p1: ote/ote/sph p2: ote - download.e- · pdf filep1: ote/ote/sph p2: ote ......

P1: OTE/OTE/SPH P2: OTEfm JWST021-Minei October 16, 2010 13:34 Printer Name: Yet to Come


‘While MPLS is in itself simple, its apparent complexity lies in the proliferation of applications,which shows no signs of ceasing. To make things worse, catching up involves reading a largenumber of documents written by various authors at various times in various styles. Hereat last is a single, all encompassing resource where the myriad applications sharpen into acomprehensible text that first explains the whys and whats of each application before goingon to the technical detail of the hows.’

Kireeti Kompella, CTO Junos, Juniper Networks

‘MPLS-Enabled Applications thoroughly covers the MPLS base technology and applicationson MPLS-enabled IP networks. It guides you to a comprehensive understanding of stan-dards, problems, and solutions in networking with MPLS. Before it had been necessary togo through material from many different sources, here we have everything in one place. Allthe MPLS protocols are covered, as are the applications of these protocols. This should be thetextbook for MPLS courses, both for training of experienced networking professionals andfor universities.’

Loa Andersson, Ericsson AB and IETF MPLS working group co-chair

‘Although over ten years old, MPLS technology continues to evolve to meet the developingrequirements of network operators and the advancing aspirations of network users. It isimportant that a book like this should continue to be updated in step with the changes toMPLS, and this new revision includes essential new material for those trying to understandthe next steps in MPLS.’

Adrian Farrel, IETF Routing Area Director

‘This book continues to be the industry and academic state-of-the-art on explaining thefoundation and nuances of MPLS technology. It is extremely well written and tackles all ofthe most modern extensions of MPLS technology. If you are interested in how the internetworks, it will be a well-worn read. It should be on every internet practitioner’s bookshelf.’

Dave Ward, IETF WG chair: BFD, Softwires, ISIS, HIP

‘This is the MPLS text that the industry has been waiting for. On one hand, the text presentsMPLS technology clearly enough that the reader can absorb its content in a few easy sittings.On the other hand, the text provides a sufficiently in-depth treatment that even an MPLSexpert can learn from it. The authors offer a clear and complete description of MPLS, its innerworkings and its applications, in a manner that could only be achieved by persons who havebeen significant contributors to the MPLS development effort. Every network operator whohas deployed or is considering the deployment of MPLS technology should read this book.It is appropriate reading for everyone from the CTO to the tier 1 NOC engineer.’

Ron Bonica, Juniper Networks, Co-director IETF Operations and Management Area

‘MPLS-Enabled Applications provides excellent insight on how recently developed solutionscan help address challenges for providing multicast in MPLS-based VPNs. The in-depthcoverage of recent advances in MPLS technology that provide multicast support in L2 and L3VPNs is essential to anyone needing to deploy both basic use cases and advanced scenariosas well.’

Thomas Morin, Network Architect at France Telecom Orange

‘This is a highly recommended book for network design engineers who want to updatethemselves with the latest MPLS development, or those who want to learn this technologythoroughly. In addition to the impressive technology coverage and depth, the book is also adelightful reading!’

Lei Wang, Department manager Mobile IP Transport, Telenor


‘MPLS-Enabled Applications is an excellent read for network engineers involved in thedesign of MPLS networks and services. It can serve as an introduction to MPLS networkingor as a reference book for the advanced engineer. It discusses practical issues that must beconsidered in the design of MPLS networks and services, including MPLS-TE, MPLS-IPVPNsand MPLS L2VPNs. It also discusses current topics that are still evolving in the industry suchas inter-AS/area MPLS-TE, point-to-multipoint LSPs and IPVPN multicast, providing a goodoverview of the issues being addressed and the current industry direction.’

Nabil N. Bitar, Principal member of Technical Staff and lead network architect, Verizon

‘MPLS-Enabled Applications: Emerging Developments and New Technologies second edi-tion, by Ina Minei and Julian Lucek, presents the current state-of-the-art in the specification,development, and application of MPLS and its related technologies. I believe, the readerswill find the book to be a very valuable resource. I am pleased to see that the third edition ofthis book covers contemporary topics in the Internet industry such as MPLS-TP.’

Bijan Jabbari, PhD, Founder of Isocore, and Professor of Electrical Engineering, George MasonUniversity

‘This is the MPLS book that I reference the most and recommend to all my colleagues. Itis written in an easy-to-follow approach that starts with basic concepts and then graduallyramps to advanced topics. It is timely in its coverage of new developments such as MPLS-TPand BGP/MPLS mVPNs, yet exhaustive by addressing all aspects of MPLS including thenewer advances. I have personally used this book to architect designs such as broadcastvideo over IP/MPLS, hierarchical video-on-demand library distribution using BGP/MPLSmVPN, and a MPLS-based network supporting triple-play services over a BGP and PIM-freeCore.’

Mazen Khaddam, Principal lead network architect, network architecture group, Cox communications

‘This book is a wonderfully comprehensive overview of not just the underlying technol-ogy, but also the many use case applications of MPLS. It’s a must have for networkingprofessionals.’

Dorian Kim, Director of Network Development, NTT America

‘MPLS-Enabled Applications takes a unique and creative approach in explaining MPLS con-cepts and how they are applied in practice to meet the needs of Enterprise and ServiceProvider networks. I consistently recommend this book to colleagues in the engineering,education and business community.’

Dave Cooper, Chief IP Technologist, Global Crossing Ltd.

‘This book presents clear, comprehensive descriptions of the various scenarios in which theMPLS toolkit can be used to provide reliable and quality connectivity. It includes backgroundinformation, detailed explanations on how to enable different services and applications, andprecise technical and operational considerations. Business drivers for emerging technologiesare discussed as well as practical and real deployment scenarios. Highlighting the hottesttrends in the industry, this invaluable book describes how best to fit the pieces of the puzzletogether to efficiently enable new applications and services.’

Nurit Sprecher, Senior specialist, Packet Transport Evolution, Nokia Siemens Networks


MPLS-Enabled Applications


WILEY SERIES IN COMMUNICATIONS NETWORKING& DISTRIBUTED SYSTEMS

Series Editors: David Hutchison, Lancaster University, Lancaster, UKSerge Fdida, Universite Pierre et Marie Curie, Paris, FranceJoe Sventek, University of Glasgow, Glasgow, UK

The ‘Wiley Series in Communications Networking & Distributed Systems’ is aseries of expert-level, technically detailed books covering cutting-edge research,and brand new developments as well as tutorial-style treatments in network-ing, middleware and software technologies for communications and distributedsystems. The books will provide timely and reliable information about the state-of-the-art to researchers, advanced students and development engineers in theTelecommunications and the Computing sectors.

Other titles in the series:

Wright: Voice over Packet Networks 0-471-49516-6 (February 2001)Jepsen: Java for Telecommunications 0-471-49826-2 (July 2001)Sutton: Secure Communications 0-471-49904-8 (December 2001)Stajano: Security for Ubiquitous Computing 0-470-84493-0 (February 2002)Martin-Flatin: Web-Based Management of IP Networks and Systems 0-471-48702-3

(September 2002)Berman, Fox, Hey: Grid Computing. Making the Global Infrastructure a Reality 0-470-

85319-0 (March 2003)Turner, Magill, Marples: Service Provision. Technologies for Next Generation Commu-

nications 0-470-85066-3 (April 2004)Welzl: Network Congestion Control: Managing Internet Traffic 0-470-02528-X (July

2005)Raz, Juhola, Serrat-Fernandez, Galis: Fast and Efficient Context-Aware Services 0-470-

01668-X (April 2006)Heckmann: The Competitive Internet Service Provider 0-470-01293-5 (April 2006)Dressler: Self-Organization in Sensor and Actor Networks 0-470-02820-3 (November

2007)Berndt: Towards 4G Technologies: Services with Initiative 0-470-01031-2 (March 2008)Jacquenet, Bourdon, Boucadair: Service Automation and Dynamic Provisioning Tech-

niques in IP/MPLS Environments 0-470-01829-1 (March 2008)Gurtov: Host Identity Protocol (HIP): Towards the Secure Mobile Internet 0-470-99790-7

(June 2008)Boucadair: Inter-Asterisk Exchange (IAX): Deployment Scenarios in SIP-enabled Net-

works 0-470-77072-4 (January 2009)Fitzek: Mobile Peer to Peer (P2P): A Tutorial Guide 0-470-69992-2 (June 2009)Shelby: 6LoWPAN: The Wireless Embedded Internet 0-470-74799-4 (November 2009)Stavdas: Core and Metro Networks 0-470-51274-1 (February 2010)Gomez Herrero, van der Ven, Network Mergers and Migrations: Junos R© Design and

Implementation 0-470-74237-2 (March 2010)Jacobsson, Niemegeers, Heemstra de Groot, Personal Networks: Wireless Networking

for Personal Devices 0-470-68173-X (June 2010)


MPLS-Enabled ApplicationsEmerging Developmentsand New Technologies

Third Edition

Ina MineiJuniper Networks

Julian LucekJuniper Networks

John Wiley & Sons, Ltd


This edition first published 2011C© 2011 John Wiley & Sons, Ltd.

Registered officeJohn Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ,United Kingdom

For details of our global editorial offices, for customer services and for information abouthow to apply for permission to reuse the copyright material in this book please see ourwebsite at www.wiley.com.

The right of the author to be identified as the author of this work has been asserted inaccordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrievalsystem, or transmitted, in any form or by any means, electronic, mechanical, photocopying,recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appearsin print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed astrademarks. All brand names and product names used in this book are trade names, servicemarks, trademarks or registered trademarks of their respective owners. The publisher isnot associated with any product or vendor mentioned in this book. This publication isdesigned to provide accurate and authoritative information in regard to the subject mattercovered. It is sold on the understanding that the publisher is not engaged in renderingprofessional services. If professional advice or other expert assistance is required, theservices of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

Minei, Ina.MPLS-enabled applications : emerging developments and new

technologies / Ina Minei, Julian Lucek.p. cm.

Includes bibliographical references and index.ISBN 978-0-470-66545-9 (pbk.)

1. MPLS standard. 2. Extranets (Computer networks) I. Lucek, Julian.II. Title.

TK5105.573.M56 2010621.382′16–dc22 2010029550

A catalogue record for this book is available from the British Library.

ISBN: 9780470665459 (P/B)ePDF ISBN: 9780470976166oBook ISBN: 9780470976173ePub ISBN: 9780470976135

Set in 10/12pt Palatino by Aptara Inc., New Delhi, India.

http://www.wiley.com


Contents

About the Authors xix

Foreword xxi

Preface xxv

Acknowledgements xxxi

Part One

1 Foundations 31.1 Historical perspective 31.2 Current trends 51.3 MPLS mechanisms 6

1.3.1 Forwarding plane mechanisms 71.3.2 Control plane mechanisms 111.3.3 Transport of IPv6 over an IPv4 MPLS core 32

1.4 Conclusion 351.5 References 351.6 Further reading 361.7 Study questions 36

2 Traffic Engineering with MPLS (MPLS-TE) 392.1 Introduction 392.2 The business drivers 392.3 Application scenarios 402.4 Setting up traffic-engineered paths using MPLS-TE 43

2.4.1 LSP priorities and preemption 432.4.2 Information distribution – IGP extensions 442.4.3 Path calculation – CSPF 462.4.4 Path setup – RSVP extensions and admission control 49


viii CONTENTS

2.5 Using the traffic-engineered paths 512.6 Deployment considerations 54

2.6.1 Scalability 542.6.2 Reservation granularity 562.6.3 Routing challenges 57

2.7 Using traffic engineering to achieve resourceoptimization 572.7.1 Autobandwidth – dealing with unknown

bandwidth requirements 582.7.2 Sharing links between RSVP and other traffic –

dealing with unknown bandwidth availability 592.7.3 Other methods for optimization of

transmission resources in MPLS networks 602.8 Offline path computation 612.9 Conclusion 642.10 References 652.11 Further reading 652.12 Study questions 65

3 Protection and Restoration in MPLS Networks 673.1 Introduction 673.2 The business drivers 683.3 Failure detection 693.4 End-to-end protection 70

3.4.1 Control over the traffic flow following a failure 713.4.2 Requirement for path diversity 713.4.3 Double-booking of resources 723.4.4 Unnecessary protection 723.4.5 Nondeterministic switchover delay 72

3.5 Local protection using fast reroute 733.5.1 Case (i): link protection, for the facility

protection case 753.5.2 Case (ii): link protection, for the 1:1 protection case 773.5.3 Case (iii): node protection, for the facility

protection case 783.5.4 Case (iv): node protection, for the 1:1 protection case 79

3.6 Link protection 813.6.1 What happens before the failure 823.6.2 What happens after the failure 87

3.7 Node protection 893.8 Additional constraints for the computation of the

protection path 913.8.1 Fate sharing 913.8.2 Bandwidth protection 933.8.3 Bandwidth protection and DiffServ 96


CONTENTS ix

3.9 Interaction of end-to-end protection and fast reroute 973.10 Deployment considerations for local protection

mechanisms 983.10.1 Scalability considerations 983.10.2 Evaluating a local protection implementation 1013.10.3 The cost of bandwidth protection 103

3.11 IP and LDP FRR 1053.11.1 The tunnel-based approach 1073.11.2 The alternate-path approach 108


4 MPLS DiffServ-TE 1134.1 Introduction 1134.2 The business drivers 1144.3 Application scenarios 115

4.3.1 Limiting the proportion of traffic from aparticular class on a link 115

4.3.2 Maintaining relative proportions of traffic on links 1174.3.3 Providing guaranteed bandwidth services 117

4.4 The DiffServ-TE solution 1174.4.1 Class types 1174.4.2 Path computation 1184.4.3 Path signaling 1214.4.4 Bandwidth constraint models 1224.4.5 Overbooking 1274.4.6 The DiffServ in DiffServ-TE 1294.4.7 Protection 1304.4.8 Tools for keeping traffic within its reservation limits 1314.4.9 Deploying the DiffServ-TE solution 132

4.5 Extending the DiffServ-TE solution withmulticlass LSPs 133


5 Interdomain Traffic Engineering 1375.1 Introduction 1375.2 The business drivers 1375.3 Setting up interdomain TE LSPs 139

5.3.1 Path setup 1405.3.2 Path computation 144


x CONTENTS

5.3.3 Reoptimization 1545.3.4 Protection and fast reroute 155

5.4 Interprovider challenges 1575.5 Comparison of the LSP setup methods 1585.6 Conclusion 1595.7 References 1605.8 Further reading 1615.9 Study questions 161

6 MPLS Multicast 1636.1 Introduction 1636.2 The business drivers 1646.3 P2MP LSP mechanisms 165

6.3.1 Forwarding plane mechanisms 1656.3.2 Control plane mechanisms 167

6.4 LAN procedures for P2MP LSPs 1766.4.1 Upstream label allocation 177

6.5 Coupling traffic into a P2MP LSP 1786.5.1 Coupling Layer 2 traffic into a P2MP LSP 1796.5.2 Coupling IP unicast traffic into a P2MP LSP 1796.5.3 Coupling IP multicast traffic into a P2MP LSP 180

6.6 MPLS fast reroute 1816.7 Ingress redundancy for P2MP LSPs 1836.8 P2MP LSP hierarchy 184

6.8.1 P2MP LSP hierarchy forwarding plane operation 1866.8.2 P2MP LSP hierarchy control plane operation 187

6.9 Applications of point-to-multipoint LSPs 1876.9.1 Application of P2MP TE to broadcast TV distribution 1886.9.2 Application of P2MP LSPs to L3VPN multicast 1916.9.3 Application of P2MP LSPs to VPLS 193

6.10 Conclusion 1936.11 References 1936.12 Study questions 195

Part Two

7 Foundations of Layer 3 BGP/MPLS Virtual Private Networks 1997.1 Introduction 1997.2 The business drivers 2007.3 The overlay VPN model 2017.4 The peer VPN model 2027.5 Building the BGP/MPLS VPN solution 205

7.5.1 VPN routing and forwarding tables (VRFs) 2057.5.2 Constrained route distribution 207


CONTENTS xi

7.5.3 VPN-IPv4 addresses and the routedistinguisher (RD) 208

7.5.4 The route target (RT) 2097.5.5 The solution so far – what is missing? 2157.5.6 VPN label 216

7.6 Benefits of the BGP/MPLS VPN solution 2217.7 References 2227.8 Further reading 2227.9 Study questions 223

8 Advanced Topics in Layer 3 BGP/MPLS VirtualPrivate Networks 2258.1 Introduction 2258.2 Routing between CE and PE 2258.3 Differentiated VPN treatment in the core 2308.4 Route reflectors and VPNs 2318.5 Scalability discussion 235

8.5.1 Potential scaling bottlenecks 2368.5.2 The cost of growing the VPN network 238

8.6 Convergence times in a VPN network 2438.6.1 Convergence time for a customer route change 2438.6.2 Convergence time for a failure in the

provider’s network 2448.7 Security issues 244

8.7.1 Can traffic from one VPN ‘cross over’ intoanother VPN? 245

8.7.2 Can a security attack on one VPN affectanother VPN? 245

8.7.3 Can a security attack against the serviceprovider’s infrastructure affect theVPN service? 246

8.8 QoS in a VPN scenario 2468.9 IPv6 VPNs 2488.10 Conclusion 2518.11 References 2518.12 Further reading 2528.13 Study questions 252

9 Hierarchical and Inter-AS VPNs 2559.1 Introduction 2559.2 Carriers’ carrier – service providers as VPN customers 256

9.2.1 ISP as a VPN customer 2579.2.2 VPN service provider as a VPN customer –

hierarchical VPN 262


xii CONTENTS

9.3 Multi-AS backbones 2669.3.1 Option A: VRF-to-VRF connections at the ASBR 2669.3.2 Option B: EBGP redistribution of labeled

VPN-IPv4 routes 2689.3.3 Option C: multihop EBGP redistribution of

labeled VPN-IPv4 routes between the sourceand destination AS, with EBGP redistributionof labeled IPv4 routes from one AS to theneighboring AS 269

9.4 Interprovider QoS 2719.5 Conclusion 2729.6 References 2729.7 Further reading 2739.8 Study questions 273

10 Multicast in a Layer 3 VPN 27510.1 Introduction 27510.2 The business drivers 27610.3 mVPN – problem decomposition 27810.4 The original multicast solution− PIM/GRE mVPN

(draft-rosen) 27910.4.1 PIM/GRE mVPN – routing information

distribution using PIM C-instances 28010.4.2 PIM/GRE mVPN – carrying multicast traffic

across the core using multicast distribution trees 28110.4.3 Properties of the PIM/GRE mVPN solution 283

10.5 NG multicast for L3VPN – BGP/MPLS mVPN(NG mVPN) 286

10.5.1 Requirements for support of PIM-SM SSM inan mVPN 286

10.5.2 BGP/MPLS mVPN – carrying multicastmVPN routing information using C-multicast routes 287

10.5.3 BGP/MPLS mVPN – carrying traffic acrossthe provider network using inter-PE MPLS tunnels 292

10.5.4 BGP/MPLS mVPN – inter-PE tunnels –inclusive and selective tunnels 292

10.5.5 BGP/MPLS mVPN – carrying traffic fromseveral mVPNs onto the same inter-PE tunnel 294

10.5.6 BGP/MPLS mVPN – creating inter-PEtunnels using BGP autodiscovery routes 295

10.5.7 Requirements for support of PIM ASM inan mVPN 299


CONTENTS xiii

10.5.8 BGP/MPLS mVPN – carrying mVPN activesource information using BGP source activeautodiscovery routes 300

10.6 Comparison of PIM/GRE and BGP/MPLS mVPNs 30310.6.1 VPN model used 30310.6.2 Protocol used in the control plane 30410.6.3 Data-plane mechanisms 30510.6.4 Service provider network as a ‘LAN’ 30610.6.5 Deployment considerations 306


11 Advanced Topics in BGP/MPLS mVPNs 31111.1 Introduction 31111.2 BGP/MPLS mVPN – inter-AS operations 31111.3 Support of PIM DM in BGP/MPLS mVPN 31611.4 Discovering the RP – auto-RP and BSR support in

BGP/MPLS mVPN 31711.5 Implementing extranets in BGP/MPLS mVPN 31911.6 Transition from draft-rosen to BGP/MPLS mVPNs 32211.7 Scalability discussion 325

11.7.1 PIM/GRE mVPN control plane scaling 32511.7.2 BGP/MPLS mVPN control plane scaling 326

11.8 Achieving multicast high availability withBGP/MPLS mVPN 328

11.8.1 Live-Standby multicast delivery usingBGP/MPLS mVPN 329

11.8.2 Live-Live multicast delivery usingBGP/MPLS mVPN 332

11.8.3 Comparison of the Live-Live andLive-Standby multicast high-availability schemes 335

11.9 Internet multicast service using the BGP/MPLSmVPN technology 335


12 Layer 2 Transport over MPLS 34112.1 Introduction 34112.2 The business drivers 34112.3 Comparison of layer 2 VPNs and layer 3 VPNs 344


xiv CONTENTS

12.4 Principles of layer 2 transport over MPLS 34512.5 Forwarding plane 347

12.5.1 ATM cell 34912.5.2 ATM AAL5 34912.5.3 Frame relay 35012.5.4 Ethernet 350

12.6 Control plane operation 35112.6.1 Original LDP signaling scheme 35112.6.2 BGP-based signaling and autodiscovery scheme 35312.6.3 LDP signaling with BGP autodiscovery 35712.6.4 Comparison of BGP and LDP approaches to

Layer 2 transport over MPLS 35812.7 Admission control of layer 2 connections into

network 36012.8 Failure notification mechanisms 36112.9 Multi-homing 362

12.9.1 BGP case 36212.9.2 LDP case 364

12.10 Layer 2 interworking 36512.11 Circuit cross connect (CCC) 36512.12 Point-to-multipoint Layer 2 transport 366

12.12.1 Point-to-multipoint CCC 36712.12.2 Layer 2 Multicast VPNs 367

12.13 Other applications of Layer 2 transport 36812.14 Conclusion 37012.15 References 37012.16 Study questions 371

13 Virtual Private LAN Service 37313.1 Introduction 37313.2 The business drivers 37313.3 VPLS mechanism overview 37513.4 Forwarding plane mechanisms 379

13.4.1 Forwarding of unicast frames 37913.4.2 Broadcast and multicast frames 382

13.5 Control plane mechanisms 38413.5.1 LDP-based signaling 38413.5.2 BGP signaling and autodiscovery 38913.5.3 Comparison of LDP and BGP for VPLS

control plane implementation 39613.5.4 IGMP and PIM snooping 39913.5.5 Use of multicast trees in VPLS 401

13.6 LDP and BGP interworking for VPLS 40613.7 Interprovider Option E for VPLS 413

13.7.1 Comparison of interprovider schemes for VPLS 415


CONTENTS xv

13.8 Operational considerations for VPLS 41613.8.1 Number of MAC addresses per customer 41613.8.2 Limiting broadcast and multicast traffic 41713.8.3 Policing of VPLS traffic 41713.8.4 VPLS with Integrated Routing and Bridging

(IRB) 41713.8.5 Learning mode 417


Part Three

14 Advanced Protection and Restoration: Protectingthe Service 42314.1 Introduction 42314.2 The business drivers 42314.3 Failure scenarios 42514.4 Existing solutions 426

14.4.1 Single homed CE 42614.4.2 Dual-homed CE 42714.4.3 Analyzing existing dual-homing solutions 432

14.5 Protecting the egress – local protection solution 43314.5.1 Protecting against an attachment circuit

failure in a pseudowire scenario – edgeprotection virtual circuit 435

14.5.2 Protecting against an egress PE failure in anL3VPN scenario 437


15 MPLS Management 44315.1 Introduction 44315.2 Management – why and what 44315.3 Detecting and troubleshooting failures 445

15.3.1 Reporting and handling nonsilent failures 44515.3.2 Detecting silent failures – MPLS OAM 44615.3.3 Troubleshooting failures 461

15.4 Configuration errors 46715.4.1 Preventing configuration errors 46715.4.2 Detecting and reporting misconfigurations 469

15.5 Visibility 473


xvi CONTENTS


16 MPLS in Access Networks and Seamless MPLS 47916.1 Introduction 47916.2 The business drivers 479

16.2.1 The transition from legacy access to Ethernet access 48016.2.2 MPLS as the technology choice for the

Ethernet access network 48316.3 Models for MPLS deployment in access networks 48616.4 Seamless MPLS Mechanisms 491

16.4.1 Extending MPLS to the Access Node 49116.4.2 Seamless MPLS scaling 49316.4.3 Scaling analysis of Seamless MPLS 49716.4.4 Seamless MPLS for multicast 501

16.5 Conclusions 50716.6 References 50716.7 Study questions 508

17 MPLS Transport Profile (MPLS-TP) 50917.1 Introduction 50917.2 The business drivers 50917.3 Requirements for a transport profile for MPLS 512

17.3.1 Characteristics of transport networks 51317.3.2 Requirements and architectural goals of

MPLS-TP 51417.4 MPLS-TP functionality 516

17.4.1 MPLS-TP as a subset of MPLS 51617.4.2 MPLS-TP resilience functions 51717.4.3 MPLS-TP OAM functions 518

17.5 Deployment considerations 52217.6 Misconceptions about MPLS-TP 52617.7 Conclusion 52717.8 References 52717.9 Study questions 529

18 Conclusions 53118.1 Introduction 53118.2 Network convergence 53318.3 Interaction with client edge equipment 53618.4 Interprovider capability 53818.5 MPLS in the data communications network (DCN) 539


CONTENTS xvii

18.6 MPLS in mobile networks 54018.7 MPLS in the enterprise 54218.8 MPLS in the transport 54518.9 Final remarks 54518.10 References 546

Appendix A: Selected Backhaul Scenarios in MPLS-BasedAccess Networks 547

Appendix B: MPLS Resources 559

Appendix C: Solutions to Selected Study Questions 561

Appendix D: Acronyms 575

Index 587


About the Authors

Ina Minei joined Juniper Networks in 2000 and is currently Director ofIP and MPLS technologies. During this time she worked on the imple-mentation of LDP and RSVP-TE, helped define new protocol extensions,and worked with numerous customers on network design. Her focus hasbeen on next-generation network technologies, in particular MPLS pro-tocols and applications. She previously worked at Cisco for two years invarious software development projects for routers and switches. Ms Mineiis an active participant in industry forums and conferences and holdsseveral patents in the area of IP and MPLS. She earned a Master’s degreein computer science from the Technion, Israel.

Julian Lucek joined Juniper Networks in 1999 and is currently aDistinguished Systems Engineer in the Europe, Middle East and Africaregion, where he has been working with many service providers on thedesign and evolution of their networks. He previously worked at BT forseveral years, at first in the Photonics Research Department and later inthe data transport and routing area. During this time, he gained a PhD inultrahigh-speed data transmission and processing from Cambridge Uni-versity. He is the holder of several patents in the area of communicationstechnology. He has a Master’s degree in Physics from Cambridge Univer-sity and holds Juniper Networks Certified Internet Expert (JNCIE) #21.


ForewordYakov Rekhter, Juniper Fellow, Juniper Networks

Multi-Protocol Label Switching (MPLS) began in the mid-1990s with justtwo modest design objectives. The first was a better integration of ATMwith IP, a goal that we hoped could be met by providing a single IP-basedcontrol plane that would span both ATM switches and IP routers. Thesecond objective was to augment the IP control plane with some additionalfunctionality, namely traffic engineering using constraint-based routingthat was already present in the ATM control plane.

Not long after it started, MPLS usage was extended to applicationssuch as Circuit Cross Connect (CCC), ATM and Frame Relay serviceover an IP/MPLS infrastructure (draft-martini), BGP/MPLS VPNs (2547VPNs) and then Virtual Private LAN Services (VPLS). The originalconstraint-based routing functionality evolved beyond traffic engineer-ing to applications such as fast reroute and Differentiated Services TrafficEngineering (DiffServ-TE).

The idea of a single control plane for both ATM switches and IP routersevolved into Generalized Multi-Protocol Label Switching (GMPLS), whichprovides a single control plane that could span not only routers and ATMswitches but SONET/SDH and optical cross connects as well.

One of the recent MPLS developments deserving of mention here isthe use of MPLS in the access network. Expanding MPLS into the accessnetwork brings with it scalability challenges. The third edition describes asolution, known as ‘Seamless MPLS’, that addresses these challenges.

Since the first edition of this book, considerable progress has been madein the area of MPLS multicast, IP multicast with BGP/MPLS VPNs, andIP multicast with VPLS. Advances in these areas were included in thesecond edition of this book. The third and current edition further ex-pands upon these developments by covering such topics as supportingmulticast extranets in BGP/MPLS VPNs and supporting Internet multi-cast over an MPLS infrastructure. This edition also presents in detail thescalability comparison between two schemes of supporting multicast in


xxii FOREWORD

BGP/MPLS VPNs – the first one based on the PIM/GRE solution (knowninformally as ‘draft-rosen’), and the second based on NG multicast forL3VPN (BGP/MPLS mVPN). As an ever-increasing array of services hasbeen developed surrounding MPLS infrastructure, the importance of highservice availability and its successful provision has come to light. Sincethe second edition of this book, significant progress has been made in thearea of scalable fast protection based on the technique of local repair; thesedevelopments and their implications are included in this edition.

One important development since the publication of the second edi-tion of this book is MPLS Transport Profile (MPLS-TP), a technologydriven by the desire of the service providers to transition their transportinfrastructure from circuit-switched based technologies (SONET/SDH) topacket-based switching technology based on MPLS. This edition providesan overview of MPLS-TP, and clarifies the motivations behind and require-ments for its adoption.

It is important to keep in mind that in all of the applications mentionedabove, MPLS is just one of the components of such applications, albeita critical one. If we look back at the time when MPLS was created, andcompare its design objectives with what MPLS is used for today, we noticeseveral things. First of all, most of the applications of MPLS that we havetoday were not conceived of during the original design of MPLS, whilesome of the applications conceived of during the original design of MPLSare no longer relevant. For example, the original design goal of a betterintegration of ATM and IP routers by having a single control plane thatspans both ATM switches and routers is a thing in the past. And whilethe ability to offer ATM service over an IP/MPLS infrastructure is stillrelevant, it becomes less and less important relative to the Ethernet serviceover an IP/MPLS infrastructure. While originally MPLS was conceived asa technology solely for the Service Providers, we see today how MPLS isgradually penetrating the enterprise environment. Additionally, over timethe whole MPLS concept evolved from Multi-Protocol Label Switching toMulti-Purpose Label Switching.

A new technology quite often generates opposition, and MPLS was byno means an exception. You may all remember how MPLS was brandedby its opponents in negative terms as ‘bad’, ‘evil’, ‘a social disease’ or ‘anightmare of unprecedented proportions’. To put this in a proper perspec-tive, we need to keep in mind that technologies exist not for their ownsake but for the purpose of solving business problems. Therefore, talkingabout ‘good’ technologies versus ‘bad/evil’ technologies has little practicalrelevance; what is of great relevance is how well a particular technologymeets business needs.

One might wonder how to judge how well a particular technology,like MPLS, meets business needs. To answer this question I would like toinvoke the words of Cervantes’ Don Quixote: ‘the proof of the pudding


FOREWORD xxiii

is in the eating’, to which I would add: ‘and not in the debate about thepudding’. That being said, the ultimate judge of how well a particulartechnology meets business needs is the marketplace. It is the judgment ofthe marketplace that determines whether a particular technology deservesto live or to die; and with respect to MPLS the market made its verdict loudand clear – MPLS is here to stay.


Preface

In the three years since we began the previous edition of this book, somany new MPLS developments have taken place that our publisher andmany readers suggested that a third edition would be useful. Two par-ticular note-worthy developments since the second edition are SeamlessMPLS – an architecture to scale networks to 100,000+ MPLS nodes – andMPLS-TP, which provides the infrastructure for MPLS-based transportnetworks. The motivation for the book remains the same: MPLS is movingso fast that some of its new applications have already been deployed inproduction networks, yet are not described anywhere in book form. Inmany cases, the only available resources are the IETF drafts which listthe extensions needed to produce interoperable implementations. Thesedocuments often assume familiarity with the problem at hand and do notdiscuss why a particular solution has been chosen or explain its pros andcons. The third edition of MPLS-Enabled Applications attempts to fill thisgap and provide the reader with an understanding of both the problemand why the solution looks the way it does.

Therefore, when we describe the mechanisms underpinning an MPLSapplication, the emphasis is on giving an overview of the protocol ma-chinery without delving into the bits and bytes of each protocol message.This allows us to convey the concepts without making it difficult to see thewood for the trees. Also, some of the mechanisms that we write about arecurrently being defined, so details of the protocol messages may change,but the concepts are less likely to. References at the end of each chapterpoint to the documents describing the message formats and processingrules. Because a lot of the content in this book deals with technologiesthat are still, literally, works in progress, several things may happen. Firstly,some proposals may be abandoned or fail to become widely adopted. Sec-ondly, different vendors may introduce the technology at different times,and finally, the solution may evolve and change as implementation anddeployment experience is gained. Therefore, the fact that we discuss a


xxvi PREFACE

particular technology in this book does not guarantee that it is availableor deployed.

Although we both happen to work for the same router vendor, the bookis not vendor-specific. Occasionally, we point out some vendor-specificquirks if they are relevant to the discussion, or aid in understanding aparticular topic. Many of the topics discussed are still under debate in theIETF, and naturally our personal views on one topic or another may bestated more strongly than the opposing view.

WHO SHOULD READ THIS BOOK?

The intended audience of this book includes employees of network op-erators and network equipment vendors, customers of service providerswho are interested in the mechanisms underpinning the services that theybuy, network professionals who want to keep up to date with the latestadvances in MPLS and students of network technology. To make this bookmore accessible to both the student and to the practitioner of MPLS, wehave added study questions at the end of each chapter.

We assume that the reader has some degree of familiarity withnetwork technology and routing protocols, in particular BGP and thelink-state IGPs, but these are not a requirement to benefit from the book.Although our main aim is to cover the cutting-edge developments ofMPLS, the Foundation chapter allows the reader unfamiliar with MPLSto get up to speed in order to benefit from the remainder of the book.Even when discussing basic topics such as traffic engineering or fastreroute, we also explore the more interesting and advanced aspects of thetechnology.

WHAT IS NEW IN THE THIRD EDITION?

In this third edition, we aim to capture the latest developments in thefield. For this reason, we added three new chapters. Chapter 11 coversadvanced topics in multicast in L3VPNs, focusing on new developmentsin the BGP/MPLS scheme, which has gained significant deployment overthe last few years. Chapter 14 discusses advanced protection schemes forthe LSP tail-end, thus enabling sub 50 ms end-to-end service restoration.Finally, Chapter 17 provides an overview of MPLS-TP, the transport profilefor MPLS, which will form the foundation for packet-switched transportnetworks. Additional material was added and updated throughout thebook. Chapter 16, covering MPLS in access networks, has new sectionsdescribing the Seamless MPLS architecture, including the solutions forboth unicast and multicast. The book also has new material covering the


PREFACE xxvii

Live-live and Live-standby schemes for multicast resilience, point to mul-tipoint pseudowires, pseudowire redundancy and VPLS InterproviderOption E. The study questions at the end of each chapter are intendedto help readers test their understanding of the topics discussed and canserve to trigger debate on the pros and cons of a particular technology toa particular deployment.

HOW THIS BOOK IS ORGANIZED

The book is divided into three parts, each containing several chapters. PartOne describes the MPLS infrastructure tools used as the foundation tobuild services, Part Two covers the MPLS-based services and Part Threeexplores advanced topics.

The structure of Part One

Chapter 1, the Foundations chapter, reviews the control plane andforwarding plane mechanisms associated with MPLS. In this chapter, wegive an overview of the LDP and RSVP signaling protocols and comparethe two.

Chapter 2 discusses MPLS Traffic Engineering, which gives serviceproviders control over the path taken by traffic through their networkand the ability to give bandwidth guarantees. In this context, we look atthe impact of TE on network scalability, as well as at solutions for TE inLDP networks.

Chapter 3 explores the topic of Protection and Restoration in MPLSnetworks, essential to allowing MPLS networks to carry mission-criticaltraffic. We cover link and node protection, their respective scalingproperties and the cost of bandwidth protection. We also explore moreadvanced topics such as fate sharing and the new developments for pro-viding fast restoration in IP and LDP networks.

Chapter 4 presents Differentiated Services (DiffServ) Aware TrafficEngineering, which allows traffic engineering to be applied with per-classgranularity, bringing QoS to the network.

Chapter 5 introduces Interdomain Traffic Engineering. Both thesignaling and computation aspects are discussed, and path-computationelements are also reviewed.

Chapter 6 is devoted to MPLS multicast functionality. This chaptercovers not just P2MP LSP setup with RSVP and LDP but also advancedtopics such as upstream label allocation and hierarchies of P2MP LSPs.MPLS multicast is currently of great interest as it allows MPLS to be used


xxviii PREFACE

in broadcast TV and IPTV applications and because it is an essential partof the next-generation L3VPN multicast solutions discussed in Part Two.

The structure of Part Two

Chapters 7, 8, 9, 10 and 11 are devoted to Layer 3 VPNs – the mostwidespread application of MPLS to date. Chapters 7 through 9 focuson unicast traffic in VPNs. Chapter 7 provides a tutorial on L3VPN andexplains the basic concepts, Chapter 8 discusses more advanced topicssuch as route target filtering and scalability analysis, and Chapter 9 covershierarchical VPNs. Chapters 10 and 11 dive into the topic of multi-cast VPNs. Chapter 10 presents and compares the PIM/GRE and theBGP/MPLS solutions for multicast VPNs, while Chapter 11 focusesentirely on advanced topics such as extranet and inter-AS support in theBGP/MPLS solution, which has gained a lot of traction in the last fewyears.

Chapter 12 describes the rapidly growing area of Layer 2 transportover MPLS, including pseudowires and Layer 2 VPNs. These allowservice providers to migrate ATM and Frame Relay services to an IP/MPLSnetwork and to offer Ethernet-based alternatives to those services.

Chapter 13 describes the Virtual Private LAN Service (VPLS). Thisallows a service provider to offer a very simple-to-use service to enter-prise customers, in which the customer’s sites appear to be attached to thesame LAN. Multicast support over VPLS, an area which has seen a lot ofchange in recent years, is also discussed.

The structure of Part Three

Chapter 14 describes advances in protection schemes aimed at providing50 ms recovery times for end-to-end services. As we show in the chapter,a critical building block is providing protection of the LSP tail end.

Chapter 15 covers some aspects of the management and troubleshootingof MPLS networks. The subject of management of MPLS networks could fillan entire book by itself and a single chapter does not do it justice. However,we attempt to show some of the challenges (such as ICMP tunneling) andsome of the available tools, such as LSPing.

Chapter 16 provides an overview of the emerging trend of using MPLSin the access network, explains why this technology is taking off anddescribes the various deployment models, as well as describing the newand increasingly popular Seamless MPLS architecture.

Chapter 17 discusses the much-debated topic of MPLS-TP, the transportprofile for MPLS. MPLS-TP is currently the most active standardizationarea in MPLS. In order to track developments in this field, it is important

p1: ote/ote/sph p2: ote - download.e- · pdf filep1: ote/ote/sph p2: ote ......

Documents