OSPF: Anatomy of an Internet Routing Protocol

Product Author Bios

John T. Moy is a Senior Consulting Engineer at Ascend Communications. He is the author of the OSPF and MOSPF protocol specifications and currently chairs the OSPF and MOSPF Working Groups in the Internet Engineering Task Force. Mr. Moy has been involved in the design and development of router software for 15 years, currently at Ascend, and previously at Proteon and at Bolt Beranek and Newman. Mr. Moy holds a Master of Arts in Mathematics from Princeton University and a Bachelor of Engineering in Mathematics from the University of Minnesota.

0201634724AB04062001

Written for TCP/IP network administrators, protocol designers, and network application developers, this book gives the most complete and practical view ever into the inner workings of Internet routing. The book focuses on OSPF (Open Shortest Path First), a common TCP/IP routing protocol that provides robust and efficient routing support in the most demanding Internet environments. A methodical and detailed description of the protocol is offered and OSPF's role within the wider context of a TCP/IP network is demonstrated.

Practical throughout, the book provides not only a theoretical description of Internet routing, but also a real-world look into how theory translates into practice. It shows how algorithms are implemented, and how the routing protocols function in a working network where transmission lines and routers routinely break down.

You will find clear explanations of such routing fundamentals as how a router forwards packets, IP addressing, CIDR (Classless Inter-Domain Routing), the routing table, Internet routing architecture, and the two main routing technologies: Distance Vector and link-state algorithms. OSPF is discussed in depth, with an examination of the rationale behind OSPF's design decisions and how it has evolved to keep pace with the rapidly changing Internet environment. OSPF topics covered by the book include the following:

• OSPF areas and virtual links
• NBMA (Nonbroadcast multi-access) and Point-to-MultiPoint network segments
• OSPF configuration and management
• Interaction with other routing protocols
• OSPF cryptographic authentication
• OSPF protocol extensions, including the Demand Circuit extensions and the multicast extensions to OSPF (MOSPF)
• An OSPF FAQ

IP multicast and multicast routing are also discussed. Methods for debugging routing problems are explained, including a catalog of available debugging tools. The book also offers side-by-side comparisons of all the unicast and multicast routing protocols currently in use in the Internet.

You will come away from this book with a sophisticated understanding of Internet routing and of the OSPF protocol in particular. Moreover, the book's practical focus will enable you to put this deeper understanding to work in your network environment.

Preface

Introduction The Internet is a global communications network. With connections in more than 100 countries, tens of millions of people use the Internet for business, education, and recreation. Electronic commerce is beginning on the Internet as businesses connect to sell their products and services. Academics collaborate over the Internet by exchanging electronic mail. People can converse using Internet phones, send faxes, participate in online chats and bulletin boards, play multiuser games, and experiment with virtual environments. Special-purpose computers called routers connect the Internet together. As data is forwarded from one place in the Internet to another, it is the routers that make the decisions as to where and how the data is forwarded. The protocols that dynamically inform the routers of the paths that the data should take are called routing protocols. It is the job of these protocols to react quickly to changes in the Internet's infrastructure, such as transmission lines going in and out of service, routers crashing, changes in network policies, and so on. Routing is what makes the Internet tick. Although many users of the Internet and the World Wide Web are unaware of the machinery underlying the network applications, routing is an interesting but complicated subject. Routing protocols are sophisticated distributed algorithms that must also be extremely robust to keep a large, decentralized network like the Internet running smoothly. Audience This book is for students of data communications, TCP/IP network administrators, protocol designers, developers of routing protocol software, and other professionals involved in the design, development, and management of TCP/IP networks. This book is a practical, hands-on description of Internet routing rather than a theoretical treatment. Although we describe how the various protocols were intended to work, we also describe how well the design has translated into practice. Internet protocol design is a practical undertaking itself, with efficiency of implementation often dictating design choices. For this reason, this book gives an in-depth treatment of how a router really works. Instead of just describing the algorithms, this book goes beyond to show how the algorithms are implemented. We often present ideas in a historical context, showing how Internet protocols have evolved. This is done for two reasons. First, you can learn a lot from the mistakes (and successes) of the past. Second, in order to participate in Internet discussion groups, many of which are dominated by old-timers, it is good to have some context. This book is not an elementary introduction to TCP/IP and its routing. Instead we assume that you have some familiarity with the TCP/IP protocol suite and some exposure to the basic concepts of routing. These assumptions allow us to explore many of the facets of Internet routing in greater detail than possible in an introductory text. Organization of This Book

This book is organized into five parts. Part I sets the groundwork for a discussion of Internet routing. After a brief description of how routing fits together with the rest of the Internet's protocols, Chapter 1 describes in depth how a router forwards packets. This discussion naturally leads to an explanation of IP addressing and CIDR, as well as of the interaction of hosts and routers. Internet routing protocols are introduced in Chapter 2, beginning with a treatment of the end product of all routing protocols: the router's routing table. Chapter 2 ends with an overview of the Internet's routing architecture and the two main routing technologies in use in today's Internet: Distance Vector and link-state algorithms.

Part II describes the Internet's OSPF routing protocol. We start in Chapter 3 with an explanation of why the OSPF protocol was developed in the first place. Chapter 4 discusses the basics of link-state routing; Chapter 5, how OSPF behaves over various subnet technologies; Chapter 6, its use of hierarchical routing; and Chapter 7, extensions to OSPF. Each chapter not only describes how OSPF works but also explains why it works that way. We explore the reasons behind OSPF's design decisions and how the OSPF protocol has evolved to keep pace with the rapidly changing Internet environment. Part II concludes with an OSPF FAQ (Chapter 8).

Part III (Chapters 9 and 10) describes TCP/IP multicast routing, including broadcast and multicast forwarding, the MBONE, and the two distinct types of multicast routing protocols: source-based trees and shared-tree algorithms. As we did with unicast routing, we go further into the subject of multicast routing through the examination of a particular multicast routing protocol: the Multicast Extensions to OSPF (MOSPF).

Part IV covers the configuration and management of Internet routing. The configuration and management of OSPF is explained in detail in Chapter 11. Chapter 12 describes the tools used to monitor and debug routing in a TCP/IP network. For each tool, we describe its use, how it works, and its advantages and drawbacks.

Part V is a comparison of Internet routing protocols. Chapter 13 compares and contrasts the routing protocols in use in the Internet: RIP, OSPF, BGP, IGRP, and IS-IS. In Chapter 14, we examine the available multicast protocols: DVMRP, MOSPF, PIM Dense and Sparse, and CBT.

Following Chapter 14 is an extensive bibliography arranged and numbered in alphabetical order. Within the text, the citation 85, for example, refers to item 85 in the bibliography.

The companion book OSPF Complete Implementation, in keeping with the Internet tradition that reveres "working code" over all else, explores even further the mechanics of Internet routing through examination of a real, working OSPF implementation. The book contains a complete implementation of OSPF on CD. Written in C++, the OSPF implementation is intended to be portable to a wide range of environments. Two sample ports are included: an OSPF routing daemon (called ospfd) for FreeBSD 2.1 and an OSPF routing simulator that can be run on Windows 95. The OSPF implementation has been developed using publicly available tools.

I would like to thank the technical reviewers who improved this book through their thoughtful and timely reviews: Ran Atkinson, Eural Authement, Fred Baker, Howard Berkowitz, Jeffrey Burgan, Joel Halpern, Mukesh Kacker, Robert Minnear, Jim Reid, and W. Richard Stevens. Thanks also to Tim Stoddard and the Arkansas Public School Computer Network (APSCN) for letting me collect OSPF statistics on the APSCN network and use that network as an example of OSPF configuration in Chapter 11, Configuration and Management. Thanks to S. Randall McLamb for drawing the figures.

I would also like to acknowledge the help of my editors at Addison Wesley Longman over the long life of this project: Carol Long, Karen Gettman, and Mary Harrington.

And special thanks to my wife, Sonya Keene, who designed the book, edited rough drafts, created the index, and gave encouragement while this book was being written.

Index

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

A

ABRs (area border routers), 122-123

definition of, 158-159

in MOSPF, 201

configuration of, 226

Abstraction, 55, 102

of broadcast subnets, 108-110

of NBMA subnets, 113

of Point-to-MultiPoint subnets, 116

Adapters in promiscuous mode, 189, 261

Add-ons for IP datagram forwarding, 13-14

add route command, 30

Address prefixes, 16-18

Address scoping, 23

Addresses

IP, 8, 16-22

lending, 18

multicast, 175

Adjacent neighbors, 90

Administrative scoping, 175

Advanced Peer-to-Peer Networking (APPN) protocol, 49

Advanced Research Projects Agency network (ARPANET), 3, 33

routing algorithm in, 48-50

sequence bug in, 56, 77

synchronization in, 89

ADVERTISER attribute, 296

Advertising routers, 75-76, 156

Aggregation, 17-18

in BGP, 290

in IGRP, 299

in IS-IS, 302

in MOSPF, 199-203

in NSSA, 144

in OSPF, 157, 283

in OSPF MIB, 218

in RIP, 278-279

All-subnets-broadcast addresses, 20

AllDRouters address, 95

AllSPFRouters address, 95

Analyzers

for debugging, 259-262

for interoperability testing, 58

integrated with IP datagram forwarding, 14

Application layer, 9

Application protocols, 5

APPN (Advanced Peer-to-Peer Networking) protocol, 49

APSCN (Arkansas Public School Computer Network) example, 228-230

Area border routers (ABRs), 122-123

definition of, 158-159

in MOSPF, 201

configuration of, 226

Area flooding scope, 132

Area IDs, 86

Areas. See also OSPF areas

in MOSPF, 199-203

in IS-IS, 301

Arkansas Public School Computer Network example, 228-230

ARPANET (Advanced Research Projects Agency network), 3, 33

routing algorithm in, 48-50

sequence bug in, 56, 77

synchronization in, 89

AS-external-LSAs, 75, 128-131, 284

in Database Overflow extensions, 145

with external-attributes-LSAs, 147-149

flooding, 131-132

with IBGP, 235

vs. Type-7-LSAs, 143

AS flooding scope, 131

AS numbers, 34

AS_PATH attribute, 295

AS paths in BGP, 286, 295

AS sequences, 290

AS sets, 235, 290

ASBR-summary-LSAs, 75, 131

ASBRs (Autonomous System boundary routers), 127-128, 133, 144

ASN.1 format, 215

ASs (Autonomous Systems), 32-34

in BGP, 147-149, 296-297

configuring, 213

routing protocols with, 146

ATOMIC_AGGREGATE attribute, 296

Authentication

cryptographic, 237-240

header fields for, 86

Automatic neighbor discovery, 106

Autonomous System boundary routers (ASBRs), 127-128, 133, 144

Autonomous Systems (ASs), 32-34

in BGP, 147-149, 296-297

configuring, 213

routing protocols with, 146

B

Backbone areas, 55, 125

Backup Designated Router field, 87

Backup Designated Routers

in database synchronization, 106-110

design rationale, 53-54

function of, 156-157

in NBMA subnets, 111-113

Bellman-Ford Vector algorithm, 36, 39

Best matches in routing tables, 28

BGP (Border Gateway Protocol), 34, 39, 284-292

with ASs, 147-149, 296-297

IBGP, 292-294

interactions with, 233-235

MIB for, 254

operational issues, 296-297

path attributes, 286, 294-296

Binary trees for routing tables, 30-31

Border Gateway Protocol. See BGP (Border Gateway Protocol)

Boundaries of OSPF areas, 157-158

Broadcast-and-prune multicast routing protocols, 310-312

Broadcast forwarding, 178-184

Broadcast IP addresses, 16

Broadcast storms, 10

Broadcast subnets, 104-105

abstraction of, 108-110

database synchronization in, 106-108

neighbor discovery and maintenance in, 105-106

problems in, 110-111

BSD UNIX, 277

Burgan, Jeff, 63

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

C

Cache hits with hash functions, 31

Caches

in multicast forwarding, 183

in MOSPF, 195-197

for routing tables, 31-32

Candidate lists, 97-98

CBQ (Class Based Queuing) algorithm, 13

CBT (Core Based Trees) routing protocol, 180, 315

CheckAge constant, 80

Checksums

in database synchronization, 94

in IP headers, 24-25

for LSAs, 78, 84, 159-160

in OSPF headers, 86

CIDR (Classless Inter-Domain Routing), 17, 67

for hierarchical routing, 121

for RIP, 277

CIDR blocks, 33

CIDR notation, 17

Cisco routers, 223-224

CIXs (Commercial Internet Exchanges), 34

Class Based Queuing (CBQ) algorithm, 13

Classes of IP addresses, 17, 19-21, 172-174

Classless Inter-Domain Routing (CIDR), 17, 67

for hierarchical routing, 121

for RIP, 277

Clients of Route Reflectors, 293

Clocks for OSPF aging, 160

Cloud subnets, 110

CLUSTER_LIST attribute, 293, 296

Clusters in BGP, 294

Coltun, Rob, 56, 60, 151, 263

Commercial Internet Exchanges (CIXs), 34

COMMUNITY attribute, 296

Compatibility

of Database Overflow, 146

of Demand Circuit, 142

when extending OSPF, 136

of external-attributes-LSA, 149

in MOSPF, 203-205

of NSSA areas, 145

of stub areas, 140

of TOS-based routing, 139

Complete Sequence Number PDUs (CSNPs), 302

Confederations in BGP, 292-294

Conferencing applications, 174

Configuration, OSPF, 222-225

difficulties in, 155-156

link costs, 225

OSPF areas, 226-227

timers, 227-228

Convergence

Distance Vector, 35-39

as OSPF design issue, 45

in routing, 27

Convergence time, 27

Core Based Trees (CBT) routing protocol, 180, 315

Core routers, 180. See also Rendezvous Points

Costs

link, 152-153, 225

in MOSPF, 163-166

Counting to infinity, 37-39, 298

Cryptographic authentication, 237-240

CSNPs (Comple

Errata

The following corrections were made for the second printing of this title.

Preface

• Page xvii: The word forthcoming was added (midpage) to indicate that the second volume of this book is not yet available.

Chapter 4

• Page 84: In line 4, the word column was changed to row.
• Page 93: In Figure 4.8 the arrow next to the label Time T3+ (at the bottom left of the figure) was changed so that the arrowhead points toward 10.1.1.3 instead of away from 10.1.1.3.

Chapter 6

• Page 125: Five lines from the bottom, the word best was changed to better.

The index was also revised for this reprint.

The following corrections will be made in the third printing of this title.

Chapter 9

• Page 172: In the second sentence of the last paragraph,the network segment 195.5.2/24 should be 192.5.2/24.