IP, Ethernet and MPLS Networksdownload.e-bookshelf.de/download/0000/8137/34/L-G-0000813734... · Adapted and updated from Gestion des ressources et des défaillances dans les réseaux

File AttachmentC1.jpg

File Attachmentcover.jpg

IP, Ethernet and MPLS Networks

IP, Ethernet and MPLS Networks

Resource and Fault Management

Andr Perez

First published 2011 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc. Adapted and updated from Gestion des ressources et des dfaillances dans les rseaux IP, MPLS et Ethernet published 2009 in France by Hermes Science/Lavoisier LAVOISIER 2009

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd John Wiley & Sons, Inc. 27-37 St Georges Road 111 River Street London SW19 4EU Hoboken, NJ 07030 UK USA

www.iste.co.uk www.wiley.com

ISTE Ltd 2011 The rights of Andr Perez to be identified as the author of this work have been asserted by him in accordance with the Copyright, Designs and Patents Act 1988. ____________________________________________________________________________________

Library of Congress Cataloging-in-Publication Data Perez, Andre. [Gestion des ressources et des defaillances dans les reseaux IP, MPLS et Ethernet. English] IP, Ethernet, and MPLS networks : resource and fault management / Andre Perez. p. cm. Includes bibliographical references and index. ISBN 978-1-84821-285-5 1. Computer networks--Management. 2. Computer networks--Quality control. 3. Resource allocation. 4. Fault-tolerant computing. I. Title. TK5105.5.P471613 2011 004.6--dc22

2011006657

British Library Cataloguing-in-Publication Data A CIP record for this book is available from the British Library ISBN 978-1-84821-285-5 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne.

Table of Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Chapter 1. Network Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1. Basic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1. Layered structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.2. LANs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.3. WANs networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.1.4. Protocol architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.5. Addressing principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.2. IP technology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.1. Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.2. The IPv4 protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.3. The IPv6 protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.3. The MPLS technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.3.1. Label switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.3.2. Tables in the LSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.3.3. The PHP function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.3.4. The format of the MPLS header . . . . . . . . . . . . . . . . . . . . . 16 1.3.5. Encapsulation of a labeled packet . . . . . . . . . . . . . . . . . . . . 18

1.4. The ICMP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.5. Ethernet technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.5.1. The physical layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 1.5.2. The data link layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Chapter 2. Characterizing Quality of Service . . . . . . . . . . . . . . . . . . . 27

2.1. Quality of service functions . . . . . . . . . . . . . . . . . . . . . . . . . . 27 2.2. Quality of network operation . . . . . . . . . . . . . . . . . . . . . . . . . 29

2.2.1. Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

vi IP, Ethernet and MPLS Networks

2.2.2. Admissibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2.2.3. Loss ratio. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.2.4. Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.2.5. Jitter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.2.6. Classes of service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.2.7. Bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.3. Requirements of applications . . . . . . . . . . . . . . . . . . . . . . . . . 36 2.3.1. Voice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 2.3.2. Video . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 2.3.3. Application and control data . . . . . . . . . . . . . . . . . . . . . . . 43

2.4. The service contract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 2.4.1. The Enterprise section. . . . . . . . . . . . . . . . . . . . . . . . . . 46 2.4.2. The Service section . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 2.4.3. The Technology section. . . . . . . . . . . . . . . . . . . . . . . . . 48 2.4.4. The Quality of Service report . . . . . . . . . . . . . . . . . . . . . . . 50

Chapter 3. Transport Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.2. The TCP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.2.1. Format of the TCP header . . . . . . . . . . . . . . . . . . . . . . . . . 52 3.2.2. Initiating and closing a connection. . . . . . . . . . . . . . . . . . . . 54 3.2.3. Data transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 3.2.4. The slow start and congestion avoidance mechanisms . . . . . . . . 56 3.2.5. The fast retransmit and fast recovery mechanisms . . . . . . . . . . 57 3.2.6. The ECN mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.3. The UDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3.3.1. Format of the UDP header . . . . . . . . . . . . . . . . . . . . . . . . 60

3.4. The RTP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.4.1. Format of the RTP header . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.5. The RTCP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.5.1. Format of the SR message. . . . . . . . . . . . . . . . . . . . . . . . . 63 3.5.2. Format of the RR message . . . . . . . . . . . . . . . . . . . . . . . . 65

3.6. The DCCP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.6.1. DCCP procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 3.6.2. Congestion control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 3.6.3. Format of the DCCP header. . . . . . . . . . . . . . . . . . . . . . . . 67 3.6.4. Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

3.7. The SCTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 3.7.1. Format of the SCTP header . . . . . . . . . . . . . . . . . . . . . . . . 71 3.7.2. Association . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 3.7.3. Data transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Table of Contents vii

Chapter 4. Implementing Operation Quality . . . . . . . . . . . . . . . . . . . 77

4.1. The architectural framework. . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.2. Implementation of resource management . . . . . . . . . . . . . . . . . . 77

4.2.1. Relative QoS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.2.2. Guaranteed QoS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.2.3. Resource reservation scenarios . . . . . . . . . . . . . . . . . . . . . . 80 4.2.4. Mechanisms associated with the user plane . . . . . . . . . . . . . . 83 4.2.5. Load balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 4.2.6. Link optimization mechanisms . . . . . . . . . . . . . . . . . . . . . . 92

4.3. Implementing fault management . . . . . . . . . . . . . . . . . . . . . . . 96 4.3.1. Network reconfiguration. . . . . . . . . . . . . . . . . . . . . . . . . . 96 4.3.2. Fault detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 4.3.3. Equipment reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . 98

Chapter 5. IP Technology Resource Management . . . . . . . . . . . . . . . 101

5.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5.2. The DiffServ model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

5.2.1. The DSCP field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 5.2.2. The DiffServ architecture . . . . . . . . . . . . . . . . . . . . . . . . . 106

5.3. The IntServ model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 5.3.1. Principles of resource reservation . . . . . . . . . . . . . . . . . . . . 109 5.3.2. The RSVP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

5.4. The ARSVP protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 5.4.1. Principles of aggregation . . . . . . . . . . . . . . . . . . . . . . . . . 123 5.4.2. The ARSVP procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . 124

Chapter 6. IP Technology Fault Management . . . . . . . . . . . . . . . . . 127

6.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 6.2. Hot Standby Router Protocol . . . . . . . . . . . . . . . . . . . . . . . . . 128

6.2.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 6.2.2. Format of the HSRP message. . . . . . . . . . . . . . . . . . . . . . . 129 6.2.3. Load balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

6.3. Virtual Router Redundancy Protocol . . . . . . . . . . . . . . . . . . . . . 132 6.3.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 6.3.2. Format of the VRRP message . . . . . . . . . . . . . . . . . . . . . . 132

6.4. OSPF protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 6.4.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 6.4.2. Format of the OSPF message . . . . . . . . . . . . . . . . . . . . . . . 135 6.4.3. Restarting the OSPF protocol . . . . . . . . . . . . . . . . . . . . . . . 139

6.5. Border Gateway Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 6.5.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

viii IP, Ethernet and MPLS Networks

6.5.2. Format of the BGP message . . . . . . . . . . . . . . . . . . . . . . . 145 6.5.3. Path attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 6.5.4. Route selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 6.5.5. BGP restart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Chapter 7. MPLS Technology Resource Management . . . . . . . . . . . . 153

7.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 7.2. Support for DiffServ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

7.2.1. Types of virtual circuits . . . . . . . . . . . . . . . . . . . . . . . . . . 154 7.2.2. Interaction between markings. . . . . . . . . . . . . . . . . . . . . . . 156

7.3. Traffic engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 7.3.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 7.3.2. The RSVP-TE protocol . . . . . . . . . . . . . . . . . . . . . . . . . . 160 7.3.3. OSPF-TE protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

Chapter 8. MPLS Technology Fault Management . . . . . . . . . . . . . . 177

8.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 8.2. The LDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

8.2.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 8.2.2. Format of the LDP PDU. . . . . . . . . . . . . . . . . . . . . . . . . . 180 8.2.3. The LDP messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 8.2.4. Restarting the LDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

8.3. The RSVP-TE protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 8.3.1. Failure detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 8.3.2. Restarting the RSVP-TE protocol . . . . . . . . . . . . . . . . . . . . 190

8.4. The FRR mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 8.4.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 8.4.2. Extensions of the RSVP-TE protocol . . . . . . . . . . . . . . . . . . 195 8.4.3. Procedure of the FRR mechanism . . . . . . . . . . . . . . . . . . . . 199

Chapter 9. Ethernet Technology Resource Management . . . . . . . . . . 203

9.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 9.2. Priority management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 9.3. Resource reservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

9.3.1. The bandwidth manager . . . . . . . . . . . . . . . . . . . . . . . . . . 207 9.3.2. The SBM protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

9.4. Flow control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 9.5. The access network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

9.5.1. Architecture of the PON . . . . . . . . . . . . . . . . . . . . . . . . . . 218 9.5.2. Priority management in EPON . . . . . . . . . . . . . . . . . . . . . . 220 9.5.3. Priority management in GPON . . . . . . . . . . . . . . . . . . . . . . 221

9.6. The aggregation network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Table of Contents ix

Chapter 10. Ethernet Technology Fault Management . . . . . . . . . . . . 227

10.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 10.2. The STP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

10.2.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 10.2.2. Format of the BPDU message. . . . . . . . . . . . . . . . . . . . . . 229 10.2.3. Procedure of the STP . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

10.3. The RSTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 10.3.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 10.3.2. Format of the BPDU message. . . . . . . . . . . . . . . . . . . . . . 234 10.3.3. Procedure of the RSTP . . . . . . . . . . . . . . . . . . . . . . . . . . 235

10.4. The MSTP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 10.4.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 10.4.2. Format of the BPDU message. . . . . . . . . . . . . . . . . . . . . . 238 10.4.3. Procedure of the MSTP . . . . . . . . . . . . . . . . . . . . . . . . . 240

10.5. Link aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 10.5.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 10.5.2. The LACP message . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 10.5.3. The Marker protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

10.6. The aggregation network . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 10.6.1. Operating principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 10.6.2. The APS protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

Preface

This book addresses two aspects of network operation quality; namely, resource management and fault management.

Network operation quality is among the functions to be fulfilled in order to offer quality of service, QoS, to the end user. It is characterized by four parameters:

packet loss;

delay;

jitter, or the variation of delay over time;

availability.

Resource management employs mechanisms that enable the first three parameters to be guaranteed or optimized. Fault management aims to ensure continuity of service.

Internet Protocol (IP), Multiprotocol Label Switching (MPLS) and Ethernet are the main technologies deployed by operators, in the case of wide area networks (WANs), and by businesses, in the case of local area networks (LANs). Initially, these technologies were not designed to deal with resource management, which partly explains their simplicity and commercial success. The features offered were sufficient at the time to offer services without QoS (best effort) constraints, such as web access and electronic messaging.

Resource management is indispensable when QoS constraints need to be taken into account. Voice and video are two applications illustrative of strong QoS requirements. Resource management can be provided using two different approaches:

xii IP, Ethernet and MPLS Networks

resources are managed node by node. No resource is allocated to a flow from end to end. In this case, there is a risk of congestion, as the resources of one node may become insufficient to meet demand. The network thus offers relative QoS;

resources are allocated to a flow from end to end. The Connection Admission Control (CAC) function allows congestion to be avoided in the network, in that case offering guaranteed QoS.

Deployed networks essentially implement relative QoS, as they have the advantage of simplicity compared with guaranteed QoS. The oversizing of current networks, associated with traffic measurement functions that feed into capacity planning, has so far enabled congestion problems to be solved in a satisfactory manner.

Such a mode of operation can prove inadequate in cases of large increases in video traffic. The latter can be classified according to two communication patterns:

broadcast video. A video source broadcasts one and the same program to several users. Resource consumption depends on the number of sources and is relatively independent from the number of users. A network offering relative QoS allows for the transport of this type of traffic, whose increase is not dramatic;

unicast video. The video signal is exchanged in real time between a source and a user (on-demand video) or between two users. Resource consumption depends on the number of users. Given the required throughput per video program, oversizing the network may not be an adequate response. A network offering guaranteed QoS may prove to be the only option for this type of traffic.

Fault management relies on devices that enable a reconfiguration of the network following a fault in a node or link, and the reconfiguration of a node following a fault in the data processing board of the control plane.

The main parameter associated with the reconfiguration of the network is the convergence time. This is relatively long for IP and Ethernet networks. Depending on network size, it can reach several tens of seconds. These values may be sufficient for businesses that deploy LANs. In contrast, operators that deploy WANs require substantially quicker times and seek values less than one second or even one tenth of a second.

The reconfiguration of the node, upon its detection by adjacent nodes, will cause a first reconfiguration of the network when the processor board is faulty, and a second reconfiguration of the network once switching to the standby board has taken place. The purpose of Graceful Restart-type mechanisms is to avoid this double switching.

Preface xiii

Consequently, this book is structured in 10 chapters; the main topics discussed therein are summarized in the following table.

Chapter Designation Description

1 Network Operation IP, MPLS, Ethernet technologies

2 Characterizing Quality of Service Operation quality parameters, requirements of applications, the service contract

3 Transport Protocols The TCP, UDP, RTP, DCCP, and SCTP protocols

4 Implementing Operation Quality Mechanisms associated with the user plane

5 IP Technology Resource Management

Relative QoS: the DiffServ model; Guaranteed QoS: the IntServ model and the RSVP protocol

6 IP Technology Fault Management

Network reconfiguration: LAN-side HSRP and VRRP protocols, WAN-side OSPF and BGP routing protocols. Node reconfiguration: the Graceful Restart mechanism

7 MPLS Technology Resource Management

Relative QoS: DiffServ support; Guaranteed QoS: traffic engineering and the RSVP-TE and OSPF-TE protocols

8 MPLS Technology Fault Management

Node reconfiguration: the Graceful Restart mechanism and the LDP and RSVP-TE protocols. Network reconfiguration: the FRR mechanism

9

Ethernet Technology Resource Management

Relative QoS: frame tagging; the PON access network. Guaranteed QoS: the SBM protocol

10 Ethernet Technology Fault Management

Network configuration: the STP, RSTP, and MSTP protocols; linear protection. Link reconfiguration: the LACP protocol

Chapter 1

Network Operation

1.1. Basic concepts

The purpose of the network is to convey data between the terminal stations, or hosts. The network is comprised of nodes that are interconnected with one another by links. The network nodes perform two elemental functions:

transmission: this function allows for the adaptation of the data to be transmitted on the transmission medium (copper pair, optical fiber, free space);

connectivity: this function allows for the transfer of the data between an input and an output of the network node.

The network includes two entities:

the local area network (LAN): this is a private network, deployed inside a company, and provides a connection between hosts (client stations, servers, telephone terminals);

the wide area network (WAN): this is a public network, deployed by operators or service providers, and provides interconnection between LANs. The WAN is comprised of access networks, aggregation networks, and a core network.

1.1.1. Layered structure

The transmitted data are structured in layers (Figure 1.1). Each layer implements an encapsulation that corresponds to a PCI (protocol control information) header including fields whose interpretation is defined by a protocol. The Internet or

2 IP, Ethernet and MPLS Networks

Transmission Control Protocol (TCP)/Internet Protocol (IP) model is composed of three layers:

the application layer (layer 7), which enables communication between distant software applications. The obtained data structure constitutes the message;

the transport layer (layer 4): the TCP and UDP (User Datagram Protocol) are the most common protocols. The transport layer enables a remote application to be addressed through a target port number. The TCP additionally makes it possible to check that the data transfer was performed properly from end to end. The Real-time Transport Protocol (RTP) complements the UDP. It is used for real-time applications (e.g. voice or video). The obtained data structure constitutes a segment;

the network layer (layer 3): the Internet model defines two versions of the IP, IPv4 and IPv6. The network layer allows data to be transferred across the network. The information used is the destination IP address. The network nodes that perform such a transfer are the routers. The network layer therefore fulfils the connection function. IP operates in a non-connected mode: the data transfer is not conditional upon the implementation of a path between the two endpoints. The obtained data structure constitutes a packet.

Physical Layer

Data Link Layer

IP

TCP - UDP 4

3

2

1

DNS, DHCP, SNMPTELNET

HTTP, SMTP, POP3FTP 7

Internet Model

PCISegment

PCIPacket

PCIFrame

PCIFrame/Bit

Message

Layer Number Structure Name Encapsulation

Header : PCIProtocol Control Information

PCI

Figure 1.1. Layered structure

Network Operation 3

The protocols for the data link layer (layer 2) and the physical layer (layer 1) depend on the type of network traversed:

the purpose of the data link layer is to delineate the beginning and end of the transmitted data structure. Other functions may be added, such as flow control or error checking. The obtained data structure constitutes a frame;

the purpose of the physical layer is to define the interface of the node (level, impedance, sensitivity, line code, or modulation). The physical layer performs the transmission function. Some protocols also define a header whose main function is to synchronize the physical frame.

1.1.2. LANs

Typical LANs deployed in large businesses consist of several blocks (Figure 1.2):

the Users block allows the client stations, i.e. the network users, to be connected;

the Server block encompasses all the companys application servers;

the Management block comprises the servers and the supervision and administration stations of the LAN;

the Access block provides access to the WAN network;

the Core network block performs the interconnection of the various blocks.

The equipment that makes up the Users, Servers, and Management blocks belongs to two types:

access switches; and

distribution switches. Distribution switches are also used in the Core and WAN Access blocks.

An access switch is a machine that employs Ethernet technology. The latter specifies the protocols for the physical layer and the data link layer. The physical layer uses a copper pair or optical fiber as the transmission medium. The data link layer allows the data to be transferred across the LAN. The information used to perform the transfer is the target MAC (Media Access Control) address contained in the Ethernet header. The Ethernet data link layer thus performs a connectivity function. The switching table of the switch is acquired through a learning process, based on the source MAC addresses of the received frames. Similar to the IP, the Ethernet operates in a non-connected mode.


The distribution switch is a multi-level or multi-layer switch (MLS). It can provide connectivity between an input port and an output port using the following layers:

the Ethernet data link layer;

the IP network layer; and

the TCP or UDP transport layer.

Level 3 connectivity is identical to that used by an IP router; the difference lies in the interfaces. An MLS has Ethernet interfaces and is less constrained resource-wise than an IP router, which has interfaces with the WAN whose throughput rates might be lower.

Level 4 (transport layer) connectivity may be used at the Servers block. Connectivity is then based on port numbers. Level 4 connectivity allows users to utilize the same IP address (that of the MLS) to reach the various servers without distinction.

WAN

AccessSwitch

DistributionSwitchMLS

CoreBlock

UsersBlock

ServersBlock

ManagementBlock

WANAccessBlock

UsersBlock

Router

MLS

MLS

Figure 1.2. Architecture of a LAN

Network Operation 5

1.1.3. WANs networks

It is customary for WANs to be structured in three entities (Figure 1.3):

the Access Network, which allows a client to be connected to the first technical site of the operator;

the Aggregation Network, which performs the collection of traffic from access networks;

the Core Network, a mesh network that provides the interconnection of the various aggregation networks and an interface with different operators networks.

The combination of the access network and aggregation network is based on the transport of Ethernet frames. Such transport is provided by the Q-in-Q, VPLS (Virtual Private LAN Service), and VPWS (Virtual Private Wire Service) Ethernet technologies.

An IP router provides an interconnection with the core network.

The MPLS (Multi-Protocol Label Switching) technology is often employed by operators in the core network. The OTN (Optical Transport Network)/DWDM (Dense Wavelength-Division Multiplexing) technology enables large throughput rates to be achieved for the interconnection of the equipment in the aggregation network and core network.

AggregationNetwork

CoreNetwork

AccessNetwork

Figure 1.3. Architecture of a WAN

The MPLS technology operates in a mixed mode for the following reasons:

IP connectivity (non-connected mode) is provided at an edge node of the network;

a virtual circuit, called an LSP (Label Switching Path), is implemented between two edge nodes. Equipment within the MPLS network performs label switching.


The VPWS and VPLS technologies allow Ethernet frames to be transported across the network using the concepts of MPLS technology. The VPWS technology enables the provision of a point-to-point link service. The VPLS technology enables the provision of a multiple-access broadcasting network emulation service. As with the MPLS technology, the connection mode is mixed for the following reasons:

Ethernet connectivity (non-connected mode) is provided at an edge node of the network;

a virtual circuit, referred to as a Pseudo-Wire, is created between two edge nodes. As with the MPLS network, equipments within the VPLS or VPWS network perform label switching.

The OTN/DWDM technology relies on two types of multiplexing:

a time multiplexing of tributaries at 1 Gb/s (Giga Ethernet Interface) or 2.5 Gbit/s (STM-16 interface), or at 10 Gbit/s (STM-64 or 10G Ethernet interfaces) into a 40 Gbit/s resultant;

a wavelength multiplexing, DWDM, into an 80- or 160-wavelength resultant, wherein each wavelength can support a 40 Gbit/s rate.

Optical fibers are used as the transmission medium employed to connect the equipment in the aggregation network and the core network, whereas three types of medium are used in the access network:

optical fibers, with point-to-point Ethernet links for the connection of large businesses. A wiring program using optical fibers for residential users is underway in France, with point-to-point Ethernet or point-to-multipoint PON (Passive Optical Network) links. Optical fibers are mainly used in dense urban areas;

copper pairs, with ADSL (Asymmetric Digital Subscriber Line) or SHDSL (Single pair High-speed Digital Subscriber Line) links for the connection of residential users, professionals, and small and medium businesses. Copper pairs are used in urban areas of high or medium density;

free space, with WiMax wireless links for rural areas, or WiFi in the case of HotSpots. WiFi technology is also present in LANs.

1.1.4. Protocol architecture

The different operations performed by network components are described in Figure 1.4.

Network Operation 7

WAN LANLAN

IP

Eth

Eth

Eth

Eth

Eth

Eth

IP

Eth

Eth

IP

L2

L1

L4

L7

Transmission

IP

Eth

Eth

L4

L7

Eth

Eth

Eth

Eth

IP

Eth

Eth

IP

L2

L1

Connectivity

switch switchrouter router

Connectivity

Host Host

Transmission

IP

L2

L1

IP

L2

L1

Connectivity Connectivity

Connectivity

Transmission Transmission Transmission Transmission

Figure 1.4. Protocol architecture

The end user (for example, a PC or server) carries out various encapsulations before transmitting the data to its access switch. Based on the received data, the switch performs the following operations:

when receiving, it unencapsulates the physical layer and the Ethernet data link layer, retrieves the destination MAC address, and searches the switching table for the output port;

when transmitting, it performs the encapsulation of the data link layer and the Ethernet physical layer and transmits the data to the router. It should be noted that the IP packet is not interpreted by the switch.

The router performs the following operations:

when receiving, it unencapsulates the physical layer, the Ethernet data link layer and the network layer IP, retrieves the destination IP address, and searches the routing table for the output port;

when transmitting, it performs the encapsulation of the IP network layer, of the data link layer and of the physical layer, and sends the data to the WAN. It should be noted that the segment is not interpreted by the router.

The physical layer and data link layer protocols, at the interface between the router and the WAN, are specified by the service provider or operator. The WANs IP routers perform the same basic functions as that which interconnects the LAN to the WAN. The operation of MPLS routers will be explained later.

Networks are partitioned into a plurality of AS (Autonomous System) domains. All the equipment in a domain is administered by a single entity.


Inside a domain, routers exchange information that allows the routing table to be dynamically populated. This information, which contains the destination networks and associated metrics, is carried by IGP (Internal Gateway Protocol) routing protocols.

The edge routers of the different domains are interconnected with one another. They, too, exchange information, which is carried by EGP (External Gateway Protocol) routing protocols.

1.1.5. Addressing principles

Two types of relationship between terminals are defined in the Internet model; the difference between them arises from the way in which the session (socket) identifiers are implemented:

the client-server relationship;

the peer-to-peer relationship.

The relationship between terminals utilizes a session identified by four identifiers:

the IP address of the source;

the IP address of the destination;

the port number of the source;

the port number of the destination.

In a client-server relationship, the session is always initialized by the client that sets up the socket:

the IP address of the source may be static (manually configured) or dynamic (obtained by sending a request to a DHCP (Dynamic Host Configuration Protocol) server);

the IP address of the destination is obtained dynamically from a DNS (Domain Name System) server. The information provided to the DNS server to obtain that IP address is an URL (Uniform Resource Locator; Figure 1.5);

the port number of the source is a random number greater than 1,024;

the destination port number depends on the application to which the client is establishing a connection.

In a peer-to-peer relationship, each peer defines a half-socket including the IP address and port number of the source. Additional messages need to be exchanged

Network Operation 9

between the two peers in order for each end to retrieve the half-socket determined by the other end. Voice transmission is an example of a peer-to-peer relationship. SIP (Session Initiation Protocol) and SDP (Session Description Protocol) signaling messages enable each entity to complete its socket.

Protocol Web Server FileDirectory

Application Domain name

URL: Universal Resource Locator

Destination @ IP : 65.12.125.75 (decimal notation)

DNS Server Query

Destination @ MAC : 00:C0:9F:B7:60:37 (hex notation)

WEB Server Queryor

Gateway Router Query

DNS Protocol

ARP Protocol

Browser

http://www.hermes.com/book/network.htm

Figure 1.5. Mechanisms for retrieving destination addresses

The identifiers used for the Ethernet data link layer are the MAC (Media Access Control) addresses of the source and destination. The MAC address of the source is generally that provided by the network card vendor. The destination MAC address is obtained from the destination IP address using the ARP (Address Resolution Protocol) protocol (Figure 1.5):

the destination MAC address is that of the recipient if the latter is in the same network as the source in the IP addressing scheme sense;

the destination MAC address is that of the router, which provides a way out of the LAN, if the source and recipient are in two different networks in the IP addressing scheme sense.

1.2. IP technology

1.2.1. Routing

The main purpose of the router is to transfer received packets to an output port. In order to carry out this function, the router needs to determine, for each received

http://www.hermes.com/book/network.htm


packet, the IP address of the next hop. That information is obtained from the routing protocols used, in order to construct two tables:

RIB (Routing Information Table). This table is located in the processor board performing the processing of data belonging to the control plane;

FIB (Forwarding Information Table). This table is located in the line card performing the processing of data belonging to the user plane.

The RIB table is directly populated from the routing protocols. The FIB table is a copy of the RIB table. When a packet is received, a decision is made directly at the line card level, without having to search the RIB table. This arrangement significantly improves the routers transfer capacity.

1.2.2. The IPv4 protocol

The IPv4 protocol resides in each end station or host in the network, and in each router involved in conveying the packet. It fulfils two basic functions: routing and fragmentation. Routing is based on the destination address contained in the header. Fragmentation consists of constructing reduced-size packets, according to the type (in the level 2 protocol sense) of network being traversed.

The IPv4 protocol, in its basic version, does not provide a reliable packet routing service. There is no acknowledgment among routers, no error checking of the data encapsulated by the IP header, and no flow control. The IP only performs error checking on the header.

The IPv4 protocol header contains the following fields (Figure 1.6).

0 1 2 30 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

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+|Version| IHL |Type of Service| Total Length |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| Identification |Flags| Fragment Offset |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| Time To Live | Protocol | Header Checksum |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| Source Address |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| Destination Address |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 1.6. Format of the IPv4 header

Network Operation 11

Version: this field comprises the version number of the IP. Its value equals four.

IHL (Internet Header Length): this field comprises the size of the IP header as a multiplier of 4 bytes. For a header with no options, the value of this field equals five.

Type of service: this field was defined in the original version of the standard in order to implement priority management, and to provide indications on the required level of delay, rate, and reliability. It has been superseded by the DSCP (DiffServ Code Point) field, which identifies a set of flows for which quality-of-service mechanisms are defined.

Total length: this field comprises the size of the packet, including the IP header and the encapsulated data.

Identification: this field comprises a value allowing the fragments of the IP packet to be reassembled upon reception.

Flags: this field comprises three bits that are used as follows:

the first bit is always set to 0;

the second bit determines whether fragmentation is enabled (bit set to 0) or disabled (bit set to 1);

the third bit identifies the last fragment (bit set to 0) or the intervening fragments (bit set to 1).

Fragment offset: this field indicates the position of the fragment within the initial packet.

Time to live (TTL): this field comprises the maximum number of routers crossed by the packet. Each crossed router decrements by one unit the value of this field. When the value equals zero, the packet is discarded and an ICMP error message is returned to the source.

Protocol: this field comprises a value allowing the identification of the type of data encapsulated by the IP header.

Header checksum: this field comprises a checksum calculated solely on the header.

Source address: this field comprises the IP address of the packet source.

Destination address: this field comprises the IP address of the packet recipient.


Options: the options defined in the original version of the standard are as follows:

loose source routing: this field contains a non-exclusive list of routers to be crossed by the IP packet (source routing);

strict source routing: this field contains an exclusive list of routers to be crossed by the IP packet (source routing);

record route: this field contains the list of all routers crossed by the IP packet;

Internet timestamp: this field contains the time at which a packet crossed a router.

1.2.3. The IPv6 protocol

The main modification brought in by the IPv6 protocol regards the size (16 bytes) allocated to the field of the source and destination addresses, which results in a larger-sized header (40 bytes).

The IPv6 protocol introduced simplifications over the IPv4 protocol, allowing for better performance in terms of packet processing by routers. The following fields were deleted from the IPv6 header:

IHL: the IPv6 header has a fixed length;

identification, flags, fragment offset: fragmentation is dealt with through an extension;

header checksum: it is considered that the transmission is of a good quality and that binary errors are infrequent.

Options were removed from the basic header and replaced by new headers, called extensions, which can be ignored by intermediary routers. Apart from the hop-by-hop option, which is processed by all intermediary routers, other options are only taken into account by the terminal stations.

The IPv6 header contains the following fields (Figure 1.7).

Version: this field presents the version number of the IP. Its value equals six.

Traffic class: this field is equivalent to the DS field in the IPv4 header.

Flow label: this field presents a unique number chosen by the source, whose purpose is to facilitate the task of the routers in implementing the quality of service functions for a particular flow.

Network Operation 13

0 1 2 30 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

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+|Version| Traffic Class | Flow Label |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| Payload Length | Next Header | Hop Limit |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| |+ +| |+ Source Address +| |+ +| |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+| |+ +| |+ Destination Address +| |+ +| |+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 1.7. Format of the IPv6 header

Payload length: this field presents the size of the data encapsulated by the IP header.

Next header: this field presents the identifier of the next header of the data encapsulated by the IP header. This field has the same role as the Protocol field in the IPv4 header. This field allows the extensions of the IPv6 header to be chained:

hop-by-hop options: this extension is processed by all routers crossed by the IP packet. It includes the router alert option, which is used when sending an RSVP message;

routing: this extension enables source routing and contains a non-exclusive list of the routers to be crossed by the IP packet (loose source routing);

fragment: this extension supports the fragmentation of an IPv6 packet. Unlike the fragmentation carried out by the router under IPv4, fragmentation in IPv6 must be carried out by the sender of the packet. For that purpose, the sender therefore has to implement a packet size discovery technique;

destination options: this extension presents options that are processed by the recipient;

Documents

IP, Ethernet and MPLS Networksdownload.e-bookshelf.de/download/0000/8137/34/L-G-0000813734... · Adapted and updated from Gestion des ressources et des défaillances dans les réseaux