
 Focus on protection switching and QoS
 Document test case results and solutions
 Document test case results and solutions
 Update the design based on test results
RZhang_MPLS_TP 96

Turn-up and Provisioning
Turn up and Provisioning
 Equipment install and turn-up
 An intermediate staging may be useful
 Operational staff training
 Final hardware testing
 Provision the equipment based on the design
The network is ready for use
RZhang_MPLS_TP 97

Summary
Summary
RZhang_MPLS_TP 98
98

Transport Network in a Transition
Transport Network in a Transition
 Convergence of multiple networks into a single transport
network
network
 Reducing CAPEX and OPEX
RZhang_MPLS_TP 99

A Converged Transport
A Converged Transport
 OTN provides the digital wrapper
MPLS T t P fil (TP) id SONET lik
 MPLS Transport Profile (TP) provides SONET like
services
OPEX
Traditional TDM services and packet based services
p
carried over a single transport network
RZhang_MPLS_TP 100

Why MPLS Transport Profile?
Why MPLS Transport Profile?
 Transport like protection
 Transport like OAM
 Transport like operation
p p
 Statistical multiplexing and oversubscription
 Interoperability with IP/MPLS
Interoperability with IP/MPLS
RZhang_MPLS_TP 101

Follow the Deployment Process
Follow the Deployment Process
 Know the differences between TDM and packet
 Understand new requirements
 Proof of concept testing
 Create designs
 Performance testing
 Turn-up and provisioning
RZhang_MPLS_TP 102

Deploy Packet Transport with MPLS-TP - nanog.pdf

Deploy Packet Transport with MPLS-TP - nanog.pdf

More Related Content

What's hot (20)

Similar to Deploy Packet Transport with MPLS-TP - nanog.pdf (20)

Recently uploaded (20)

Deploy Packet Transport with MPLS-TP - nanog.pdf