Networking Explained Second Edition Michael Gallo

Networking Explained Second Edition Michael
Gallo pdf download
https://ebookgate.com/product/networking-explained-second-
edition-michael-gallo/
Get Instant Ebook Downloads – Browse at https://ebookgate.com

Get Your Digital Files Instantly: PDF, ePub, MOBI and More
Quick Digital Downloads: PDF, ePub, MOBI and Other Formats
Raspberry Pi Networking Cookbook Second Edition Golden
https://ebookgate.com/product/raspberry-pi-networking-cookbook-
second-edition-golden/
Hands On Networking Fundamentals 2nd Edition Michael
Palmer
https://ebookgate.com/product/hands-on-networking-
fundamentals-2nd-edition-michael-palmer/
Wireless Networking with Microsoft Windows Vista 1st
Edition Michael Miller
https://ebookgate.com/product/wireless-networking-with-microsoft-
windows-vista-1st-edition-michael-miller/
The European Union Explained Second Edition
Institutions Actors Global Impact Andreas Staab
https://ebookgate.com/product/the-european-union-explained-
second-edition-institutions-actors-global-impact-andreas-staab/

Handbook Of Geriatric Assessment 4th Edition Joseph J.
Gallo
https://ebookgate.com/product/handbook-of-geriatric-
assessment-4th-edition-joseph-j-gallo/
Quantitative Ecotoxicology Second Edition Michael C.
Newman
https://ebookgate.com/product/quantitative-ecotoxicology-second-
edition-michael-c-newman/
The Linux Cookbook Second Edition Michael Stutz
https://ebookgate.com/product/the-linux-cookbook-second-edition-
michael-stutz/
3G HSPA and FDD versus TDD Networking Smart Antennas
and Adaptive Modulation Second Edition L. Hanzo
https://ebookgate.com/product/3g-hspa-and-fdd-versus-tdd-
networking-smart-antennas-and-adaptive-modulation-second-edition-
l-hanzo/
Measuring Workplace Performance Second Edition Michael
J. O'Neill
https://ebookgate.com/product/measuring-workplace-performance-
second-edition-michael-j-oneill/

Networking Explained,
Second Edition

This Page Intentionally Left Blank

Networking Explained,
Second Edition
MICHAEL A. GALLO, PH.D.
Florida Institute of Technology
WILLIAM M. HANCOCK, PH.D.
Exodus Communications Incorporated
Boston • Oxford • Auckland • Johannesburg • Melbourne • New Delhi

Copyright © 2002 by Butterworth–Heinemann
A member of the Reed Elsevier group
All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of the publisher.
Recognizing the importance of preserving what has been written, Butterworth–Heinemann
prints its books on acid-free paper whenever possible.
Gallo, Networking Explained, Second Edition
ISBN 1-55558-252-4
The publisher offers special discounts on bulk orders of this book.
For information, please contact:
Manager of Special Sales
Butterworth–Heinemann
225 Wildwood Avenue
Woburn, MA 01801–2041
Tel: 781-904-2500
Fax: 781-904-2620
For information on all Butterworth–Heinemann publications available, contact our
World Wide Web home page at: http://www.bh.com
10 9 8 7 6 5 4 3 2 1
Printed in the United States of America

To Dee Dee Pannell . . .
a terriﬁc computer systems and network supervisor,
and a remarkable human being

Contents
PREFACE xiii
1 FUNDAMENTAL CONCEPTS OF COMPUTER NETWORKS AND
NETWORKING 1
Concepts and Related Issues of Computer Networking 1
Deﬁnition and Components of Computer Networks 3
Network Protocols 5
Network Media 6
Applications Versus Application Protocols 6
Interoperability 9
Internet, internet (Internetwork), Intranet, Extranet 9
LAN, MAN, WAN, PAN, GAN, SAN 10
Decentralized Versus Centralized Versus Distributed Systems 14
Client/Server, Peer-to-Peer, Servent, and WWW Network Models 15
Network Appliances 20
Serial and Parallel Communications 21
Synchronous, Asynchronous, and Isochronous Communications 22
Simplex, Half-Duplex, and Full-Duplex Communications 24
Network Standards 25
2 NETWORK TOPOLOGIES, ARCHITECTURES, AND THE OSI MODEL 31
Concept of Network Topologies 31
Point-to-Point Networks: Star, Loop, Tree 32
Broadcast Networks: Bus, Ring, Satellite 35
Multidrop Networks 38
Circuit- and Packet-Switched Networks 38
Network Architectures 42
The OSI Model 42
Connection-Oriented and Connectionless Services 48
The TCP/IP Model 50
3 THE INTERNET AND TCP/IP 55
Deﬁnition of the Internet 56
Internet History 56
vii

viii Networking Explained, Second Edition
The Current Commodity Internet 62
Other Internet Initiatives: vBNS/vBNS+, Internet2, Next Generation Internet 64
Internet Administration, Governance, and Standards 66
History of TCP/IP 68
TCP/IP Application Layer Protocols: SMTP, MIME, POP, IMAP, TELNET, FTP,
and HTTP 71
TCP/IP Transport Layer Protocols: UDP and TCP 101
TCP/IP Network Layer Protocol: IP 113
IPv4 Addresses and Subnetting 118
IP Address Assignments 123
IP Name Resolution 126
IPv6 129
Internet Services, Resources, and Security 137
4 PHYSICAL LAYER CONCEPTS 139
Physical Layer Issues 139
Analog Versus Digital Communications 142
Bandwidth Versus Throughput and Data Rate Versus Baud Rate 147
Noise 148
Shannon’s Limit 149
Multiplexers and Multiplexing Strategies 150
Switching Strategies 155
Physical and Electrical Characteristics of Wire 156
UTP, STP, and IBM Cable 161
Coaxial Cable 164
Fiber-Optic Cable 166
Wireless Media 170
Satellite Communications 176
5 DATA LINK LAYER CONCEPTS AND IEEE STANDARDS 181
Data Link Layer Overview and IEEE’s Perspective 181
Framing 184
Ethernet/802.3 Frames 185
Flow Control and Flow Control Protocols 187
Error Control 196
MAC Sublayer 202
Random Access and Token Passing Protocols 204
Data Prioritization and Quality of Service (QoS) 208
6 NETWORK HARDWARE COMPONENTS (LAYERS 1 AND 2) 213
Connectors 213
Transceivers 216
Repeaters 219
Media Converters 222
Network Interface and PC Cards 222
Bridges 226
Switches 231

Contents ix
7 WANS, INTERNETWORKING, AND NETWORK LAYER CONCEPTS
AND COMPONENTS 237
The Concept of Internetworking 237
WAN Circuits 241
SONET 248
Layer 3 Concepts and Issues 250
Router Protocols and Routing Algorithms 252
Routing Versus Switching 268
Virtual Private Networks (VPNs) 272
Multiprotocol Label Switching (MPLS) 283
8 ETHERNET, ETHERNET, AND MORE ETHERNET 287
History of Ethernet 1
Ethernet Versus IEEE 802.3 288
The Initial IEEE 802.3 Protocol: 10 Mbps Ethernet 290
10-Mbps Ethernet Performance Issues: Network Diameter and Collision
Domain 293
Partitioning, Switched Ethernet, and Virtual LANs 303
100 Mbps Ethernet: Fast Ethernet and 100VG-AnyLAN 313
1 Gigabit Ethernet 323
10 Gigabit Ethernet 332
IsoEthernet 335
9 TOKEN RING 337
Deﬁnition and Operation 337
Frame Formats 342
Priority Scheduling 342
Monitor Stations 344
Token Ring Versus Token Bus 345
Advantages and Disadvantages of Token Ring Networks 346
Switched, Dedicated, and Full-Duplex Token Ring 346
High-Speed Token Ring 348
Token Ring’s Future 349
10 FIBER DISTRIBUTED DATA INTERFACE (FDDI) 351
General Information 351
Data Link Layer Issues 356
Operation and Conﬁguration Issues 357
FDDI and Ethernet/802.3 361
FDDI and ATM 363
CDDI 363
Future of FDDI 364
11 INTEGRATED SERVICES DIGITAL NETWORK (ISDN) 365
History of ISDN 365

x Networking Explained, Second Edition
General Overview and Components 368
Channel Types 372
BRIs, PRIs, and SPIDs 373
Line Sets and Feature Sets 375
ISDN Protocols 378
AO/DI and B-ISDN 380
Alternative Implementation Strategies 381
12 FRAME RELAY 383
Frame Relay Overview 383
Frame Relay’s Physical Layer: Virtual Circuits and Committed Interface Rates
(CIRs) 384
Technical Aspects and Operation of Frame Relay 394
Frame Relay’s Data Link Layer 395
Voice Over Frame Relay 402
Frame Relay Versus ATM, SMDS, and Gigabit Ethernet 403
Frame Relay in the News 404
13 SWITCHED MULTIMEGABIT DATA SERVICE (SMDS) 407
SMDS Overview 407
SMDS and Local and Inter-Exchange Carriers 408
The DQDB and SMDS Protocols 410
SMDS Addressing 416
SMDS Applications 418
SMDS Versus Frame Relay and ATM 418
Current Status and Future of SMDS 420
14 ASYNCHRONOUS TRANSFER MODE (ATM) 423
Deﬁnition and History of ATM 423
General Concepts and Operation of ATM 424
ATM Interfaces and the Anchorage Accord 426
ATM Components and Addressing 429
ATM Cells, Switches, and Virtual Connections 430
ATM Adaptation Layer (AAL) 435
Data Types Insights 437
ATM Versus Gigabit Ethernet 438
ATM in LAN Environments 439
ATM, Frame Relay, and SONET 441
15 DIALUP AND HOME NETWORKING 443
Dialup Networking Concepts and Issues 443
Modem Concepts: Analog and 56K Modems 446
xDSL Connections 452
Cable Modem Connections 458
Home-Based Internet Connections 462
Home-Based LANs: Concepts and Issues 466
16 NETWORK SECURITY ISSUES 471
Network Security Overview 471

Contents xi
Threat Assessment and Risk Analysis 472
Social Engineering, Denial of Service, and Applications 477
Network Security Preparations and Measures 480
Firewalls 483
Cryptography and Encryption: DES, AES, RSA, PGP 490
Authentication: Digital Certiﬁcates, Smart Cards, and Kerberos 501
Internet Security and Virtual Private Networks (VPNs) 503
17 NETWORK CONVERGENCE 505
Network Convergence Overview 505
Impact of Convergence on Network Media 509
Network Convergence and Multimedia 516
Impact of Convergence on Businesses 521
Network Convergence at Home 527
Network Convergence and Voice over IP (VoIP) 528
18 WIRELESS NETWORKING 537
Wireless Communications Overview and History 537
Wireless Data Transmission Methods 539
Cellular Telephone Networks 540
Paging Networks 550
Wireless Data Networks 552
Bluetooth Concepts Versus Wireless LANs 558
Advantages and Disadvantages of Wireless Communications 559
The Future of Wireless Communications 560
APPENDIX A: VENDOR ETHERNET/802.3 PREFIXES 563
APPENDIX B: USING PARITY FOR SINGLE-BIT ERROR CORRECTION 565
APPENDIX C: GUIDELINES FOR INSTALLING UTP CABLE 567
APPENDIX D: NETWORK DESIGN AND ANALYSIS GUIDELINES;
NETWORK POLITICS 569
APPENDIX E: X.25 575
GLOSSARY 581
BIBLIOGRAPHY 631
INDEX 653

Preface
Welcome to the second edition of Networking Explained. In preparing this new edition,
our goal was threefold: (1) to update the book so that it reflects, as best as possible, the
current state of computer networking; (2) to enhance the discussion of key concepts so that
you will have a greater understanding of how certain protocols work; and (3) to maintain
the elements and features of the previous edition that contributed to its success. We believe
that we have succeeded in accomplishing our goals.
Key Content Changes
Chapter 1: Fundamental Concepts of Computer Networks and Networking now contains
an overview of networking and a discussion of the main issues involved in networking. We
also provide additional information about storage area networks (SANs), discuss the
World Wide Web as a networking communications model, and include information about
file-sharing programs such as Napster.
Chapter 2: Network Topologies, Architectures, and the OSI Model includes new historical
information about the development of the OSI model. We also rewrote the material on
packet-switched networks to improve the clarity of the presentation.
Chapter 3: The Internet and TCP/IP was extensively rewritten and now contains informa-
tion about TCP/IP’s application layer protocols, including SMTP, MIME, POP, IMAP,
TELNET, FTP, and HTTP. We also extended our discussion on TCP/IP’s two transport
layer protocols, UDP and TCP, and provided more detailed information about subnets and
subnet masking.
Chapter 4: Physical Layer Concepts contains updated information on UTP cables and
wireless media. We also revised the discussion on IEEE 802.11 and included new material
about HomeRF.
Chapter 5: Data Link Layer Concepts and IEEE Standards contains new material on var-
ious flow control protocols, including stop-and-wait, sliding window, and go-back-n, and
includes a new boxed text item that demonstrates how CRC checksums are calculated.
Chapter 7: Internetworks and Network Layer Concepts and Components contains a new
boxed text item that discusses Nyquist’s theorem and includes additional information about
xiii

xiv Networking Explained, Second Edition
the pulse code-modulation (PCM) process. We also extended our coverage on the tunneling
concept and included new information about multiprotocol label switching (MPLS) and
virtual private networks (VPNs). Finally, we moved information about VPN security pro-
tocols from Chapter 16 and placed it in this chapter.
Chapter 8: Ethernet, Ethernet, and More Ethernet contains updated information on Giga-
bit Ethernet and includes new material on 10 Gigabit Ethernet.
Chapter 15: Dialup and Home Networking contains expanded coverage on cable modem
service and the various digital subscriber line (DSL) protocols. We also updated our dis-
cussion on modem protocols to include V.44 and V.92, and rewrote the material dealing
with establishing a home network.
Chapter 16: Network Security Issues was rewritten to improve topic clarity and organiza-
tion. The chapter was also updated and now contains new information about denial of ser-
vice (DoS) attacks, public key infrastructure (PKI), the advanced encryption standard
(AES), and a new boxed text item that demonstrates the RSA algorithm.
New Chapters and Appendix
In addition to the content changes cited above, we also wrote two new chapters and one
new appendix.
Chapter 17: Network Convergence contains a thorough discussion of the convergence con-
cept from several perspectives, including its impact on networks, network media, businesses,
and the home. The chapter also contains several examples of convergence applications,
including Enum and the Bluetooth project, and provides information about multimedia net-
working from a convergence perspective as well as information about voice over IP (VoIP).
Chapter 18: Wireless Networking extends the wireless communications discussion from
Chapter 4 and includes information about cellular telephone networks and protocols, pag-
ing networks, and wireless data networks. This chapter also expands Chapter 4’s discussion
on IEEE 802.11b and 802.11a and compares wireless LANs (WLANs) to the Bluetooth
technology. Finally, the chapter concludes with a brief discussion of the future of wireless
communications.
Appendix E: X.25 provides an overview of the X.25 standard for data communications.
Retained Features
One of the most redeeming features of the ﬁrst edition of this book was its question and
answer format, which we modeled after a terriﬁc book on statistics by James Brewer:
Everything You Always Wanted To Know about Statistics But Didn’t Know How to Ask.
This question-answer presentation was retained for the second edition. As was the case
with the previous edition, we structured the question-answer format to emulate a conversa-
tion between a networking professional and the reader. The questions are representative of
those asked by individuals who are interested in computer networks and those who wish to
gain additional understanding of the subject. The answers are intended to give the reader a
broad foundation in networking concepts.

Preface xv
A second retained feature is the collection of informative figures, tables, and boxed text
items. We have included 32 new figures, 26 new tables, and 6 new boxed text items. In all
total, this new edition contains nearly 300 pieces of art. The figures graphically represent
key concepts, the tables summarize salient information, and the boxed text items explore
topics in greater detail without getting in the way of the main discussion.
A third feature we retained in this edition is the glossary. A comprehensive 700-term
glossary complements the book’s nucleus.
A fourth feature we retained is the overall structure of the chapters. As was the case
with the first edition, all the chapters are structured logically to follow each other and build
on previous learned knowledge. At the same time, however, the chapters may be read in
any order because we wrote them to also be independent of each other. Throughout every
chapter, key terms or concepts that were either presented in an earlier chapter or discussed
in a later chapter, are explained in the current context with either a forward or backward
reference that directs the reader to the appropriate chapter where the term or concept is
discussed more completely. These cross-references enable readers to take a break from the
current discussion so they can either refresh their knowledge of previously presented
material or jump ahead to gain further insight into a specific topic. All chapters are also
formatted consistently. Each chapter begins with an introduction that includes a bullet list
of the major topics discussed. Topics also contain corresponding question numbers,
thereby making it easy to quickly locate information about a particular topic. Questions
are categorized hierarchically by subject so that busy readers who seek answers to specific
questions can find them easily. End-of-chapter commentaries are also provided. These
commentaries consist of transitional material that identifies other chapters in the book
containing additional information related to the current discussion.
Finally, in writing this new edition, we were careful to maintain our overall philoso-
phy on which the first edition was grounded: to provide current and future network manag-
ers and administrators with an accurate and easy to read book on data communications
standards and emerging networking technologies that is accessible by readers of all back-
grounds. Given its format and level of coverage, we equate the book to a snorkeling
adventure. We primarily stay at the surface to examine the features, attributes, technical
issues, and concepts of networks. Occasionally, we hold our breath and dive under to
explore a particular concept in more detail. We do not, however, discuss any particular
topic in too much depth. Such an undertaking is better left for more technical books,
which are analogous to a scuba diving expedition.
What This Book Is and Isn’t About
After reading this book, you will accrue a greater understanding of and appreciation for
networks and networking. This book will help you understand basic networking terminol-
ogy, components, applications, protocols, architectures, standards, and implementation
strategies. The reader is cautioned that this is not a “how-to” book. We do not provide spe-
cific information relative to network management or configurations. Thus, the material
contained here will not help you perform such tasks as setting up a domain name server,
configuring a network printer, or installing or managing an office network. However, your
knowledge, appreciation, comprehension, and awareness of the concepts involved in such
activities will be more acute after reading this book.

xvi Networking Explained, Second Edition
Secondary Audiences
In addition to its primary audience, this book lends itself to three secondary audiences.
First, the book is appropriate for computer-networking hobbyists or nonprofessionals who
desire to gain a working knowledge of the vocabulary, concepts, and current technologies
related to networking. Second, the book is suitable for individuals who have a working
knowledge of networks, but lack an understanding of the fundamental concepts and theo-
retical underpinnings of networks. Finally, the book can be used as a companion resource
in an academic setting.
Acknowledgments
Many people contributed considerably, either directly or indirectly, in the preparation of
this material. It is therefore justiﬁable that these individuals be acknowledged. First of all,
we are grateful to the authors of the articles, books, RFCs, and other reference material
listed in the Bibliography. These publications served as invaluable resources for conﬁrm-
ing that our illustrations and material were accurate, complete, and up-to-date. Next, sev-
eral people reviewed various aspects of the material in one form or another. They include
Prasad Aloni of Drexel University, Ron Fulle of Rochester Institute of Technology, Arnold
Meltzer of George Washington University, Krishna Sivalingam of Washington State Uni-
versity, Eugene Styer of Eastern Kentucky University, and Michael Whitman of Kennesaw
State University. The feedback we received from these individuals was invaluable and we
are grateful for their assistance. It is also with pleasure that we acknowledge and thank
John Rhoton, who is a wireless technology consultant at Compaq Global Services. John
provided us with immense support and timely assistance in the preparation of the wireless
networking chapter. We are also extremely appreciative of the support and guidance pro-
vided by our editor, Pam Chester, and by Karen Forster, production director, for keeping
this project on schedule. Thanks Pam, and thanks Karen. It was a pleasure working with
you both. A sincere note of appreciation is also extended to Kristin Merz, who provided
tremendous assistance with the artwork. Finally, on a personal note, one of us is indebted
to Jane for always believing that “this will be the last one.” Without her support, patience,
understanding, and love, this project would never have been completed. Thank you
sweetie. I love you.

Chapter 1
Fundamental Concepts of Computer
Networks and Networking
In this chapter, we provide an overview of the various topics and concepts surrounding
computer networks and networking. We begin by discussing some of the key issues related
to networking. We follow this discussion with a formal definition of computer networks
along with a description of the various terms related to this definition. As part of this dis-
cussion, we also introduce basic networking vocabulary as well as several key networking
concepts. Much of the material we present here is done so from a general perspective and
is elaborated in subsequent chapters. Understanding this material is important because it
serves as the underlying foundation of the book. An outline of the terms and concepts we
define and discuss follows:
• Concepts and Related Issues of Computer Networking (Questions 1–2)
• Definition and Components of Computer Networks (Questions 3–9)
• Network Protocols (Questions 10–12)
• Network Media (Questions 13–14)
• Applications vs. Application Protocols (Questions 15–19)
• Interoperability (Questions 20–21)
• Internet, internet (Internetwork), Intranet, Extranet (Questions 22–24)
• LAN, MAN, WAN, PAN, GAN, SAN (Questions 25–27)
• Decentralized vs. Centralized vs. Distributed Systems (Questions 28–30)
• Client/Server, Peer-to-Peer, Servent, and WWW Network Models (Questions 31–41)
• Network Appliances (Questions 42–44)
• Serial and Parallel Communications (Questions 45–48)
• Synchronous, Asynchronous, and Isochronous Communications (Questions 49–53)
• Simplex, Half-Duplex, and Full-Duplex Communications (Questions 54–55)
• Network Standards (Questions 56–64)
1

2 Networking Explained, Second Edition
1. The title of this book is Networking Explained. Please explain the term “network-
ing” and distinguish it from the term “network.”
Before we respond directly, consider for a moment what some people do when they
attend a professional conference or meeting—they usually engage in some form of net-
working. That is, they seek out people in an attempt to establish contacts so they can
exchange or share information. Collectively, these individuals form a human network.
(Have you ever heard someone say, “She has her own network of friends”?) Things are not
that much different in the computer world. Speciﬁcally, the term “networking” refers to
the concept of connecting a group of systems for the expressed purpose of sharing infor-
mation. The systems that are connected form a network.
2. What are some of the issues involved in networking?
Computer networking involves many issues.A brief overview of some key ones follows.
• Communication methodology and protocols. In order for network members to
understand each other’s communication, they must first know and follow specific
communication rules. In the world of computer networks, communication meth-
odology and protocols provide these rules of communication. We will discuss
the concept of communication methodology and protocols later in this chapter.
• Topology and design. A key issue of networking involves how to connect the
individual systems that comprise a network. Thus, the topology and design cri-
teria are important topics in any discussion of computer networks. Topology
and design are the focus of Chapter 2.
• Addressing. Addressing describes how systems are assigned addresses and how
systems locate each other within a network. It involves assigning a network
node a unique address so that other systems or devices are able to locate it. It is
similar to a house’s street address—knowing the street helps to find where we
want to go, but having the house number means we will eventually find the
exact location of our destination. Addressing is first introduced in this chapter
and is then revisited in nearly every subsequent chapter.
• Routing. Related to addressing is the concept of routing, which describes the
manner in which data are transferred from one system to another across a net-
work. Routing involves determining the path a message takes as it travels
between source and destination nodes. Routing is usually performed by special
dedicated hardware units called routers. The “best” route packets should take is
a function of a specific criterion (or criteria), which is called a metric. Some of
the issues involved in routing include the path a message follows, how a path is
determined, and various policy issues that govern the control of a path. Routing
is discussed in Chapter 3 and then again in more detail in Chapter 7.
• Reliability. Reliability refers to data integrity, which has to do with ensuring that
the data received are identical to the data transmitted. Computer networks are
nontrivial systems and are not infallible. In fact, most networks are highly frag-
ile and easily disrupted. Thus, it is important that they be designed with the
capability to resolve errors. A common error-control strategy is to provide

Chapter 1: Fundamental Concepts of Computer Networks and Networking 3
enough information in the transmitted data to enable the destination node to
detect an error. This is called error detection. Once a destination node has
detected an error, it can then do one of two things: request a retransmission from
the sending node or determine what the correct data should be and change them
accordingly. Both methods are forms of error correction. We call the first error
correction through retransmission and the second autonomous error correc-
tion. Note that error correction implies error detection. Autonomous error
correction is very expensive to implement. Hence, most networks today are
designed with error-detection capabilities, and error correction is done by hav-
ing the destination node request the sending node to retransmit the data in
question. The concept of and issues related to reliability are first discussed in
Chapter 2 and then re-examined from different perspectives in nearly every sub-
sequent chapter. The two most popular error-detection strategies—parity check
and cyclic redundancy check (CRC)—are discussed in Chapter 5.
• Interoperability. In any free enterprise system, consumers want the freedom to
determine which products to purchase. Unfortunately, this can sometimes lead
to disastrous results when purchasing networking products from different ven-
dors. Thus, it is important that interoperability issues are considered. This refers
to the degree in which software and hardware products developed by different
vendors are able to communicate successfully with each other over a network.
The topic of interoperability is first introduced later in this chapter and then
highlighted in most subsequent chapters, particularly Chapter 7.
• Security. Whenever we engage in any networking activity, we would like some
assurances that our communications are secure. Thus, security issues are of par-
amount concern and pertain to the safekeeping or protection of all components
of a network. Network security refers to the proper safeguarding of everything
associated with a network. This includes data, media, and equipment. Security
involves administrative functions such as threat assessment (“What do you have
and who would want it?”), technical tools and facilities such as cryptographic
products, and network access control products such as firewalls, which are spe-
cial hardware devices that serve to protect an internal network from the outside
world. Security also involves making certain that network resources are used in
accordance with a prescribed policy and only by people who are authorized to
use these resources. Chapter 16 is dedicated to network security issues.
• Standards. The development and implementation of networks rely on the estab-
lishment of specific rules and regulations to be followed. This is the role of
networking standards. We introduce the concept of network standards later in
thischapterandthenincludeadiscussionofitsimpactineverychapterthereafter.
3. OK. Now that I have some idea of what networking is, what exactly is a computer
network?
A computer network is a collection of computers and other devices that use a common
network protocol to share resources with each other over a network medium.

4. That’s a mouthful. Can you break this down for me, please? There are several
terms I do not understand.
Sure. Where do you want to start?
5. First of all, you say a computer network is a collection of computers and other
devices. What other devices?
In our definition we use the term device to represent any entity that is connected to a
network. Such entities may be terminals, printers, computers, or special network-related
hardware units such as communication servers, repeaters, bridges, switches, routers, hand-
held units such as the PalmPilot, and various other devices, most of which are discussed in
detail in later chapters. Devices can be either local or remote. The device originating com-
munication across a network is called the local device or sending device, and any device
within the network that is accessed from this local device is called the remote device or
receiving device. In a telephone network, the telephone handsets we all use are devices. So
is the interconnecting hardware at the phone company that allows handsets to talk to each
other. A network requires many diverse types of devices in order to work.
6. So device is a generic term.
Correct.
7. What about node? I see this term used frequently.
The word node is commonly used interchangeably with device; both terms refer to any
equipment that can access a network. You also frequently see the term station as well as
appliance used in the literature. Both are synonymous with device and node. All four
terms refer to any equipment that can access a network.
8. If device and node are generic terms that refer to any entity connected to a net-
work, then why do you use the term computers in your definition?
We prefer to distinguish between devices and computers. As network devices, comput-
ers are called hosts (or servers) or workstations (also called desktops or clients). This ter-
minology refers to computer systems that have their own operating systems (such as
Windows). Thus, a workstation might be a personal computer such as an Apple Macintosh
or a Windows-based machine such as a Compaq, Dell, or Gateway PC; a graphic worksta-
tion such as those manufactured by Sun Microsystems, Silicon Graphics, IBM, or Hewlett
Packard; a super-minicomputer such as an IBM AS/400 system; a super-microcomputer
like Compaq’s Alpha; or perhaps a mainframe such as an IBM ES-9000.
9. With so many different types of nodes connected to a network, what mechanism do
they use to find each other?
You are referring to the concept of “addressing,” which was mentioned earlier in our
response to Question 2 and is discussed in more detail in later chapters. For now, though,
suffice it to say that a network node is assigned a unique address that allows other systems

or devices to locate them. It is similar to a house’s street address—knowing the street
helps to find where you want to go, but having the house number means you will eventu-
ally find the exact location of your destination. Another analogy is the phone system you
use all the time. Each phone (a node) has an area code and a number (an address). The area
code provides information about the node’s location within the global telephone network,
and the telephone number is the device’s specific identification number within that locale.
Systems and call “routers” at the phone company are programmed to provide information
to other network devices to get the call from your phone handset to the proper destination
(the phone number you are calling).
10. OK. Next is network protocol. What does this mean?
Let’s take this in two parts, starting with the term protocol. From a general perspective,
a protocol is an accepted or established set of procedures, rules, or formal specifications
governing specific behavior or language. For example, when eating in a fancy, expensive
restaurant, patrons are usually required to observe a specific dress protocol (e.g., men typ-
ically have to wear a jacket and tie). Other restaurants, such as the ones frequented by the
authors, may have different dress protocols: no shoes, no shirt, no service. If you were to
meet the Queen of England, once again you would need to observe a certain protocol.
When applied to networking and data communications, a network protocol is a formal
specification that defines how nodes are to “behave” or communicate with each other.
Among other things, network protocols define how data are to be formatted, how data
integrity is to be maintained, and how data are transmitted and received between nodes. In
short, a network protocol specifies the vocabulary and rules of data communication.
11. Can you give me an example of a network protocol?
Yes. A good example consists of the individual protocols that are part of the TCP/IP
suite. TCP/IP stands for “Transmission Control Protocol/Internet Protocol,” which serves
as the basis of the Internet. (See Chapters 2 and 3 for more information about TCP/IP and
the Internet.) Although TCP/IP specifies two particular protocols (TCP and IP), it is also
used to name the set of protocols that includes not only TCP and IP, but also many others.
This set of protocols is called the TCP/IP suite. (When a group of related and interoperat-
ing protocols are put together in a package on a system, we call it a suite.)
Another protocol that is part of the TCP/IP suite is FTP, or “File Transfer Protocol,”
which specifies how to do file transfers. HTTP, the Hypertext Transport Protocol, is used
for the World Wide Web (WWW), and defines how servers need to transfer documents
(Web pages) to clients (Web browsers). Three protocols used for electronic mail (e-mail)
with which you might already be familiar are the Post Office Protocol (POP), the Simple
Mail Transfer Protocol (SMTP), and the Internet Mail Access Protocol (IMAP). All of the
foregoing are network protocols that are also part of the TCP/IP suite. Today’s networks
employ a great multitude of protocols, ranging from very simple to quite complex. Proto-
cols are the glue that binds together computer networks because they define how specific
operations are to be performed. We will provide detailed examples of how some of these
protocols actually work in later chapters of the book.

12. Are there other sets (or suites, as you called them) of network protocols?
Yes. One you might be familiar with is AppleTalk, which is a network protocol suite used
by Apple Computer, Inc., originally in its line of Macintosh computers and now available
in many other operating systems. Another example is the set of protocols that are part of
Microsoft Corporation’s Windows NT operating systems. Sometimes, computer networks
are named by their protocols. For example, a network that consists of devices supporting
AppleTalk is referred to as an AppleTalk network. Similarly, a TCP/IP network implies a
set of devices linked together that uses the TCP/IP suite as its set of rules for communication.
13. OK. Getting back to your deﬁnition of computer networks, I now understand
what you mean by devices and network protocol. One last thing you mentioned
that I am unclear about is “network medium.”
Besides protocols, nodes have to be connected to each other in some manner to share
resources or receive services via a network. The physical environment used to connect
members of a network is referred to as a medium (the plural of which is media). Network
media come in two broad categories: cable and wireless. Examples of cable include
twisted-pair, coaxial, and ﬁber-optic cable. Examples of wireless include radio waves
(including microwave and satellite communication) and infrared radiation. Network media
will be discussed in more detail in Chapter 4.
14. So, computer networks require media and protocols because without a link,
resources cannot be shared, and without protocols, communication cannot be
understood even if a link exists. Is this correct?
Right. Network media provide an environment in which communication can take
place, while protocols are necessary to ensure that communications are understood. This is
similar to a telephone conversation between one person who speaks only Italian and
another who speaks only Russian. If a telephone circuit (i.e., network link) for this conver-
sation is not available, then these two individuals cannot speak to each other (i.e., they
cannot share resources). Given a circuit, the two individuals can now speak and hear each
other’s voices (i.e., transmission), but communication cannot take place because neither
individual is capable of understanding the other’s message—they speak different lan-
guages. Networking happens when a common wiring infrastructure connects nodes that
share a common protocol infrastructure—just like human communication.
15. Regardless of the protocol used, don’t most computer networks support similar
network-related applications like e-mail?
Yes they do. Although the function of these applications across different networks is
similar, the manner in which they are implemented is protocol-dependent. For example,
e-mail messages can be exchanged between hosts connected to a TCP/IP network because
they speak the same language. Similarly, e-mail messages can be exchanged between hosts
of a Windows NT network because, once again, they speak the same language. However,
e-mail messages cannot be exchanged directly between a host connected to a TCP/IP net-
work and a host connected to a Windows NT network because they may use different appli-

cation protocols for electronic mail. Consequently, although different networks might be
functionally equivalent in that they support similar applications (e.g., TCP/IP and NT net-
works both support electronic mail), the manner in which these functions are implemented
is not the same. As an example, UNIX systems support the e-mail protocol we previously
mentioned, called SMTP, which is a component of the TCP/IP suite of protocols. NT sys-
tems support a different e-mail system called Exchange. By default, they cannot interoper-
ate directly even though they are both connected to same network. However, Exchange can
be configured to communicate not only “native” with its own protocols, but also to simul-
taneously support TCP/IP and SMTP so that UNIX and NT users can exchange e-mail.
A network may have many protocols and many applications. Not all of them necessar-
ily talk directly with each other. Software functions called gateways (explained later)
allow conversion (like a linguistic translation) between some application protocols. In
other areas the problems are solved simply by supporting more than one application proto-
col at the same time. While this sounds complex, usually a little care in the planning cycle
makes everything work well when activated for use.
16. So what you are saying then is that there is a difference between an application
like e-mail and the protocol that defines it.
Exactly! To help understand this better, consider the hierarchy in Figure 1.1. At the
root layer we have a network protocol suite (TCP/IP). The next layer shows three network
applications that are part of this suite (e-mail, file transfer, and virtual terminal). The third
layer contains the protocols that define these applications (SMTP and POP for e-mail, FTP
for file transfer, and TELNET for virtual terminal). The last layer identifies a specific pro-
gram that users can use for these applications. A similar tree diagram can be drawn for NT,
or any other protocol suite. In short: A network protocol suite provides the specifications
for network applications such as e-mail. These applications have specific protocols that
define how the application is to be implemented on the network. The application protocols
also include specific user programs that we use to interact with the application.
FIGURE 1.1 A protocol suite contains specific network applications (e.g., e-mail), which in turn
are defined by specific application protocols (e.g., SMTP). These application protocols are part of
specific application programs (e.g., UNIX mail) that provide a user with an interface to interact with
an application. The application protocol also defines the manner in which an application is to be im-
plemented between two hosts connected to a network.
UNIX mail/mailx Eudora Fetch (Macintosh) NCSA Telnet
Application Program
SMTP POP FTP TELNET
Application Protocol
Electronic Mail File Transfer Virtual Terminal
General Function
TCP/IP
Protocol Suite

17. Can you give me an example to make this a little more concrete?
Sure. Let’s use e-mail as our example. E-mail is a network application that involves
creating, sending, receiving, and storing messages electronically. These activities are per-
formed by using a “mail program,” which provides a utility for users to (among others)
compose, read, save, delete, and forward e-mail messages. This mail program is an appli-
cation program that resides on a host. A mail program is also concerned with issues such as
how a host accepts or rejects mail, how mail is stored on a system, how a user is notified of
the arrival of new mail messages, and so forth. A mail program does not, however, manage
the network exchange of e-mail messages between two hosts. Instead, the method by which
e-mail is transferred from one host to another is handled by an e-mail application protocol.
Examples include SMTP, IMAP, and POP, which are part of the TCP/IP protocol suite.
Other network applications are similar to e-mail. They consist of an application pro-
gram that provides the user with an interface to interact with the application, and they con-
tain a related application protocol that defines the manner in which an application
communicates over a network. Thus, file transfer programs provide users with an interface
for copying files to or from a remote host, and virtual terminal programs provide users
with an interface for establishing a login on a remote host. These applications also have
corresponding protocols (FTP and TELNET, respectively) that define the rules local and
remote hosts must follow to run the application across the network.
18. Is there a one-to-one correspondence between an application program and net-
work application?
No. Some network applications support more than one application protocol. For exam-
ple, the public domain package NCSA Telnet, supports both the virtual terminal protocol
(TELNET) and the file transfer protocol (FTP). As another example, consider the software
product from Netscape Communications, Netscape Communicator. This product supports
several protocols including those for network news (NNTP), e-mail (SMTP, POP, and
IMAP4), the World Wide Web (HTTP), and file transfers (FTP).
19. I can understand why nodes have to use the same protocol, but do users also have
to use the same program to communicate with one another? For example, does
everyone I send e-mail to have to use the same e-mail application I use?
No, not at all. Remember, an application program simply provides the user with an
interface to interact with the application. Behind this application is an associated protocol,
which is transparent to the user. As long as the application’s corresponding protocol under-
stands another application’s protocol everything should work out. Thus, you might use
Eudora as your e-mail package, but someone else might choose to use the e-mail program
that is part of Netscape Communicator, and still a third person might use the e-mail appli-
cation that is part of Microsoft’s Internet Explorer. It doesn’t matter—all these applica-
tions support protocols that understand each other. It’s similar to a superhighway. Not all
vehicles are the same, but they all have rules on what they can do in the lanes, how fast
they can go, and how they get on and off the road. We all know what happens when some-
one violates the rules on a road—it’s messy. The concern you have is a valid one. It also is
part of a much larger network-related issue known as interoperability.

20. Interoperability? What’s that?
Earlier we stated that interoperability refers to the degree in which products (software
and hardware) developed by different vendors are able to communicate successfully with
each other over a network. During the heyday of proprietary (private, vendor-specific, or
in-house) networks, interoperability really wasn’t an issue as long as one stayed with a
specific vendor’s products and protocols. Occasionally a third-party vendor would set up
shop and develop an application that had more bells and whistles (called valued-added
features) than your vendor was offering. To do so, though, this third-party vendor had to
receive permission from the primary vendor; which usually implied paying a licensing fee.
Today, however, with TCP/IP being an “open” standard, and with the Internet’s extremely
rapid growth, vendors who want to write and sell TCP/IP-based applications are free to do
so without fear of violating any proprietary copyrights. Although the protocol specifica-
tions for the TCP/IP suite of applications are freely available, the interpretation of these
protocols by different vendors is not always the same. This, coupled with the fact that
there is no governing body to oversee the development of TCP/IP-based products, some-
times leads to incompatible products.
21. Thanks for the tip. Is this something I really need to be concerned about?
You should be cognizant of it, particularly if you are a network manager responsible
for the applications that run across your network, at whom fingers will point when things
go wrong. Most computer vendors strive for interoperability with other vendors’ products.
In fact, one of the largest networking trade shows in the world is called Networld+Interop
(for interoperability). Each year at Networld+Interop competing vendors convene to dis-
play their products and to demonstrate how they can interoperate with other vendors’
products. Still, the issue of interoperability is paramount, and you should exercise care
when considering using network products from different vendors.
22. Since we’re talking about it, is the Internet a computer network?
Although it might appear that way, the Internet is not a computer network. Recall our
response to Question 3. We said that a computer network “is a collection of computers and
other devices.” The Internet does not consist of a collection of computers and other
devices. Instead, it consists of a collection of computer networks.
Just as computers can be connected to one another to form a network, computer net-
works can be connected to one another creating what is known as a network of networks,
or an internet. For example, a network located in an office on one floor of a building can
be connected to another network located on a different floor of the same building. Collec-
tively, these two interconnected networks represent an internet.
23. I noticed that you wrote “internet” and not “Internet?” Is there a difference
between the two?
The term internet is an abbreviation for internetwork, which refers to a collection of
interconnected networks that functions as a single network. When used as a proper noun
and spelled with an uppercase I, the Internet refers to the world’s largest internetwork,

which consists of hundreds of thousands of interconnected networks worldwide and has
associated with it a certain culture. The Internet also implies a set of networks that support
the same network protocol, namely, TCP/IP. Thus, the Internet is a collection of computer
networks based on a specific set of network standards (TCP/IP), which describe how the
computers of each individual network are to communicate with each other. The Internet
allows individual, autonomous networks to function and appear as a single large network.
The Internet and TCP/IP are discussed in more detail in Chapter 3.
24. OK, I’m really confused now. I’ve heard of something called an intranet. How is
that different from an internet?
It’s easy to get confused with all the network buzzwords. An intranet is the internal
network implementation of traditional Internet applications within a company or an insti-
tution. Examples of applications that run on corporate or institutional internets are Web
servers, e-mail, newsgroups. There are many others. It is, in the strictest sense, still an
internet (notice the lack of an uppercase I), but it is easier to understand that the speaker is
referring to the internal corporate network by calling it an intranet. To make things even
more confusing, a popular networking term for an interconnection from the internal intra-
net to a customer or noncompany network that is not the Internet connection is called an
extranet connection. This may involve a leased-line connection or some other network
type of connection; it may also involve the use of a secure protocol to “tunnel” through the
Internet. This is discussed further in Chapter 7.
In summary, an intranet is an internal company network that implements traditional
Internet services; an extranet is a network connection to noncompany entities that are not
being accessed via an Internet connection; and the Internet is a series of worldwide net-
work services available from an Internet service provider (ISP), which is discussed in
more detail in Chapter 3.
25. Are the networks connected to the Internet called local area networks or LANs?
Some are and some are not. Computer networks frequently are classified by the geo-
graphical area they encompass. One classification is local area network (LAN). Another is
wide area network (WAN). A LAN generally interconnects computing resources within a
moderately sized geographical area. This can include a room, several rooms within a
building, or several buildings of a campus. Since the term “moderately sized” is not well
defined, some people quantify a LAN’s range by restricting it from a few feet to several
miles or kilometers (the IEEE usually relates this to 10 km or less in radius). In contrast to
a LAN, a WAN interconnects computing resources that are widely separated geographi-
cally (usually over 100 km). This includes towns, cities, states, and countries. Following
the quantification of a LAN’s range, a WAN would span an area greater than five miles
(eight kilometers). A WAN can be thought of as consisting of a collection of LANs.
26. So the Internet consists of a collection of WANs and LANs?
That’s correct. You should note, though, that some people make further distinctions
between LANs and WANs. One such distinction is metropolitan area network (MAN),

which interconnects computing resources that span a metropolitan area. For example, con-
sider a large business organization with buildings located throughout a local county or
city. If each building has its own independent LAN, and if these LANs were intercon-
nected to one another (thus forming an internet), the resulting network could be consid-
ered a MAN since all of the buildings are located within the same metropolitan area,
namely, the local county. MANs generally refer to networks that span a larger geographi-
cal area than LANs but a smaller geographical area than WANs.
Another classification is personal area network (PAN), which refers to the small com-
puter networks that are found in private homes. The relatively low cost of computers and
the resulting growing number of multicomputer homes is driving the need for PANs as
home computer users begin to realize the convenience of interconnecting their computers.
For example, PANs can interconnect multiple home computers to the same printer, thereby
eliminating the need to purchase separate printers for each computer. PANs can also
enable home-based computer users to use a file server on which application software and
user data can reside but are accessible from any machine connected to the home network.
PANs also provide all members of a household with convenient access to home-based
shared computing resources from their private rooms (e.g., a child’s bedroom, home
office, or kitchen). Still another classification is global area network (GAN), which refers
to a collection of WANs that span the globe. For example, many businesses such as
McDonald’s Restaurants have operations in many different countries throughout the
world. The interconnection of these individual business locations makes up a GAN.
Finally, there is storage area network (SAN), which is a network dedicated exclusively for
storing data. Figure 1.2 provides a pictorial representation of some of these types of net-
works. Note that many of the devices shown in Figure 1.2 are discussed later in this chap-
ter or in subsequent chapters.
As an interesting note, IBM has created a device that is worn like a pager by humans
and is connected to the wearer’s skin via small sensors. Using the conductivity of the skin
of the wearer as the network medium, the device can communicate with another wearer of
a comparable device to exchange information between the personal devices. IBM calls this
type of network a personal area network as well, although it is somewhat different than the
above definition of a PAN.
27. Could you expand on the concept of storage are networks for a moment?
Sure. A SAN can be thought of as a “back-end” network that connects data storage
devices via specialized channels such as the small computer systems interface (SCSI, pro-
nounced “scuzzy”) or Fibre channel. We will discuss these channels later in Chapter 4.
For the time being, you may think of these channels simply as a method for connecting
devices. A SAN can be configured in a centralized or distributed manner. A centralized
SAN configuration (Figure 1.3a) links multiple hosts to a single storage system. Consistent
with the centralized concept, a centralized SAN makes data storage management simpler
and easier but also represents a single source of failure. A distributed SAN configuration
(Figure 1.3b) connects multiple hosts with multiple storage devices. Finally, and for com-
pleteness, there is also something called network attached storage (NAS), which inter-
connects hosts and storage devices via a network device such as a a repeater hub or network
switch (Figure 1.3c). We will discuss repeater hubs and switches later in Chapter 6.

Centralized
Backbone
Switch
Building 1 Building 2 Building 3
High-Speed Link
Internet
Ethernet Switch Ethernet Switch Ethernet Switch
(a) Sample Local Area Network (LAN)
LAN
CSU / DSU
Router / Switch
Orlando, FL Dallas, TX
Router / Switch
CSU / DSU
High-Speed Link
(b) Sample Wide Area Network (WAN)
LAN
Application Servers
Mainframe
Storage
Backup
Array
(c) Sample Storage Area Network (SAN)
FIGURE 1.2 Examples of various types of networks. In (a) a local area network (LAN) intercon-
nects the computing resources housed in several buildings of an organization. Note that if the build-
ings were located throughout a metropolitan area such as a city or county, then we would have a
metropolitan area network (MAN). In (b) a wide area network (WAN) interconnects LANs that are
located in different states. If LANs, MANs, and WANs that span the globe are interconnected, then
we have a global area network (GAN). In (c) a storage area network (SAN) is a back-end network
that connects storage devices to multiple application servers or mainframes.

Application Servers
(a)
Ethernet Switch
Application Servers
Disk Storage
Disk Array
Raid Array
(b)
Fibre Channel
or SCSI Switch
Application Servers
Disk
Disk Array
Raid Array
Fibre Channel
or SCSI Links
(c)
FIGURE 1.3 A storage area network (SAN) comprises several components, including an interface
such as a Fibre channel or SCSI link, an interconnect such as a switch or hub, a protocol such as SCSI
or IP, and storage devices such as disks or tapes. All are connected to multiple application servers. In
(a) a centralized SAN (channel attached) configuration is shown; in (b) a distributed SAN (channel
attached is shown; and in (c) a networked attached SAN is shown.
TCP/IP
Storage
Array
Disk
Storage
Array

28. In Figure 1.2(a), I see the word “centralized” is used. Sometimes when I hear peo-
ple talk about computer networks, they often use terms like centralized, decentral-
ized, and distributed. What do these terms mean?
The terms centralized, decentralized, and distributed are part of the old MIS (manage-
ment information systems) vocabulary and denote specific data communications models.
These terms are more applicable to the use of computing and not networking resources.
However, they are still used today to describe various communication models and hence war-
rant consideration. To understand the differences among them, think of an organization’s
typical computing needs. For example, there is accounting, inventory, and maintenance of
personnel records. Computer communication models can be designed to handle these func-
tions in several ways. In a decentralized model, offices or departments have their own sys-
tems, independent of each other, and maintain separate databases germane to their specific
activities. In a centralized model, a single computer provides all the computing resources for
all offices and departments within an organization. Finally, in a distributed model, computers
are linked together to provide, in a transparent manner, the required computing resources and
information processing needs of the entire organization. Distributed systems bear the great-
est resemblance to computer networks. A pictorial description that distinguishes among
these three terms is given in Figure 1.4. Note that although these terms are still used today,
their corresponding computer communication models have since evolved into various net-
working models. These include client/server, peer-to-peer, Web-based, and servent. We dis-
cuss these models later in the chapter.
Users Users
Users
ecentrali e entrali e i tri ute
FIGURE 1.4 In a decentralized environment, users are “happy” but managers are not; in a centralized
environment, managers are “happy” but users are not; in a distributed environment, both managers and
users are “happy.” Source: Adapted from Distributed Processing, by F. Kaufman, p. 14. Copyright ©
1977 Coopers and Lybrand. Found in Data Center Operations: A Guide to Effective Planing,
Processing, and Performance, 2e, by H. Schaeffer, p. 272. Copyright @ 1987 Prentice-Hall.
Management Management
Management
Proce ing Proce ing Proce ing

29. In what way are distributed systems similar to networks?
A distributed system consists of independent computers interconnected to one another.
The primary difference between the two is that in a distributed environment resources are
made available to the user in a transparent manner.What this means is that users are unaware
thatthecomputersareinterconnected.Fromauser’sperspective,adistributedsystemappears
as if it were a single system. Using specially designed software, all functions of a distributed
system are handled without users ever having to explicitly request a specific service. In a net-
worked environment, though, users must explicitly identify what it is they want done.
30. I need an example of how a distributed system differs from a computer network.
No problem. Consider the task of editing a file that resides on a remote system. In a
distributed environment a user would simply “call up” the file to be edited and the system
would make the file available. In a computer network, however, the user must first know
on what remote host the file resides, and then either (a) transfer the file to the local host
(which involves running a file transfer program), or (b) establish a login to the remote host
on which the file is located (which involves running a virtual terminal program to log into
the remote host). Thus, in a distributed system the file appears local to the user regardless
of where the file actually resides within the system, whereas in a computer network the
user must be cognizant of the file’s residence and then explicitly perform some function to
gain access to the file. Although not exactly a computer network, there is considerable
overlap between distributed systems and computer networks. Suffice it to say that a dis-
tributed system represents a special case of a network, with the major distinction being the
software as opposed to the hardware.
31. Now that I have some understanding of these legacy communication models, tell
me about the current networking models that you mentioned earlier, namely, cli-
ent-server, peer-to-peer, Web-based, and servent.
OK. We’ll begin with client/server. Most network communications and applications
today are based on a paradigm called the client/server model. This model describes net-
work services (e.g., file transfers, terminal connections, electronic mail, and printing) and
the programs used by end users to access these services. The client/server model can be
thought of as dividing a network transaction into two parts: The client side (or front end)
provides a user with an interface for requesting services from the network, and the server
side (or back end) is responsible for accepting user requests for services and providing
these services transparent to the user. Both terms—client and server—can be applied to
either application programs or actual computing devices.
32. Please give me an example.
Okay. Let’s assume you are using Microsoft Word on a networked PC. Let’s further
assume that the printer you will use to print this document is a networked device accessi-
ble by your PC. Thus, we have a common scenario in which an end user using an applica-
tion program (Word) needs to access a specific network service (printing). In this context,
the application program becomes a client when it relays the print request to the printer,

while the printer is the server, which accepts and services this request. It’s similar to being
served in a restaurant.You are the client who issues a request (you order a salad), and your
waiter or waitress is the server who services your request (he or she brings you the salad).
33. I understand the concept. What I was hoping for was a little more detail. For
example, would you walk me through an Internet-based client/server example?
Sure. A typical Internet-based client/server interaction works as follows:
• A server process is started on a host. This process notifies the host that it is ready
to accept client requests. The server process then waits for a client process to con-
tact it to request a specific network application service.
• Independent of the server process, a client process is started. This process can be
invoked either on the same system that is hosting the server process, or on another
computer that is connected to the same network to which the computer supporting
the server process is connected. Regardless of which system is involved, a client
process is usually initiated by a user through an application program.A request for
service is sent by the client process to the host that is providing the requested ser-
vice and the server program running on that host responds to the request.
• When the server process has fully honored the client’s request, the server returns
toa“wait”stateandwaitsforanotherclientrequestfromthesameoranotherclient.
On UNIX systems, a server process is commonly referred to as a daemon and is desig-
nated by the letter d at the end of a program’s name. (Note: On operating systems such as
Windows NT, OpenVMS, and OS/400, different nomenclature is used.) For example, the
virtual terminal program telnet represents the client side and its companion telnetd (pro-
nounced “telnet dee”) is the server side. Similarly, the file transfer program has both a cli-
ent and a server side, ftp and ftpd. An example of an Internet-based client/server
interaction is shown in Figure 1.5. Host A is running an HTTP server process (httpd), and
a user on B is requesting a specific document from this server using a Web browser such as
Microsoft’s Internet Explorer or Netscape’s Communicator. The Web browser is the appli-
cation that supports the HTTP client protocol. When the server receives the request, it pro-
cesses it by transferring the requested Web page or document to the client.
FIGURE 1.5 Example of an HTTP client/server process. Host A is running a Web server process,
and host B is running a Web client process (i.e., a Web browser). When a user on host B opens a lo-
cation (i.e., a Web address or URL), a connection is established to the machine at that address. The
server accepts this request and services it by transferring the requested document to the client ma-
chine that made the request.
(Web page from hostA.com)
httpd
Client Request
Server Response
http://www.hostA.com
Host A (Web Server) Host B (Web Client)

Computer systems that run specific server programs are commonly referred to by the
application service they provide. For instance, a host that accepts and provides mail service
is known as a mail server; a computer that provides users access to files remotely is known
as a file server; a computer running httpd is known as a Web server; and a computer that
runs a network news protocol (e.g., NNTP) is known as a news server.
34. Thanks. That was helpful. How does the client/server model compare to the
peer-to-peer model?
Peer-to-peer is another model on which some network communications and applica-
tions are based. In a peer-to-peer environment, each networked host runs both the client
and server parts of an application. (Contrast this to a client/server environment in which a
host is capable of running only client-based applications, only server applications, or both
client and server applications, thereby acting as a server for one application but a client for
another application.) This is accomplished by installing the same network operating sys-
tem (NOS) on all hosts within the network, enabling them to provide resources and ser-
vices to all other networked hosts. For example, each networked host in a peer-to-peer
environment can allow any other connected host to access its files or print documents on
its printer while it is being used as a workstation. Once again, the key distinction between
peer-to-peer computing and client/server computing is the former requires each networked
host to run both client and server parts of all applications. Examples of peer-to-peer net-
works include Microsoft’s Windows NT and Windows 95/98/2000, Apple’s AppleTalk,
Artisoft’s LANtastic, and Novell’s Personal NetWare. Peer-to-peer networks are relatively
inexpensive to purchase and operate, and are fairly easy and straightforward to configure.
They represent a cost-effective model for small organizations or departments that want to
enjoy some of the benefits of networking but do not have the requisite resources (financial,
human, or equipment). However, peer-to-peer networks can be less reliable than cli-
ent/server based networks. They usually also require the use of more powerful worksta-
tions for certain activities (e.g., sharing a database) than a client/server based network.
35. Time out. What is a network operating system and how different is it from a reg-
ular operating system?
The term network operating system (NOS) refers to software that is installed on a sys-
tem to make it network-capable. Examples include IBM’s LAN Server, Banyan’s VINES,
and Novell’s NetWare (also known as IntranetWare). In each of these cases, the NOS is
independent of a computer’s native operating system—it is loaded “on top” of the com-
puter’s operating system and provides the computer with networking capability based on a
particular protocol. If, on the other hand, a computer’s native operating system includes
built-in network support, then a NOS refers to that particular OS. Examples include Sun
Microsystems’ Solaris Operating System, Hewlett-Packard’s HP-UX Operating System,
and Microsoft’s NT Server. You can think of the NOS in this case meaning networkable
operating system.

36. OK. Getting back to the peer-to-peer model, if my PC is running Windows
95/98/2000 and is connected to a network, then any other PC running Windows
95/98/2000 that is connected to the same network can access files on my machine
because of this peer-to-peer thing. Is that correct?
Yes, provided your machine has been configured for this access. It’s just like Apple
Macintoshes, which support AppleTalk. If you go into the Control Panel and invoke “file
sharing,” you can make your Macintosh accessible to any other machine connected to the
same network that also is running AppleTalk. In Windows, you select My Computer, fol-
lowed by the disk you are interested in sharing, and then right-click the mouse. If your net-
work Control Panel applet has been configured for file sharing, you will see the word
“Sharing” in the menu and you simply select it and then set up the parameters for sharing
that will be allowed. (If sharing is already enabled on a Windows 95/98/2000 system, you
will see a picture of a hand holding the device icon.) This is the concept of peer-to-peer
networking—it enables users to easily share resources on a network.
Network printing is implemented in a similar manner. For example, a locally con-
nected printer in Windows NT 4.x can be configured as a network device. During the ini-
tial configuration process, the “Add Printer Wizard” provides two radio buttons—Shared
and Not Shared. Selecting “Shared” makes the printer accessible network-wide. (You also
have to give it a name and identify the operating systems that will print to this printer so
the proper device drivers are installed.) Now, assuming users have the proper access per-
missions, this printer appears as an available resource whenever users browse network
resources in Windows NT Explorer.
37. This sounds like the way Napster works.
Exactly. As you probably know, Napster was the brainchild of Shawn Fanning, who, as
a 19-year old college student, was interested in making it easy for music listeners (espe-
cially other college students) to share copies of their favorite recordings. Prior to Napster’s
development, electronic versions of songs were accessible via the Internet and could be
downloaded using a Web browser and TCP/IP’s file transfer protocol (FTP). Recordings
were saved using the MPEG-1 audio layer III (MP3) compression format (see Chapter
17). This approach was deficient, though, because there was no current list that identified
which recordings were available and where they were located.
Napster addressed this deficiency. Using a locally installed client application along
with a custom networking protocol, users specify to a Napster server their local host’s
Internet address and the names of the recordings they want to share. Other Napster clients
can then search this server for the files they want. The actual file transfer, however, does
not involve the Napster server. Instead, the transfer is between two Napster clients—the
host on which the requested file resides and the host that is requesting the file. Thus, the
Napster server simply brokers the file transfer between clients.
Using our previously developed vocabulary, in the Napster model the local and remote
systems are both client and server, similar to the peer-to-peer model. In this model, how-
ever, all the Napster hosts are concurrently servers and clients, and the term that is used to
denote this is servent, which is a combination of SERVer and cliENT. Thus, unlike the
peer-to-peer model, which is one-to-one, or the client/server model, which is one-to-many,

the Napster model is many-to-many. The servent concept is expected to be expanded into a
general file-sharing model that will enable all interconnected hosts to exchange any type
of file residing on their hard drives.
38. Isn’t this sort of what Gnutella is like? How does Gnutella relate to Napster?
Yes. Gnutella is similar to Napster. It too is an open, decentralized, peer-to-peer file
search system. Unlike Napster, though, Gnutella is not a company, it is not an application,
and it is not a Web site. It is simply a name for a file-sharing technology.
39. What’s the future of the servent model given the court’s ruling against Napster?
This is a difficult question to answer because of the problematic nature of current law
as applied to the Internet. Nevertheless, some believe that once all the copyright related
issues and attending legal matters are resolved, the servent concept is expected to be
expanded into a general file-sharing model that will enable all interconnected hosts to
exchange any type of file residing on their hard drives.
40. Is this sharing model concept also true for client/server? For example, can other
machines access my machine if they are running the same NOS?
No. In the client/server model, it is important to note that a network service can only
be provided if a server program responsible for servicing a request is running on a particu-
lar host. For example, look back at Figure 1.5. If host A was not running the httpd process,
then the request from host B would not be honored. This is why PCs used to access Inter-
net services such as e-mail or the World Wide Web are relatively secure from being com-
promised by “outside” users. These machines usually run client versions of
Internet-related applications. For example, Eudora is an Internet-based mail client pro-
gram that requests mail service from a mail server. It makes a connection to another
machine running the server process to retrieve mail for a user. If you are running Eudora
on your PC, Eudora users on other machines cannot connect to your machine just because
you are running Eudora as well; your machine is not running an appropriate mail server
program (Eudora is a client, not a server). Although some PC- or Macintosh-based Inter-
net applications can be configured as servers (e.g., there are mail, ftp, gopher, and WWW
server applications available for PCs and Macintoshes), most users only run the client side
of these applications. As a result, without a server process running on a system, a network
connection to that system cannot be made by another machine.
41. OK. I now understand the difference among client/server, peer-to-peer, and ser-
vent. And I even see how they are different than the earlier computer communi-
cations models of decentralized, centralized, and distributed. What I still don’t
know, though, is where the World Wide Web fits into all of this.
The World Wide Web (WWW or Web for short) is yet another networking model that
emerged as a result of the Internet. From a communications model perspective, the Web
can be viewed as a collection of Internet-based clients and servers that speak the same lan-

guage, namely, the Hypertext Transfer Protocol (HTTP). The Web is accessible through
Web browsers such as Netscape’s Communicator or Microsoft’s Internet Explorer. With the
introduction of a programming language called Java in the mid-1990s by Sun Microsys-
tems, animated and interactive Web pages can be designed in addition to the standard fare
of displaying text and images. A direct result of the Web’s evolution is a Web-based com-
munications model that fosters the notion of The network is the computer, a phrase coined
in the late 1980s by Sun Microsystems’ president, Scott McNealy. “The network is the
computer” implies that by making resources available to users via a network, the network
essentially becomes the single, most powerful computer accessible. Thus, the network
gives users access to more computing power than their desktop models. Although this com-
munications model is still client/server based, it deserves separate recognition because of
its potential for reshaping the manner in which resources are provisioned to the end user.
42. Is this the networking model that is used by all of those speciality networking
devices you see today like dedicated e-mail devices?
Yes, and for the record, those devices are formally called network appliances. (Recall
from an earlier response we stated that an appliance was another term commonly used to
denote a network node or device.) Other names commonly used include netappliances,
information appliances, or Internet appliances. Unlike traditional laptop or desktop com-
puters, network appliances usually support a single, dedicated application such as Web
browsing or e-mail, and in some cases do not have keyboards or monitors. Nevertheless,
network appliances are powerful computing devices designed for the average consumer
who wants Internet access but does not want to be burdened with the attending problems
and maintenance issues often related to personal computer ownership.
Network appliances usually rely on application service providers (ASPs) (also known
as content service providers or CSPs) to furnish users with computing resources via the
Internet. Examples of network appliances include: handheld devices such as the PalmPi-
lot; two-way “smart pagers” that can send and receive e-mail and provide Web access;
“smart phones,” which are specially designed corded desktop phones or wireless devices
that provide standard voice communications as well as Internet access for e-mail service
and Web browsing; and TV/set-top boxes such as WebTV, which make Web access and
e-mail service available to consumers via standard television sets. Other network appli-
ances currently being developed or deployed as of this writing include: automobile dash-
board-installed Internet connections and service; refrigerators with an in-door
screen-based PC and modem for Internet access, which will enable the device to monitor
food quantities and automatically order food from Internet-based grocers; and Inter-
net-enabled fax machines that can send and receive e-mail. (See http://devices.inter-
net.com for additional information about the rapidly evolving consumer network
appliance marketplace.)
43. Would you consider a network computer a netappliance?
Yes. A network computer (NC) is a business-based network appliance instead of a con-
sumer-based appliance. A NC promotes the concept of thin client computing. It is an inex-
pensive ($500 or less) network access device with functionality that allows some

applications to be run, but not as complete as what would typically be found on a PC or a
workstation of some sort. Notice that we use the term “device” here.
Although the term “computer” is part of their name, NCs are not computers; they have
a specialized, proprietary (and highly restricted) operating system and are usually diskless
(i.e., most have no hard disk drives for local storage). NCs are stripped-down systems that
use the network to access their applications dynamically. For example, if you need a word
processor, a copy of a word processing application is downloaded from a network server to
your NC and stored in its memory (RAM) for you to use. Any documents you create are
uploaded to and saved on the server. The idea behind NCs is to offer businesses a tremen-
dous reduction in cost-of-ownership for each desktop location where a more expensive tra-
ditional terminal or PC would otherwise be used. By incorporating a massive server, or
server “farm,” with user NCs, companies can save money compared with purchasing
fully-loaded PCs for each user and dealing with their management and maintenance.
44. Given what I’ve been able to understand so far, this sounds like centralized com-
puting to me. Is this correct?
In a sense it is—what goes around comes around. What do you call a computing
device that relies on the network for its application? A terminal. The NC concept is very
reminiscent of the era when terminals (dumb or otherwise) were connected to a main-
frame. (The MIS people called this a network; we know better now.) It also is similar in
concept to diskless UNIX workstations and X-terminals.
45. OK. What’s next?
Since we just discussed the various communication models, this is probably a good
time to extend our discussion to include the various communication service methods and
data transmission modes. We’ll begin with serial and parallel communication.
46. I am pretty familiar with these two terms already. Serial communication means
sending data one bit at a time; parallel communication means sending data in
parallel, like eight bits at a time. Is this right?
Yes. Should we skip these terms then?
47. No. Go ahead and review them for me. It can’t hurt.
As you said, serial communication (also referred to as serial transmission) is a data
transmission method in which the bits representing a character of data are transmitted in
sequence, one bit at a time, over a single communications channel. Serial transmission is
limited to the speed of the line. Parallel communication (also called parallel transmission)
refers to the simultaneous transmission, each on a separate channel, of all the bits repre-
senting a character. In contrast to serial communications, a parallel link transmits a group
of bits at one time. The number of bits varies from device to device. Consequently, assum-
ing the line speeds were the same, in the same amount of time required to transmit one bit
of information to a remote node over a serial line we can transmit eight bits (or more) of
data over a parallel line. (See Figure 1.6.)

48. If parallel communication is so much faster than serial communication, why are
most network links serial-based?
Although parallel communication is capable of transmitting data more quickly than
serial communication, it does have its limitations. For instance, parallel communication
requires a relatively complex communication link, which is achieved through the use of
large, multiwire cables. Also, the longer the parallel link, the worse the degradation of the
electrical signal from the most distant nodes. Consequently, in most networking applica-
tions, parallel communication is limited to peripherals directly connected to a system and
for communication between systems that are relatively close to each other (in many cases,
within a few yards or meters of each other). Serial communication, on the other hand, with
its simpler data path, is slower but enables data transmission to occur over existing com-
munications systems that were not originally designed for such transmission. As a result,
serial communications are seen nearly everywhere, including in terminal-to-systems con-
nections, via leased phone lines for data transfers, dialup lines, and satellite links.
49. Wait a minute. How does the receiver of a serial transmission know when a com-
plete unit of data has been received? For example, in Figure 1.6, a single charac-
ter, E, was transferred. What happens if a second character (say, X) is
transferred immediately after the ﬁrst? How does the receiving node identify the
beginning (or ending) of a character when all it is seeing is a stream of bits (0s
and 1s)?
That’s a good question. Obviously, without some way of identifying the beginning (or
ending) of a character, the transmitted data would be indecipherable, resulting in some sort
of communication breakdown. Two methods can be employed to resolve this problem: We
can either synchronize the sending and receiving nodes so that the receiving node always
knows when a new character is being sent, or we can insert within the bit stream special
“start” and “stop” bits that signify the beginning and end of a character. The former tech-
nique is called synchronous communication and the latter is called asynchronous commu-
nication. Synchronous communication is also tied to the clocking inherent on the link.
FIGURE 1.6 Serial and parallel communications. Here, the character E, which is 01000101 in bi-
nary, is transmitted 1 bit at a time (serial communication) and 8 bits at once in parallel.
0 →1 → 0 → 0 → 0 → 1 → 0 → 1 →
Serial ommunication Parallel ommunication
(1 channel transmits 1 bit at a time)
(8 parallel channels transmit 1 bit)
0 →
1 →
0 →
0 →
0 →
1 →
0 →
1 →

50. Oh yes. I recall these two terms, but I always have trouble remembering which is
which. Can you help?
You bet. Synchronous communication implies that communication between two nodes
is monitored by each node. That is, all actions resulting in data transmission (and general
link conditions) are closely synchronized between the nodes. If data are to be transmitted
or received, then the nodes are aware of this transmission almost immediately and prepare
for the exchange based on ordered data rates and sizes. Thus, the sending and receiving
nodes are “in sync” with each other. Asynchronous communication (commonly referred to
as async) is characterized by the encapsulation of data within special start and stop bits.
Hence, asynchronous communication is sometimes called start-stop transmission. A
direct consequence of the inclusion of these start-stop bits in the bit stream is that data can
be transferred at any time by the sending node without the receiving node having any
advance notiﬁcation of the transfer. Thus, a receiving node does not necessarily know
when data are being sent or the length of the message. An example of async communica-
tions is a computer terminal (sender) connected to a system (receiver). The system does
not know when someone will begin entering data on a terminal. As a result the system
must always be in a “ready” state. Async communication lines remain in an idle state until
the hardware on the line is ready to transmit data. Since the line is idle, a series of bits have
to be sent to the receiving node to notify it that there are data coming. At the conclusion of
a transmission, the node has to be notiﬁed that the transmission is complete so that it can
return to an idle state, hence the stop bits. This pattern continues for the duration of the
time the link is operative. We like to view the difference between these two terms from the
perspective of a mugging on a television crime drama. Are you interested?
51. In getting mugged, no. In hearing how they can be related to a mugging, yes.
In an asynchronous mugging, you know when the actor is going to be attacked and
hence are ready for it, but you do not know when it will occur. In a synchronous mugging,
you not only know the actor is going to be mugged, but you also know when, so again, you
are ready. Note that the term asynchronous is commonly used in the context of distance
education. Through distance education technologies, education can be delivered “asyn-
chronously,” namely, at any time or place.
52. Where is each type of communication found with respect to networks?
A variety of both types of communications is found on most computer networks. Most
terminals, dialup modems, and local links are asynchronous in nature. Synchronous com-
munication tends to be more expensive than asynchronous because of the need for sophis-
ticated clocking mechanisms in the hardware. However, synchronous communication can
eliminate up to 20 percent of associated overhead inherent in asynchronous communica-
tion. (This overhead percentage is easy to compute. If we need two bits—one start and one
stop bit—for each eight bit character transmitted on a serial link, then two out of every ten
bits are “wasted” on overhead functions.) This allows for greater data throughput (i.e., the
amount of real data that can be transferred in a given period) and better error detection.
Synchronous communications are typically seen in higher speed connections.

53.Whataboutisochronouscommunication?I’mstartingtohearingalotaboutit.What
is it and how does it relate to the synchronous and asynchronous communications?
Isochronous communications is a term used to describe the delivery of time sensitive
data such as voice or video transmissions. Networks that are capable of delivering isochro-
nous service preallocate a specific amount of bandwidth over regular intervals to ensure
that the transmission is not interrupted. Isochronous communications was originally
intended to service the requirements of constant and complete delivery of video communi-
cations over a transmission medium. For example, television signaling in the United States
requires that 30 frames per second of video be delivered to the receiver for full motion
video. Not 29, not 31, but exactly 30. By establishing in the communications path that a
session is going to require a specific data transmission rate on the path and also communi-
cating what the data rate will be, a continual and uninterrupted flow of data can be real-
ized. This is critical for the delivery of applications such as video, which requires a
constant-bit rate (CBR) of information to be sent and delivered over the communications
interface. Isochronous communications makes this happen. Networks such as ATM,
SONET and a special full-duplex version of Ethernet have isochronous capabilities. To
extend the previous mugging analogy, isochronous is like being mugged all the time, but
the mugger and muggee have agreed on the mugging interval and constancy before the
mugging actually started.
54. All this talk about serial, parallel, synchronous asynchronous, and isochronous
communications reminds me to ask about simplex and duplex communications.
Serial, parallel, synchronous, asynchronous, and isochronous communications repre-
sent different techniques for transferring data. Associated with these techniques are three
different modes of data transmission used for communication purposes; each corresponds
to a specific type of circuit—simplex, half-duplex, and full-duplex. These modes specify
the protocols sending and receiving nodes follow when transferring data. Figure 1.7 con-
tains a summary of these three transmission modes.
Simple ran mi ion
al uple ran mi ion
Full uple ran mi ion
Sender Receiver
Sender
or
Receiver Sender
or
Receiver
Only Only
One way at a time
Sender
and
Receiver Sender
and
Receiver
FIGURE 1.7 Simplex is a “fixed” one-way transmission; half-duplex is a two-way transmission, but
only one way at a time; full-duplex is a simultaneous two-way transmission.
Simultaneously

55. Would you please review these for me?
Simplex communications imply a simple method of communicating. In a simplex com-
munications mode, data may flow in only one direction; one device assumes the role of
sender and the other assumes the role of receiver. Furthermore, these roles may not be
reversed. An example of a simplex communication is a television transmission—the main
transmitter sends out a signal (broadcast), but it does not expect a reply since the receiving
devices cannot issue a reply back to the transmitter. It’s like a very boring person telling
you his or her life story and you can neither interrupt it nor get away from it. . . . Sigh!
In half-duplex transmission, data may travel in either direction, but only one unit can
send at any one time. While one node is in send mode, the other is in receive mode.
Half-duplex communication is analogous to a conversation between two polite peo-
ple—while one talks, the other listens, but neither talks at the same time. An example of a
half-duplex communication is a citizens band (CB) transceiver. Users of a CB transceiver
can either be senders or receivers but not both at the same time. Another example of
half-duplex communication is like a very polite game of tag between two people. Only
two can play, only one can be “it” at a time, and you know who is going to be “it” next,
don’t you?
A full-duplex transmission involves a link that allows simultaneous sending and receiv-
ing of data in both directions. Imagine, if you can, two people talking at the same time and
each one understanding the other one perfectly. Compound this idea with the added benefit
of not having to talk about the same thing. This is the realization of full-duplex commu-
nications—two separate but parallel transmissions occurring simultaneously. A full-duplex
line can be thought of as the combination of two simplex lines, one in each direction.
56. OK. I am starting to get a headache and really need some time to assimilate all of
this material. Before we take a break, though, I would like to know something
about network standards.
A veritable plethora of network standards has been developed defining such things as
hardware interfaces, communication protocols, and network architectures. Network stan-
dards establish specific rules or regulations to be followed. Standards also promote
interoperability among different hardware and software vendors’ products.
57. How are network standards created and by whom?
Standards are developed in several ways. First, they can be developed through formal
standards organizations. These organizations can be classified into four major categories:
(1) National, (2) Regional, (3) International, and (4) Industry, Trade, and Professional. A
list of some of the influential organizations within these categories is given in Table 1.1.
Standards organizations are composed of delegates from the government, academia, and
from vendors who will be developing products based on the proposed standards.
58. How long does it take to get a standard approved?
The formal standards process, which is designed to ensure that a consensus is reached,
is often lengthy and sometimes can take years before a proposed standard is approved. The
process also is politically charged. A summary of this process is given in Table 1.2.

59. What kind of standards are there?
Standards can be viewed from four different perspectives. There are de jure standards,
de facto standards, proprietary standards, and consortia standards.
60. Wow. I didn’t realize there were so many. Could you summarize them for me?
Sure. De jure standards are approved by a formal, accredited standards organization.
(“De jure” is Latin for “by right, according to law.”) Examples include modem protocols
developed by the International Telecommunications Union (ITU), the EIA/TIA-568 Stan-
dard for Commercial Building Telecommunications Wiring developed by the Electronic
Industries Association (EIA) and Telecommunications Industries Association (TIA), and
standards for local area networks developed by the Institute for Electrical and Electronic
Engineers (IEEE). (Note: These standards are discussed in subsequent chapters.)
TABLE 1.1 Network Standards Organizations
Source: Adapted from Conrad, 1988.
National Standards Organizations
(Generally responsible for standards within a nation and usually participate in that nation’s
international activity)
• American National Standards Institute (ANSI)
• British Standards Institute (BSI)
• French Association for Normalization (AFNOR)
• German Institute for Normalization (DIN)
Regional Standards Organizations
(Restrict their activity to a specific geographical region but generally influence standards outside
their regions)
• Committee of European Posts and Telegraph (CEPT)
• European Committee for Standardization (CEN)
• European Computer Manufacturers’Association (ECMA)
International Standards Organizations
(Promote standards for worldwide use)
• International Standards Organization (ISO)
• International Telecommunications Union (ITU)—consists of ITU-T, which is responsible
for communications, interfaces, and other standards related to telecommunications; and
ITU-R, which is responsible for allocating frequency bands in the electromagnetic
spectrum for telecommunications, and for making recommendations relating to radio
communications. (Note: ITU-T is the former CCITT—Consultative Committee for
International Telephony and Telegraphy.)
Industry, Trade, and Professional Standards Organizations
(Restrict their activity to member interest areas but generally influence other areas)
• Electronic Industries Association (EIA)
• Telecommunications Industries Association (TIA)
• Institute for Electrical and Electronics Engineers (IEEE)
• Internet Engineering Task Force (IETF)

De facto standards are those that have come into existence without any formal plan by
any of the standards organizations. Rather, they are developed through the industry’s
acceptance of a specific vendor’s standard, which is placed in the public domain. (“De
facto” is Latin for “from the fact.”) One example is Network File System, (NFS), a de
facto file-sharing protocol standard developed by Sun Microsystems. Sun placed the spec-
ifications of this protocol in the public domain so that other vendors could implement it.
This resulted in widespread use of NFS and established NFS as a de facto standard. NFS is
now implemented on a variety of UNIX systems (including those from Sun, IBM, Silicon
Graphics, and HP), as well as Macintosh and Intel-based systems. Another de facto stan-
dard is Java, which was also developed by Sun Microsystems.
Proprietary standards are those developed in a manufacturer-specific manner. This
implies that their specifications are not in the public domain and are only used and accepted
by a specific vendor. In the early days of networking, proprietary standards were the rule of
the day. Although such standards are now frowned upon, many still exist. Some of the most
well-known are from IBM (e.g., IBM’s Systems Network Architecture or SNA). Novell’s
IPX protocol, which is based on Xerox’s XNS protocol, is also proprietary in nature. Pro-
prietary standards lock a customer into a vendor-specific solution and make it difficult for
customers to use products (software or hardware) from other vendors (see Box 1.1).
Consortia standards are similar to de jure standards in that they too are the product of a
formal planning process. The difference is the planning process and development of such
standards are not conducted under the auspices of a formal standards organization. Instead,
specifications for standards are designed and agreed upon by a group of vendors who have
formed a consortium for the express purpose of achieving a common goal. These vendors
pledge their support for the standards being developed by the consortium and also develop
TABLE 1.2 The Network Standardization Process
Source: Adapted from Conrad, 1988.
Planning Phase
• Proposals submitted by vendors or organizations are examined.
• A determination is made whether there is a need to establish a standard.
• If a need is found, the development of a project is authorized and assigned to a technical
committee.
Development Phase
• Committee prepares a working paper describing the scope of the proposed work.
• Liaisons with other standards groups are established.
• A draft proposal of the standard is produced.
• Draft is voted on and all negative comments are addressed.
• Draft is submitted to parent organization for discussion and approval.
Approval Phase
• All members of the organization vote on draft.
• Draft is made available to the public for review.
• Draft is ultimately approved as a standard.
Publication Phase
• The new standard is published.

and market products based on these mutually agreed upon set of standards. An example of
consortium-based standards include Fast Ethernet, the early efforts for Asynchronous
Transfer Mode (ATM Forum), and Gigabit Ethernet, all of which are discussed in later
chapters of this book.
61. What about Internet standards? You listed the Internet Engineering Task Force
in Table 1.1 but didn’t say anything about how these standards are developed.
You’re right. This is because the Internet Engineering Task Force (IETF) is not a tradi-
tional standards organization; it only recommends the standardization of protocols and
protocol usage in the Internet. We listed IETF in Table 1.1, though, because many of the
protocol specifications IETF produces become standards.
Internet standards are initially developed by IETF, which, according to The Tao of
IETF (see http://www.ietf.org/tao.html), “is a loosely self-organized group of people who
make technical and other contributions to the engineering and evolution of the Internet and
its technologies.” The Internet standards development process involves the generation of
special documents called Request for Comments (RFCs), which initially were written com-
ments about resolving certain Internet-related problems. Today, however, RFCs are formal
documents that comprise two sub-series. The first contains For Your Information (FYI)
documents, which provide, in a less technical manner, general overviews and introductory
information about various Internet topics. The second sub-series, STD, references those
RFCs that specify Internet standards. Before an RFC is published, it is first developed as an
Internet Draft, which enables the Internet community to read and comment on proposed
Internet-related documents before they are officially published as an RFC. Internet Drafts
are considered temporary documents and have a shelf life of only six months; hence, they
are not archived. To facilitate the dissemination process and to maintain a spirit of open-
ness, RFCs and current Internet Drafts are available online at http://www.rfc-editor.org. It
should be noted that approved Internet standards are promoted internationally by ISO.
BOX 1.1: Open Standards versus Closed Standards
• Open is for everyone.
• Closed is vendor specific.
• Open means that everyone has a chance to implement and benefit from the same stan-
dards.
• Closed means that the vendor feels that its standard has value and will not share the tech-
nology with other vendors.
• Open allows for the creation and modifications of standards by committee.
• Closed means that the vendor can effect repair or modifications without agreement of
other vendors and without comment from customers.
• Open implies that any modifications take a long time.
• Closed implies that modifications are made in a timely manner but they are nonstandard.

Discovering Diverse Content Through
Random Scribd Documents

SOMMARIO
DELLE MATERIE CONTENUTE NEL TERZO VOLUME
—
PARTE PRIMA.
LIBRO QUINTO.
Capitolo I.
an.
970-1011. Cagioni esteriori della caduta
della dominazione musulmana
in Sicilia. Movimento nazionale
nella Terraferma italiana.
Imprese navali dei Pisani
contro i Musulmani Pag. 1
1015. Mogêhid usurpatore di Denia 4
» La Sardegna infestata
precedentemente 5
» Mogêhid a Luni e in Sardegna 7
1016. È sconfitto e ricacciato in Spagna 9
» Contese de’ Pisani co’ Genovesi 10
1016-1114. Altre fazioni contro i Musulmani 13
» I Normanni 14
» Loro tradizioni 20
1078-1086. Croniche de’ Normanni d’Italia.
Amato 21

» Guglielmo di Puglia 22
» Malaterra 23
» Leone d’Ostia e Lupo 24
» I Normanni a Salerno 25
1017-1021. Melo 26
» Compagnia Normanna 29
1040-1041. Argiro e Ardoino 30
» Battaglia dell’Olivento ed altre
vicende 33
1043. Nuovo ordinamento della
Compagnia 37
» La casa di Hauteville 38
1051. Rivolta contro i Normanni 40
1055-1058. Roberto Guiscardo 42
1059. Ruggiero. Espugnazione di
Reggio 49
» Condizioni della Compagnia
Normanna 52
Capitolo II.
1060. Disposizioni de’ Cristiani
messinesi 55
» Supposta congiura 56
» Correria sopra Messina 61
» Ibn-Thimna 62
1061. Nuova fazione 63
» Presa Messina 66
» Rametta 70
» Tripi, Frazzanò, Maniace, Centorbi 71
» Paternò, Emmelesio, Sanfelice;
battaglia di Castrogiovanni 72

» Scorreria a Girgenti. Tregua con
Palermo 75
» Ritirata 76
» Castel di San Marco. Dominazioni
diverse nelle province 78
Capitolo III.
» Rivolgimento in Palermo 79
» Condizioni degli Ziriti 80
» Aiuti di Mo’ezz 81
» Scorreria di Ruggiero sopra
Girgenti 82
» Patti co’ Trainesi 83
1062. Ruggiero sposa Giuditta di Evreux 84
» Correrie in Sicilia. Morte d’Ibn-
Thimna 85
» Brighe di Ruggiero con Roberto 87
» Rivolta di Traina 89
» Vittoria di Ruggiero 91
1063. Nuova spedizione affricana 92
» Scorrerie di Ruggiero 94
» Battaglia di Cerami 96
» Fazione de’ Pisani in Palermo 101
» Fazioni de’ Normanni a Collesano,
Brucato, Cefalù.
Combattimento presso Girgenti 105
Capitolo IV.
1064. Vano assedio di Palermo 106
» Bugamo presa: scontro presso
Girgenti 107

1064-1068. Aiûb ed Ali, figliuoli di Temim,
occupano la Sicilia occidentale 108
» Guerra civile; partenza degli
Affricani ed emigrazione 110
1066. Ruggiero a Petralia 111
1068. Battaglia di Misilmeri 113
1068-1071. Assedio di Bari 114
» Armamento contro Palermo 115
» Presa Catania 116
» Assedio di Palermo 118
» Assalti 124
1072. Resa della città 130
» E di Mazara 133
Capitolo V.
» Distribuzione de’ conquisti ivi
» Morte di Serlone 134
» Roberto ordina il governo in
Palermo 136
1072-1085. Ritorna in Terraferma. Suoi doni
alla Badia di Montecassino 139
» Contrasta co’ suoi baroni 141
1072-1085. E co’ principi di Salerno e Capua 142
» Roberto e Gregorio VII 143
» Imprese di Grecia e di Roma 144
» Morte di Roberto 146
Capitolo VI.
1072. Condizioni de’ Normanni in Sicilia 147
» E dei Musulmani 148
» Benavert 149

1073-1075. Progressi lenti di Ruggiero 150
» Vittoria di Benavert 151
1076. Ruggiero dà il guasto al Val di
Noto 153
1077. Prende Trapani ed altri paesi 154
1078. E Taormina 156
1079. Rivolta di Cinisi e Giato 159
1081. Ruggiero padrone di Messina 161
» Catania presa da Benavert e
racquistata 162
1082. Rivolta di Giordano 163
1085. Scorreria di Benavert in Calabria 164
1086. Ruggiero prende Siracusa 165
1087. Impresa navale degli Italiani
sopra Mehdia 168
» Ruggiero occupa Girgenti e la
provincia 172
» Ibn Hammûd gli dà
Castrogiovanni 173
1089-1091. Prese Butera e Noto. Urbano II a
Traina 176
» Conquisto di Malta 177
Capitolo VII.
1093. Morte di Giordano e rivolta di
Pantalica 180
1085-1093. Cresciuta potenza del conte
Ruggiero 181
» Aiuta il nuovo duca di Puglia, il
quale gli concede metà di
Palermo 182
1091-1094. Imprese di Cosenza e Castrovillari 184

1096. Assedio di Amalfi. La prima
Crociata 185
1098. Ruggiero assedia Capua co’
Musulmani 186
» E impedisce la loro conversione 187
» Aneddoto attribuitogli da Ibn-el-
Athîr 188
» Scuola di monaci statisti 190
» Relazioni del conte con Urbano II 191
» Privilegio dell’Apostolica legazione 193
1101. Morte del conte 194
» Famiglia della contessa Adelaide 196
» La Marca aleramica 198
» Bonifazio del Vasto 199
Capitolo VIII.
» Condizioni dell’isola dopo il
conquisto 200
» Diplomatica siciliana dell’XI e XII
secolo. Falsa pergamena
arabica dell’archivio di Napoli 201
1101. Diplomi arabici e greci 202
» Diplomi latini 204
» Varie schiatte. Antichi abitatori 206
» Distribuzione geografica delle
nuove schiatte 207
» Ebrei 209
» Tribù arabe e berbere 210
» Normanni e altri Francesi 213
» Colonie della Terraferma italiana 218
» Lombardi 222
» Baroni aleramidi 225

» Dialetto de’ Lombardi di Sicilia 227
» Caltagirone 228
» Origini di altre città 231
» Della famiglia Bonello 232
Capitolo IX.
» Condizioni de’ vinti. Schiavi 233
» Villani 237
» Sinonimo di Rustici 238
» Due maniere di villani 242
» Domini di Maks 243
» Platee 245
» Doveri e diritti de’ villani 246
» Borghesi 250
» Non soggetti alla gezia 253
» Borghesi delle antiche schiatte 256
» Prete Scholaro 257
» I Greci non hanno titoli di nobiltà 259
» Musulmani. Kaid, titolo di nobiltà,
d’Ufficio o meramente onorifico 260
» Origine di tutte queste condizioni 267
Capitolo X.
» Se il conte di Sicilia sia stato
vassallo del duca di Puglia 271
» Costituzione politica 274
» Ruggiero prende il titolo di Gran
Conte e poi di Console 277
» Istituzioni municipali messe in
forse dal Gregorio 278

» Memorie delle municipalità
cristiane nella guerra
normanna 280
» E sotto il principato. Arconti 281
» Anziani 284
» Buoni Uomini 286
» Maestri de’ Borghesi 289
» Municipalità diverse nella stessa
città. Anche de’ Giudei. Gema’ 291
» Forma generale de’ comuni
siciliani 292
» Franchige 296
» Municipii di Palermo e di Messina 297
» Ricerche da farsi. Feudalità 299
» Feudi ecclesiastici 301
» Autorità di Ruggiero nella
gerarchia 302
» Legazia apostolica 306
» Rifatte le diocesi dal principe ivi
» Circoscrizione territoriale politica.
Iklîm 309
» Ufiziali del principe. ’Amil,
Stratego e Vicecomite 315
» Magistrati giudiziali 318
» Entrate pubbliche 319
» Platee 320
» Diwâni 322
» Defetarii 324
» Rivendicazione de’ beni demaniali 326
» Dazii e gabelle 327
» Colletta; diritto di marineria;
tratta de’ grani 331
» Servigio militare e navale 333

» Costituzione dell’armata 335
» Avanzi d’incivilimento. Chiese e
fortezze 338
» Strade militari 339
» Monete del conte Ruggiero 342

Correzioni ed Aggiunte.
Pag. lin.
12 3 n. 5. della stessa opera dello stesso volume
25 » n. 1. volume volume. Contuttociò si
vegga il De Meo,
nell'_Apparato
cronologico agli Annali
del regno di Napoli_,
Napoli, 1785, pag.
385, segg. ed una
nota posta ne' _Regii
Neapolitani archivii
Monumenta_, vol. IV,
pag. VI, nella quale è
citato un diploma del
1008.
36 7 n. 2. potessero potessero. Si riscontri
presso Trinchera,
_Syllabus graecorum
membranarum_, etc.,
Napoli, 1865, pag. 53,
un diploma del 1054,
nel quale Argiro
s'intitola: _Magister
Vestis et dux Italiae,
Calabriae, Siciliae,
Paphlagoniae_, etc.
48 27 n. al principio alla fine
56 11 e del
milledugentottantadue
del
milledugentottantadue

e del
milleottocensessanta.
63 4 n. 5. aprile. Malaterra aprile. Edrîsi, nella
descrizione della
Sicilia, _Bibl. arabo-
sicula_, testo pag. 26,
fa cominciare il
conquisto nel 463
dell'egira, cioè dal 26
gennaio 1061 al 15
gennaio 1062.
Malaterra
75 5 discosta discosto
102 8 n. 2. dell'autore del traduttore
» 10 » 1603 1063
133 2 tributo. tributo annuale.
136 25 s'addimandò fino al
1860
s'addimanda ancora
169 1 n. 1. vol. II, p. 139, 367 vol. II, pag. 139, 355,
segg. e 547
» 2 » vol. III, p. 80, 81 vol. III, pag. 80, 81,
158.
173 9-10 n. figliuolo o nipote nipote o bisnipote
181 4 » 612. 618.
206 8 Pacione. Dond'e' Pacione, Mohammed-
Ibn-Coco. Dond'e'
219 3 Lentini e i nomi Lentini e Ragusa, e i
nomi
» 2 n. 3. secolo. secolo. Per Ragusa si
vegga Amico,
_Dizionario
topografico_, sotto
quel nome.

220 12 n. Firenze. Firenze alle radici di
Monte Morello ed
un'altra presso Bagno
a Ripoli. V'ha anco un
_Paterno_ in provincia
di Roma, presso
Albano
305 5 1093, alle quali 1093 e Malta nello
stesso tempo, com'e'
pare, alle quali

NOTE:
Si vegga il Libro IV, cap. VI, pag. 311, del vol. II.
Chronicon Pisanum, presso Muratori, Rerum Italicarum Scriptores, tomo VI, p. 101, e
Breviarium pisanæ historiæ a p. 167; e Marangone, nell’Archivio Storico Italiano, tomo
VI, parte II, pag. 4, tutti nell’anno pisano 1005. Il Breviarium, compilato alla fine del
XIII secolo, aggiugne che i Saraceni avevano minacciato Roma, fatto poco probabile,
finto com’io credo per vantare i meriti dei Pisani appo la corte papale e rincalzare la
supposta concessione della Sardegna. I compilatori pisani più moderni mano mano
confusero la narrazione, ponendo questo assalto lo stesso anno della battaglia di
Reggio, e proprio nell’assenza dell’armata; poi la scena si ravvivò con Mogêhid
(Musetto), con la Chinzica eroina, con le esortazioni del Papa, le arringhe dei consoli
pisani, i quali furono supposti con date, nomi e cognomi ec. Si veggano cotesti romanzi
nel Sardo, Cronaca Pisana; e nel Roncioni, Storie Pisane, nell’Archivio Storico Italiano,
tomo VI, parte II, pag. 76, e parte I, pag. 49, 51, e si riscontri il Muratori, Annali
d’Italia, 1005, il quale con sana critica rigetta tutti quegli episodii. Quanto all’origine
arabica del nome Chinzica, supposta dal Muratori, mi accordo col Wenrich che la mette
in forse. Rerum ab Arabibus ec., lib. I, cap. XIII, § 115. In ogni modo quella voce non
ha che fare coll’avvenimento del 1004, poichè le carte pisane innanzi il mille fanno
menzione d’un quartiere di tal nome. Si vegga l’avvertenza dei dotti editori del
Roncioni, op. cit., pag. 63, nota 1.
Quel che si sa della battaglia di Reggio è stato riferito da noi nel Libro IV, cap. VII, pag.
341, del vol. II. La supposizione della pia gesta dei Pisani è nata in questo modo. I
Benedettini della congregazione di Saint Maur pubblicarono tra le epistole di Gerberto
(Recueil des Historiens des Gaules, tomo X, pag. 426, nº. CVII) una del 999, indirizzata
non si sa a chi e molto oscura, nella quale il Papa, lamentando Gerusalemme profanata
dai Pagani, esorta lo sconosciuto cristiano: «Enitere ergo, miles Christi, esto signifer et
compugnator, et quod armis nequis, consilii et opum auxilio subveni;» nelle quali parole
in vero si trova l’idea immatura d’una crociata e la domanda di oblazioni per la santa
impresa. I dotti editori aggiungono in nota che i Pisani subito si messero in mare e
andarono a combattere. Si cita per questo, Muratori, Rerum Italicarum Scriptores, III,
400, ma in fondo non si trova altra fonte che un moderno panegirico municipale dei più
avventati, voglio dir le lunghissime note di Costantino Gaietani alle vite dei papi di
Pandolfo Pisano, pubblicate a Roma il 1638, e ristampate dal Muratori nel detto volume.
Torniamo dunque al Tronci e peggio, e si spezza il legame tra l’epistola di Gerberto del
999 e la battaglia di Reggio del 1005, si dilegua la crociata, e resta ai Pisani la
industria, la civil prudenza, e la virtù di guerra navale.

Chronicon Pisanum; e Marangone, II. cc., anno 1012.
Ne’ Mss. d’Ibn-el-Athîr si legge erroneamente Abu-Hosein, per uno scambio di lettere e
punti diacritici molto facile ad avvenire nelle copie. Abu-l-Geisc (Padre dell’esercito)
significa il soldato per antonomasia.
Rumi. Così il chiama Marrekosci, The history of the Almohades, testo arabico, pag. 52.
Può significare schiavo greco o italiano, e, in Spagna, uom delle schiatte sottomesse dai
Musulmani.
Almansor si chiamava Ibn-abi-Amir.
Dhobbi, Ms. della Soc. Asiat. di Parigi e Ibn-Bassâm, Ms. della Bibl. di Gotha, entrambi
all’articolo Mogêhid. Debbo questi estratti alla cortesia, l’uno del Prof. Dozy di Leyda, e
l’altro del Dottor Weil di Heidelberg. Ibn-el-Athîr dice che Mogêhid e il figliuolo Alì, suo
successore, furono entrambi «uomini di dottrina, amicissimi e benefici verso i dotti, cui
ricercavano nei paesi vicini e lontani.» Marrekosci fa le stesse lodi del solo figlio. La
voce ch’essi usano (’ilm) è in generale, scienza, ma più specialmente il diritto con sue
vaste ramificazioni. Dell’articolo di Dhobbi ho data una versione italiana nella Nuova
Antologia di Firenze, maggio 1866, vol. II, p. 61. Si vegga anco Ibn-Khaldûn,
Prolegomeni, testo arabico, Parte II, nelle Notices et Extraits, tomo XVIII, p. 389, e
Makkari, Analectes de l’histoire de l’Espagne, testo arabico stampato a Leyda, Vol. I, p.
280, 523, 524 e vol. II, 117, 129, 415, 433, 511, 526, dove sono narrati alcuni
aneddoti, della generosità di Mogêhid verso illustri filologi.
Ibn-el-Athir, ediz. Tornberg, tomo IX, p. 205, anno 407, nel cenno su i piccioli Stati che
nacquero in Spagna. Ho data la traduzione italiana nella Nuova Antologia di Firenze,
vol. II, p. 60, maggio 1866. Uno squarcio del testo si legge nella mia Biblioteca Arabo-
Sicula, pag. 271. Questo Capitolo con poche varianti è trascritto da Nowairi, Ms. di
Parigi, A. F., 647, fog. 108 recto; il quale chiama Mo’aiti Abu-Mohammed-Abd-Allah.
Quanto ai principii della signoria di Mogêhid a Denia, seguo piuttosto il racconto
verosimile dell’annalista musulmano, che quello del Conde, Dominacion de los Arabes
en España, cap. CIX, il quale del nome proprio Mogêhid, fece un titolo Mogêhid-ed-din
“Guerrier della Fede:” ma ciò non si adatta alle usanze di Spagna in quel tempo.
Marrekosci, loc. cit., dà appena il nome e pochissimi cenni di Mogêhid. Egli attribuisce
al costui figlio Alì, successore suo nel principato di Denia e Majorca, il titolo di Mowaffek
“Favorito (da Dio)” che Ibn-el-Athîr, Dhobbi, Nowairi e Conde danno a Mogêhid stesso,
e ch’egli forse prese quando restò solo signore, dopo la morte di Mo’aiti.
Si vegga il Libro I, cap. VII, e X, nel vol. I, pag. 170, 175, 227 e il Libro III, cap. VIII,
vol. II, pag. 180. Le scorrerie dell’816, e 817, si ritraggono da Ibn-el-Athîr nella Bibl.
Arabo-Sicula, pag. 221, 228, del testo. Entrambe mossero d’Affrica. Nella prima non
pochi Musulmani, dopo aver fatto preda, si perdettero per fortuna di mare. Quegli
andati alla seconda impresa «or vinsero, or furono vinti, e se ne tornarono.»

Così leggiamo in Edrisi, autore del XII secolo, nella Biblioteca Arabo-Sicula, testo, pag.
20 e 21, e presso Di Gregorio, Rerum Arabic., p. 112. Il passo relativo ai Sardi, ch’è
mutilato nella Geographia Nubiensis, seguita dal Di Gregorio, corre così: «Gli abitatori
della Sardegna sono di origine Rûm-Afarika, berberizzati, nemici di ogni altro ramo della
schiatta dei Rûm: uomini prodi e di saldo proponimento che non lascian mai l’armi.»
L’appellazione Rûm, nota ai nostri lettori, qui significa evidentemente gente italiana. Gli
Afarika erano le popolazioni cristiane dell’Affrica, di schiatta fenicia, come accennammo
nel Libro I, cap. V, tomo I, pag. 105. Berberizzati non può qui significar altro che misti
coi Berberi; e ci ricorda i notissimi Barbaricini dei tempi di San Gregorio in Sardegna.
Ibn-el-Athîr sotto l’anno 92 (710-11) raccoglie la storia di tutte le scorrerie dei
Musulmani in Sardegna, in unico capitolo, del quale io ho pubblicato il testo nella
Biblioteca Arabo-Sicula. Quivi si legge a pag. 217 «L’anno 135 (752-3) osteggiò
quest’isola Abd-er-Rahmân-ibn-Habib-ibn-abi-’Obeida-el-Fihri, il quale vi fe’ grande
strage. Ma poi fermò pace con gli abitatori, a patto che pagassero la gezia; la quale fu
riscossa e durò. Nè altri dopo Abd-er-Rahmân molestò quest’isola; talchè i Rûm
ristorarono le cose di quella.» Accennato poi alla scorreria del 935 e in ultimo
all’impresa di Mogêhid del 1016, avverte in fine: «nè fu mai più combattuta la Sardegna
(dai Musulmani) dopo questo tempo.» In questo capitolo Ibn-el-Athîr dimentica le
fazioni dell’816 e 817 ch’ei narra altrove come si è accennato. La menzione che si fa dei
Giudici di Sardegna nell’865 (veggasi Muratori, Dissertat. Antiq. Ital. medii ævi, II, p.
1077, Diss. XXXII) si attaglia, come dicemmo, alla testimonianza d’Ibn-el-Athîr. Si vegga
anco Manno, Storia di Sardegna, lib. VII, pag. 333 e seg. dell’ediz. di Capolago, 1840,
vol. I, e Wenrich, Rerum ab Arabibus etc., lib. I, cap. XIII, § 112, 113. Questi due
diligenti compilatori avrebbero smesso ogni dubbio, leggendo il citato capitolo d’Ibn-el-
Athîr.
Breviarum, ec., presso Muratori, Rerum Italicarum Scriptores, tomo VI, pag. 167, anno
pisano 1002. Marangone nè l’altra cronica non ne fanno menzione, e la data mal si
accorda con quella, sì precisa, degli autori arabi.
Si riscontrino: Ibn-el-Athîr nei citati due capitoli del 92, e del 407, nella Biblioteca
Arabo-Sicula, pag. 218, e 271; Ibn-Khaldûn, Prolegomeni, testo, nella Biblioteca Arabo-
Sicula, pag. 461, e nelle Notices et Extraits des MSS., tomo XVII, parte I, pag. 36;
Makkari, Mohammedan Dynasties in Spain, versione inglese del prof. Gayangos, tomo
II, pag. 258; Conde, l. c.
Si riscontrino: Ditmar, Chronicon, lib. VII, cap. 31, presso Pertz; Scriptores, tomo III,
pag. 830; Marangone, nell’Archivio Storico Italiano, tomo VI, parte II, pag. 4;
Chronicon Pisanum e Breviarium presso Muratori, R. I. S., tomo VI, pag. 107, 167, sotto
l’anno pisano 1016; e il poema di Lorenzo Vernese, presso Muratori, stesso volume,
pag. 124, dove si accenna che Mugeto l’anno innanzi la sconfitta finale (cioè 1016, del
conto comune) s’era dato alla fuga vedendo venire l’armata pisana. Le croniche pisane
laconicamente portano che i Pisani e Genovesi, fatta guerra in Sardegna con Mugeto, il
vinsero. Ditmar vescovo di Mersebourg, morto il 1018, scrisse in fin della sua cronica in
luogo che risponde al 1016, come i Saraceni venuti con l’armata in Longobardia

occupavano «Lunam civitatem;» cacciatone il vescovo s’impadronivano delle case e
mogli de’ terrazzani; come papa Benedetto chiamava alle armi i rettori e difensori della
Chiesa; come il grande navilio ch’egli adunò stringeva i Saraceni nel porto. Il re allor
fugge in barchetta; i suoi assaliti da’ Cristiani, per tre dì hanno l’avvantaggio; poi sono
rotti e passati a fil di spade; presa la regina e troncatole il capo, il papa vuol per sè la di
lei corona d’oro gemmato, e manda all’imperatore mille libbre d’oro per parte del
bottino. Ma il re saraceno facea dono al papa d’un sacco di castagne minacciando di
tornare con altrettanti uomini; Benedetto gli rimandava il sacco pieno di miglio
aggiungendo: tanti uomini e più troverai vestiti di corazze per accoglierti. E il cronista,
come scandalezzato di così fatta risposta, conchiude: Iddio giudica gli uomini; e noi
preghiamolo che allontani tal flagello da quel paese, e gli accordi la pace.
Or ognun vede che si tratti d’unico fatto, di cui Ditmar scrisse le novelle che correano in
Germania, cioè l’insulto degli Infedeli sopra una città imperiale, e la vendetta che
n’avean presa i sudditi dell’imperatore; e i cronisti pisani notarono quel che loro
premea, cioè la vittoria del navilio italiano. E però il primo ristringe il fatto a Luni; i
secondi lo pongono in Sardegna; ai quali dobbiam credere come meglio informati,
ancorchè non contemporanei. Tanto più che Ditmar, con quella fuga del re, prigionia
della moglie, e data del 1016, ci mostra aver confuso le fazioni di questo e del 1015,
come or or si vedrà nei racconto della fuga secondo gli autori arabi. Da un’altra mano
non si può supporre che Ditmar abbia sbagliato il nome della città e provincia assalita.
Dunque i Musulmani al tempo dell’impresa di Sardegna fecero una scorreria a Luni,
prima o dopo la vittoria sopra Malôt, credo piuttosto prima che dopo; i Pisani e
Genovesi gli diedero una rotta navale nello stesso anno 1015 e un’altra nella state del
1016.
Marangone e le altre Croniche Pisane, dicono «homines Sardos vivos in cruce murare.»
Lo spiega Lorenzo Vernese, narrando che Mogêhid, nel fabbricare una sua fortezza,
adoperava i Sardi da manovali, e poi li facea seppellir vivi dentro le mura.
Marangone e Croniche Pisane. Dhobbi nella biografia citata di sopra dice che Mogêhid
“occupò la maggior parte della Sardegna ed espugnò le fortezze.”
Dhobbi, Conde.
Conde e le Croniche Pisane.
La data si ritrae da Ibn-el-Athîr, che nota Mogêhid scacciato dalla Sardegna in su la fine
del quattrocentosei (8 giugno 1016). Lo stesso autore in altro luogo lo dice combattuto
e sconfitto. Le croniche Pisane accennan solo alla fuga, ma Lorenzo Vernese afferma:
«Rex fugisse (fugæ sese?) datur, multis jam marte peremptis; Barbarus abscessit,
capto cum coniuge nato»
Dhobbi, loc. cit. e Conde, il quale lo copia inesattamente.
Ibn-el-Athîr.

Lorenzo Vernese, il quale aggiunge un lungo racconto sul riscatto del figliuolo.
Si riscontrino i due citati capitoli d’Ibn-el-Athîr, anni 92 e 407, nella Biblioteca Arabo-
Sicula, pag. 218 e 271; Dhobbi, l. c. il quale narra alcuni particolari della sconfitta con
le parole di un testimonio oculare; Nowairi, Storia di Spagna, l. c.; Ibn-Khaldûn, loc.
cit., il quale dice che i Cristiani «ripigliarono immantinenti la Sardegna;» Conde,
Dominacion ec., parte II, cap. 110; Marangone nell’Archivio Storico, vol. cit., p. 4; e il
Chronicon Pisanum, e il Breviarium ec. presso Muratori, Rerum Ital., tomo VI, pag. 107
e 167, sotto l’anno pisano 1017. Lorenzo Vernese, autore del XII secolo, nel poema su
la impresa di Majorca del 1114, presso Muratori, Rer. Ital. S. VI, 124, racconta in versi
la guerra di Sardegna come l’avea intesa da’ vecchi della sua città, e s’accorda bene
con gli annalisti arabi. «Mugelus rex Baleæ et Dianæ» (Denia e le Baleari; gli altri
Pisani, anche Marangone, lo suppongono Africano) occupa la Sardegna. Vengono i
Pisani con l’armata ed egli fugge (probabilmente nelle parti occidentali dell’isola). Torna
l’anno appresso nel regno Calaritano con suoi Mori e fabbrica una fortezza. Incrudelisce
nei Cristiani. Assalito dalle armi di Pisa, fugge di nuovo lasciando prigioni il figlio e la
moglie; e i principi dell’isola rimangon sudditi dei Pisani.
Marangone, Chronicon Pisanum, e Breviarium ec., ll. cc.
A tal concetto mi portano i pochi fatti che abbiamo della Storia di Sardegna nell’XI e XII
secolo, i quali si leggono nel Manno, op. cit., lib. VII. Lorenzo Vernese nel luogo citato
del suo poema scrive:
Erepti Sardi jugulis, tutique fuerunt;
Indeque tota manent Pisanis subdita regno.
Sardiniæ: docuere senet quæcumque retexo;
Quæsitis Sardis, non hæc tibi vera negabunt.
Le quali parole, con le testimonianze non richieste che allega il poeta, mostrano che
nella prima metà del XII secolo i Pisani non pretendeano per anco la piena signoria
della Sardegna, ma un protettorato con gli abusi che ne seguitano. D’altronde non si
comprenderebbe in qual altro modo avrebbero potuto signoreggiare in Sardegna i nobili
e mercatanti che non governavano per anco Pisa. E si veggono molto più antichi della
fuga di Mogêhid, i giudici che Benvenuto da Imola, presso Muratori, Antiq. Ital. Medii
Ævi, tomo I, p. 1089, secondo le idee del XIV secolo, supponeva istituiti dai Pisani. La
concessione dell’isola per Benedetto VIII è invenzione del XIII secolo, quando la corte
di Roma avea dato lo scandalo di infeudare a questo ed a quello la Sicilia e la Sardegna
stessa; nè alcuno ha prodotto mai il testo di quel privilegio; nè lo si allegò mai nelle
contese fra i Genovesi e i Pisani presso Federigo Barbarossa, le quali si leggono
distintamente nella continuazione di Caffari, anno 1164, presso Muratori, Rerum
Italicarum Scriptores, tomo VI, p. 294, 295.
È da avvertire che il Saint Marc, Abrégé chronologique de l’histoire d’Italie, anni 1017 e
1021, tenendo per guida il Muratori, nega la concessione papale e la dominazione
pisana, senza particolareggiare gli argomenti.

Il Manno (tomo I, p. 381, dell’edizione di Capolago) non osa troncare la difficoltà nè
rigettare apertamente la narrazione riferita dal Gaietani nelle annotazioni alle vite dei
Papi (Muratori, Rerum Italicarum Scriptores, tomo III, p. 401); il quale, nel 1638,
affermava averla tolto da Lorenzo Bonincontro da San Miniato che scrisse, dice egli, più
di dugent’anni addietro. Bonincontro o Gaietani, dava con nomi e cognomi, la divisione
della Sardegna tra Pisani, Genovesi e Spagnuoli dopo la sconfitta e prigionia di Musetto.
Basterebbe la menzione delli Spagnuoli, per dimostrarla fattura del XV secolo.
Caffari, Annales Januenses; e continuazione presso Muratori, Rerum Italicarum
Scriptores, tomo VI, anni 1162 e 1164; Marangone nell’Archivio Storico Italiano, tomo
VI, Parte II, p. 38, anno 1165. Su le guerre tra quelle due città si vegga Marangone,
op. cit., p. 8 e segg., fin dal 1119 (1118). Si vegga anche il Manno, Storia di Sardegna,
lib. VII.
Cotesta falsa tradizione nacque nel XIII secolo, trovandosi nel Breviarium ec., presso
Muratori, Rerum Italicarum Scriptores, tomo VI, p. 167, anni 1017, 1020, 1050, non già
nelle due croniche del XII secolo, cioè l’anonima del Muratori e quella di Marangone. I
Genovesi a lor volta nella lite del 1164 affermavano audacemente dinanzi il Barbarossa
che i lor maggiori avessero preso il Muzaito e il vescovo di Genova lo avesse mandato
all’imperatore.
Ibn-el-Athîr, capitolo dell’anno 92, nella Biblioteca Arabo-Sicula, testo, p. 218. Ibn-
Khaldûn riferisce altre scorrerie degli Ziriti d’Affrica nel regno di Iehia-ibn-Temîm (1108
a 1116), Biblioteca Arabo-Sicula, testo, p. 482, e Histoire des Berbères, versione di M.
de Slane, tomo II, p. 25.
Ibn-el-Athîr, ediz. Tornberg, tomo IX, p. 205, anno 407.
Ademari Cabanensis Chr., nel Rec. des Hist. des Gaules, X, 156.
Gayangos, The Moham. dynasties in Spain, tomo II, p. LXXXVIII. Dozy, Hist. des
Musulmans d’Espagne, tomo IV, p. 290, 304, Cf. p. 21 della stessa opera e Dozy
medesimo, Recherches, 2ª ediz. I, 245.
Così nell’impresa del 1035 che si ritrae da Rodolfo Glabro e che or si narrerà. Si è
veduto che i Genovesi nel 1164 davano lo stesso vanto ai lor maggiori. Le supposte
imprese del 1019 e 1049 nella compilazione pisana del XIII secolo provano che durasse
la terribile leggenda di Mogêhid. È da notare che, all’infuori del poeta Lorenzo Vernese,
tutti supponeano Mugeto re d’Affrica. Quest’errore è durato fino al Manno. Il Wernich,
Rerum ab Arabibus in Italia ec., lib. I, cap. XIII, § 113 a 119, rattoppa col supposto che
Mogêhid fosse il principale dei regoli musulmani di Sardegna e che avesse chiesto aiuti
in Affrica. Del resto ei segue la tradizione pisana; se non che riconosce l’identità del
fatto di Luni e della prima vittoria dei Pisani e Genovesi.
Si vegga il Libro IV, cap. VIII, pag. 364 del vol. II.

Marangone, nell’Archivio Storico Italiano, vol. cit., p. 5, anno pisano 1035; Chronicon
Pisanum, stesso anno, presso Muratori, Rerum Italicarum Scriptores, tomo VI, p. 108. Il
Breviarium, nello stesso volume del Muratori, p. 167, finge la occupazione di Cartagine
e le corone dei due re, di Bona e Cartagine, mandate in dono dai Pisani all’imperatore.
Rodolfo Glabro, Historiarum, lib. I, cap. VII, nel Recueil des Historiens des Gaules ec.,
tomo X, p. 52, narra che i Saraceni d’Affrica perseguitavano i Cristiani per terra e per
mare; che entrambi si accordarono di combattere giuste battaglie; che i Cristiani
vinsero con grande strage, dicendosi anche ucciso il principe saraceno Motget; e che
ragunate le preziose armadure nemiche del prezzo di parecchi talenti d’argento, le
dettero per voto a Odilone abate di Cluny, il quale investì il valsente in arredi sacri e
limosine. Rodolfo era contemporaneo e famigliare degli abati di Cluny; ma testa
bislacca e gran contatore di favole. L’offerta votiva al monastero mi fece pensare
dapprima a un’impresa di Provenzali, ma fattone parola al savio autore delle Invasions
des Sarrazins en France, mi ha convinto che questa fazione, di certo navale, non potè
compiersi se non che da armate italiane. Però suppongo il voto di qualche ausiliare
provenzale ed una delle solite esagerazioni di Rodolfo Glabro. Si tratta probabilmente
dell’assalto di Bona, e vi risponde la data, poichè Rodolfo non osservando l’ordine
cronologico, pone questo fatto tra la morte di Roberto duca di Normandia (22 luglio
1035) e la ecclissi solare del 29 giugno 1033. Nelle Invasions des Sarrazins en France,
p. 221, il dotto autore, M. Reinaud, accettò che Mogêhid fosse il condottiero dell’armata
vinta; ma so ch’egli sarà per considerare il fatto altrimenti sulla nuova edizione che
apparecchia.
Par che la prima denominazione indicasse particolarmente gli uomini di Norvegia, e la
seconda quei di Danimarca. Ma spesso si confondeano gli uni con gli altri. Come ognun
sa, in Francia si chiamarono Normanni, e in Inghilterra Dani, tutti gli occupatori
scandinavi.
Questa impresa intessuta di moltissime favole si legge in Dudone di Saint Quentin, De
Moribus Normannorum, cap. I, presso Duchesne, Historiæ Normannorum Scriptores, p.
64, 65; Guglielmo di Jumièges, Historia Normandiæ, lib. I, cap. X, XI, ib., p. 220, 221;
Benoit, Chroniques des ducs de Normandie, in versi francesi, tomo I, p. 47 a 69; Wace,
Roman du Rou, versi 472 a 732. Si vegga anche Muratori, Antiquitates Ital. Medii Ævi,
tomo I, p. 25, e si riscontri la critica del fatto in Depping, Histoire des Expéditions
maritimes des Normands, edizione del 1843, p. 140, segg.
Non occorrendo citazioni distinte dei luoghi d’opere moderne dai quali ho cavati i
primordii dei Normanni, indicherò quelle che mi sono riuscite più utili. Nel sentimento
storico ho avuto a sicura guida la Conquête de l’Angleterre par les Normands, di
Augustin Thierry, alla cui memoria debbo d’altronde amore, riverenza e gratitudine. Le
minuzie dei fatti sono fornite in abbondanza dalla citata opera di Depping; e molte
critiche avvertenze si rinvengono in Lappenberg, A history of England under the
Norman kings, versione inglese con aggiunte del traduttore Benjamin Thorpe.
Importanti e novelli fatti su la società primitiva degli Scandinavi si ritraggono dalla
prefazione di Samuele Laing alla Heimskringla di Snorro Sturleson, versione inglese.

Gli storici francesi pongono vagamente la data tra l’896 e l’898, non trovandola precisa
nei cronisti, e dovendo tenere questa occupazione come diversa da quella che i cronisti
riferiscono al 17 novembre 876, cioè avanti l’assedio di Parigi. Si riscontrino le opere
citate di Depping, lib. III, cap. III; di Thierry, lib. II; e di Lappenberg, versione inglese,
p. 7, segg. I cronisti normanni in prosa e in versi confusero le tradizioni, volendo dare a
Roll, nello assedio di Parigi e nella prima occupazione di Rouen, la parte principale che
di certo non v’ebbe.
Al messaggero di Carlo il Semplice, che innanzi la battaglia dell’898 domandava il capo
loro, i Normanni risposero: «Non n’abbiamo; siam tutti eguali».
Hrôlfr, con le mutazioni eufoniche di Rolf, Roll, Rou.
Rispondeva, secondo Depping, all’odierno dipartimento della Bassa Senna e parte di
quello dell’Eure.
Wace, Roman du Rou, passim. I Francesi vendicavansi con un calembourg, più antico al
certo del XII secolo quando visse l’autore: Francheis dient ke Normandie Ço est la gent
de North mendie, versi 119, 120.
Si vegga il Libro IV della presente Storia, cap. X, p. 580 del secondo volume.
Wace, op. cit., verso 2108, accenna le tradizioni ritmiche, le quali in sua fanciullezza
avea inteso cantare a’ giullari (jugléors, oggi jongleurs).
Dudonis super Congregationem Sancti Quintini decani, De Moribus Normannorum,
presso Duchesne, Historiæ Normannorum Scriptores, p. 56 a 59. Si vegga la critica di
Lappenberg, A history of England under the Norman Kings, versione del Thorpe, p. xx.
Guglielmo di Jumièges (Wilelmus Gemmeticensis), detto Calculus (1137); Odorico
Vitalis (1141); Wace di Jersey, Roman du Rou (1184), e molti altri che si veggano in
Lappenberg, op. cit., p. xxi a xxviii.
L’Ystoire de li Normant et la Chronique de Robert Viscard par Aimé moine du Mont-
Cassin, pubblicata da M. Champollion-Figeac, Paris, 1835. L’editore con molta sagacità
ha provato irrefragabilmente il nome e nazionalità dell’autore e la data dell’opera.
Prolégomènes, p. xxxiii, segg. M. Gauttier d’Arc aveva usato fino dal 1830 un MS.
imperfetto di Amato nella Histoire des Conquêtes des Normands en Italie ec.
Le interpolazioni che non cadono in dubbio furon messe tra parentesi dal dotto editore.
Se ne può supporre delle altre, come parmi; ed anche qua e là qualche taglio, per
esempio nell’infelice fine di Dato, lib. I, cap. XXV. Nella Cronica di Roberto Guiscardo,
della quale abbiamo il testo latino, il traduttore frantende alcune frasi, fin dai primi
righi, dove leggendo d’una dama nec minus facie quam vitæ integritate formosa,
squadernò: belle de face et de touts membres entière. Similmente parmi che nella
battaglia di Canne del 1019 Amato abbia messo il nome del luogo, là dove il traduttore
scrive: et sont veues les lances estroites come les canes sont en lo lieu où il croissent.

Urbano secondo, francese, fu papa dal 1088 al 1099; Ruggiero, figlio di Roberto
Guiscardo, regnò in Puglia dal 1085 al 1111.
L’incontro fortuito di Melo e dei Normanni al Monte Gargano mi pare episodio classico
posto a capo del poema. I fendenti di Roberto Guiscardo alla battaglia di Civitella,
vengono a dirittura dalla Tavola Rotonda. Lo stratagemma di Roberto, infintosi morto e
messosi nella bara per occupare un castello in Calabria del quale non si dà il nome, è
copia della fazione di Hastings a Luni, favola scandinava ripetuta da Dadone di San
Quintino alla fine del X secolo (presso Duchesne, op. cit., p. 64, 65) e replicata nella
saga di Aroldo il Severo, come accennammo nel Libro IV, cap. X, p. 385, 386 del
secondo volume.
Tiraboschi, Storia della Letteratura Italiana, lib. IV, cap. III, § 8, si voltò con gran
collera contro i Benedettini di Saint-Maur, i quali nella Histoire Littéraire de la France,
tomo VIII, p. 488, ci rapivano questo Guglielmo di Puglia. Il signor Ruggiero Wilmans,
tedesco, fa opera a rendercelo per varie ragioni accennate nella prefazione alla detta
cronaca presso Pertz, Scriptores, tomo IX, p. 239, e più largamente discorse
nell’Archivio Storico di Pertz, tomo X, p. 93, segg. Contuttociò Guglielmo, al nome ed
alla parzialità sua contro i Longobardi, i Greci e gli abitatori della Puglia, mi sembra
chierico venuto di Francia o nato in Italia in casa francese. Quel che parrebbe in bocca
sua biasimo de’ Normanni, si trova a tanti doppii nel francese Malaterra, e suonava lode
a usanza loro.
Il Malaterra, lib. I, cap. XXV, nota che in Calabria una volta il conte Ruggiero con
quaranta suoi fedeli masnadieri plurimum penuriarum passus est, sed latrocinio
armigerorum suorum in multis sustentabatur; quod quidem ad ejus ignominiam non
dicimus, sed ipso ita præcipiente, adhuc viliora et reprehensibiliora de ipso scripturi
sumus, ut pluribus patescat quam laboriose et cum quanta angustia a profunda
paupertate ad summum culmen divitiarum vel honoris attingerit. In fondo dunque il
vecchio conte Ruggiero se ne vantava.
Questa è la cronica che il Caruso pubblicò nella Bibliotheca Sicula, p. 827, segg., col
titolo di Anonymi Historia Sicula; indi il Muratori, Rerum Italicarum Scriptores, tomo
VIII, p. 740, segg., col titolo di Anonymi Vaticani Historia Sicula. La versione in antico
francese che se ne trovava nello stesso MS. di Amato, è stata data alla luce da M.
Champollion, op. cit., col titolo di Chronique de Robert Viscard. Non si può affatto
assentire al dotto editor francese che l’autore sia Amato stesso. Se ne dee togliere in
vero, come notava M. Champollion, tutta la parte che corre dal 1101 al 1283. Ma ciò
che precede è compilazione scritta verso il 1146, come lo mostran le parole (presso
Caruso, p. 856) Huic successit ille hominum maximus.... Rogerius.... rex Siciliæ, Tripolis
Africæ.... le cui lodi l’autore, com’ei dice, non osava intraprendere. La continuazione
comincia immediatamente dopo questo passo con le parole: Post mortem comitis
Rogerii, prout confitetur in chronica, successit Rogerius ec.
Pongo la data del 1146, poichè vi si accenna il conquisto di Tripoli, non quel di Mehdia
e di tutta la costiera che seguì il 1149. La diversità degli autori ch’io sostengo, è

provata anche dalla incompatibilità di alcuni racconti, per esempio la diserzione di
Ardoino, il tempo in cui Guglielmo Braccio di Ferro ebbe il comando di tutta la banda a
Melfi ec.
Si vegga il Libro IV, cap. VII, p. 343, segg., del secondo volume.
Tale Gilberto Drengot, o Buatère, coi fratelli Rainolfo, Rodolfo, Anquetil ed Ormondo, su
i quali si veggano: Amato, op. cit., lib. I, cap. XX; Rodolfo Glabro, Historiarum, lib. III,
cap. I, nel Recueil des Historiens de la Gaule, tomo X, p. 25; e Guglielmo di Jumièges,
lib. VII, cap. 30, presso Duchesne, Historiæ Normannorum Scriptores, p. 284. Gilberto
aveva ucciso un Guglielmo Repostel che si vantava d’avergli sedotta una figliuola. I
nomi son dati diversamente dai tre cronisti. Debbo avvertire che Amato qui dice
regnante il duca Roberto di Normandia, onde il fatto andrebbe posposto al decennio
1026-35. Ma è da supporre sbagliato il nome anzichè il tempo.
Si vegga il Libro IV, cap. VII, p. 340 e 342 del secondo volume.
Secondo il biografo di Arrigo II, Acta Sanctorum, 14 luglio, p. 760, l’imperatore elesse
Melo duca di Puglia, il quale morì a Bamberg. Lupo Protospatario, anno 1020, fa ricordo
di Melo col titolo di duca di Puglia, che probabilmente gli era stato dato dai popoli o da’
suoi partigiani in Italia. Il monaco Ademaro della nobile casa di Chabanois, nella
cronaca terminata verso il 1029, scrive che al tempo di Riccardo II duca di Normandia
un Rodolfo con molti altri Normanni andavano armati a Roma, e, connivente papa
Benedetto, assaltavano e guastavan la Puglia, vincean tre battaglie; poi sconfitti dai
Russi e altri soldati dell’impero bizantino, molti n’erano condotti prigioni a
Costantinopoli; e che per tre anni i Bizantini, per rancore o sospetto de’ Normanni,
vietarono ai pellegrini occidentali il passaggio di Gerusalemme, senza dubbio per l’Italia
meridionale. Nel Recueil des Historiens des Gaules, ec., tomo X, p. 156, Rodolfo Glabro,
che scrisse verso il 1044, narra le prime imprese dei Normanni in Italia in questo modo:
che il guerriero Rodolfo perseguitato da Riccardo di Normandia, andava a Roma; si
appresentava a papa Benedetto; era confortato da lui a combattere i Greci nell’Italia
meridionale; cominciava gli assalti; era rinforzato di innumerevoli Normanni vegnenti
alla spicciolata con piacere del conte Riccardo; guadagnava due battaglie; ma dopo la
terza, vedendo scemati i suoi, andava a chiedere aiuti all’imperatore ch’indi passò in
Italia (1022). Dunque in Francia, una ventina d’anni dopo, si attribuiva al papa l’origine
di questa guerra. Si vegga la storia di Glabro, lib. III, cap. I, nel Recueil des Historiens
des Gaules ec., tomo X, p. 25, 26. Il guerriero Rodolfo è un de’ fratelli di Gilberto, di cui
dicono Amato e Leone d’Ostia.
I cronisti non dicono espressamente di due fazioni a Bari, se non che nella guerra del
1051 e nell’assedio del 1071, quando l’occuparono i Normanni. Ma i casi di Melo,
seguito dai Baresi, poi abbandonato, costretto a fuggire, e la moglie e il figliuolo di lui
mandati dai cittadini a Costantinopoli, mostrano incominciate fin dal principio del secolo
quelle fazioni che pur erano inevitabili. La plebe doveva essere amica dei Bizantini, e i
nobili nemici.

Amato, lib. I, cap. XX, e Leone d’Ostia che lo copia, lib. II, cap. 37, dicono con molta
brevità che i Normanni, invitati già a venire in Italia dal principe di Salerno, incontraron
Melo a Capua, e che les coses necessaires de mengier el de boire lor furent données,
de li seignor et bone gent de Ytalie. Il velo è molto trasparente. Guglielmo di Puglia, sia
per render omaggio alle Muse, sia perchè la corte di Guiscardo dopo la iniqua
occupazione di Salerno non amava a sentirsi ripetere che i principi di Salerno avessero
chiamato i primi Normanni, esordisce dall’incontro fortuito dei pellegrini al santuario di
Monte Gargano con uno straniero vestito di strane fogge, il quale scopre sè esser Melo,
e agevolmente li persuade a far venire lor compatriotti ai suoi stipendii. Questo par di
tutto punto un episodio poetico, contrario alla tradizione di Amato.
Leon d’Ostia, lib. II, cap. 37.
Si riscontrino: Amato, lib. I, cap. XXI, segg.; Guglielmo di Puglia, lib. I; Lupo
Protospatario, anni 1017 a 1019; Annales Beneventani, 1017, presso Pertz, Scriptores,
tomo III, p. 178; Leone d’Ostia, lib,. II, cap. 37, 38. I cronisti non si accordano sul
numero delle battaglie vinte dai Normanni, e Amato solo narra la seconda sconfitta. Il
traduttore di Amato, non comprendendo bene il testo, nel cap. XXII, suppone che
tremila Normanni fossero venuti di Salerno dopo la battaglia di Canne; ma parmi
inverosimile, e da correggersi come ho fatto.
Si riscontrino: Amato, lib. I, cap. XXIV, segg., e lib. II, cap. I a VII; Guglielmo di Puglia,
lib. I; Lupo Protospatario, anno 1021, segg. Il Malaterra, tacendo le imprese dei
Normanni prima della venuta di Guglielmo di Hauteville, spiega pur molto precisamente
nel lib. I, cap. VI, l’indole delle compagnie normanne innanzi il 1040.
Dopo la battaglia di Canne (1019) scrive Amato: Et de li Normant non remainstrent se
non cinc cent et vj grant home de li Normant remainstrent, de liquel ij remainstrent
avec Athenulfe ec., lib. I, cap. XXII. L’Imperatore Arrigo I, nel 1022, avea lasciato in un
castello dei nipoti di Melo ventiquattro cavalieri normanni capitanati da un Trostaino.
Amato, lib. I, cap. XXIX e XXXII. Nel 1040 i 300 Normanni venuti d’Aversa in aiuto
d’Ardoino, ubbidivano come innanzi diremo a dodici condottieri uguali tra loro. Dunque
nel primo caso una compagnia somma ad 80 cavalli, e nei due secondi a 25.
Libro IV, cap. X, p. 380 e 389, segg., del secondo volume.
Si ricordino le fazioni di Rayca accennate da noi nel Libro IV, cap. VII, p. 345 del
secondo volume.
Si veggano gli Annali di Bari, e Lupo Protospatario, anni 1039, 1040 e 1041, in Pertz,
Scriptores, tomo V, p. 56, 57.
Et vous i habitez comme la sorice qui est en lo pertus.... que sachiez que je vous
menerai à homes feminines, c’est à homes comme fames, liquel demorent en moult
riche et espaciouse terre. Amato, lib. II, cap. XVII, p. 43.
Cum terra sit utilitatis,

Fœmineis Græcis cur permittatur haberi?
Guglielmo di Puglia, lib. I.
Amato: Et estut li conte (il conte) xij pare à liquel ec. Cap. XVIII, p. 43. Guglielmo di
Puglia... comitatus nomen honoris Quo donantur erat.
Amato, lib. II, cap. XIX, p. 44.
Et quant il oïrent ensi parler Arduyne, se consentirent à lui et font sacrement de fidelité
de chascune part de paiz se la terre non avoit autre seignor que ou à cui face tribut se
clame tributaire. Et en ceste regne se clame terre de demainne et se a autre seignorie
se clame colonie come sont en ceste regne la terre qui a autre seignorie. Et sanz lo roy
estoit seignor Arduyne et en celle part se clament colone. Amato, lib. II, cap. XIX, p.
44, 45. Il passo che ho notato in caratteri tondi è guasto al certo, e ciò che segue è
nota interpolata dal traduttore, spiegando a suo modo il diritto pubblico napoletano del
XIII secolo; poichè Amato non potea scrivere nell’XI le voci regno e re. Leone d’Ostia
tralascia questo importantissimo fatto, e però non possiamo ristabilire il testo d’Amato.
Ma il significato necessariamente è che i Melfitani non ubbidissero a feudatario e non
prestassero servigi feudali, nè pagassero tributo se non che allo stato: il che dopo il
conquisto normanno si chiamò in Sicilia e in Puglia: stare in demanio.
Gli avvenimenti che ristringo in questo paragrafo, dal ritorno di Ardoino in terraferma
sino all’occupazione di Melfi, son tratti da Amato, lib. II, cap. XIV, segg.; Guglielmo di
Puglia, lib. I, Aversam subito venit Hardoinus; Malaterra, lib. I, cap. VIII; Leone d’Ostia,
lib. II, cap. LXVI; Cedreno, tomo II, p. 545 della edizione di Bonn; Annali (ossia
anonimo) di Bari e Lupo Protospatario, anni 1040, 1041. Oltre le discrepanze di minor
momento, se ne scorge una che occorre di notare. Amato, seguendolo Leone d’Ostia,
dice che Ardoino dopo l’ingiuria di Maniace rimase al servigio bizantino, suscitò
occultamente i Pugliesi, e andò ad Aversa pretestando un viaggio di devozione a Roma.
Guglielmo di Puglia lo fa insultare e rivoltare a Reggio, e correr di lì dritto ad Aversa.
Malaterra, con poco divario, reca l’ingiuria in Sicilia, l’aperta ribellione appena ripassato
il Faro, e non parla punto degli aiuti d’Aversa. Nelle due tradizioni dunque, la prima
d’Amato e Leone, la seconda di Guglielmo e Malaterra, si dà essenzialmente diverso il
modo e tempo dell’ammutinamento di Ardoino con la banda normanna. Or covaron essi
l’onta parecchi giorni, o parecchi mesi? Chiarironsi disertori nel novembre 1040 in
Calabria, ovvero nei principii del 1041 a Melfi? Guglielmo di Puglia fin dà il numero di
cinquanta soldati uccisi dai Normanni alla schiera bizantina mandata a inseguirli,
quando lasciarono il campo a Reggio. Amato, all’incontro, particolareggia la
dissimulazione di Ardoino: com’ei corruppe Doceano con molt’oro; come fu preposto al
governo di parecchie terre in Puglia; come incominciò ad accarezzare e convitare i
maggiori cittadini, a compiangere gli aggravii della dominazione greca, a promettere
che farebbe opera a liberarli; come infine tolse commiato, sotto specie d’andare alle
perdonanze a Roma, e andò ad Aversa.
Or dovendosi necessariamente tacciare di bugia l’una o l’altra tradizione, ammettendo
anche la sincerità di chi la scrivea, le condizioni dei due cronisti e l’indole di loro opere

accusano Guglielmo, anzi che Amato. Del Malaterra non parlo, il quale in questo
periodo ripeteva un romanzo di casa Hauteville, tacea gli aiuti di Aversa, facea capitano
dei Normanni Guglielmo Braccio di Ferro, che lo fu tre anni dopo. Quella fuga inoltre
con le armi alla mano dal centro della Sicilia secondo Malaterra, e da Reggio secondo
Guglielmo di Puglia, infino a Melfi, è molto men credibile che la prolungata simulazione
dei Normanni e che il favor di Doceano ad Ardoino, non disertore ma guerriero
ingiuriato ingiustamente da Maniace. Infine il fatto riferito da Lupo e dagli Annali
Baresi, che Doceano tornava di Sicilia di novembre 1040 per domare i sollevati di
Puglia, dà luogo al supposto che i Normanni passassero con le forze di Doceano e
fossero da lui posti a presidio in qualche terra non lontana da Melfi. Qual maraviglia che
a capo di cinquanta o sessant’anni questo cambiamento di guarnigione, com’or
diremmo, si raffazzonò nelle brigate dei principi e nobili normanni alla foggia che ci
rappresentano Guglielmo di Puglia, e Malaterra, esagerando il valore ed attenuando la
perfidia della passata generazione?
Pertanto mi appiglio alla tradizione d’Amato e cancello quel che scrissi in contrario nel
Libro IV, cap. X, p. 389 del secondo volume, seguendo Guglielmo e Malaterra e tutti gli
istorici moderni che loro credettero, i quali non aveano sotto gli occhi Amato. Che se
altri mi tacci di leggerezza per questo, mi spiacerà meno del ricusar testimonianza al
vero una volta ch’io ne sia convinto.
Gli Annali di Bari col privilegio del «si dice» fanno montare i Greci a 18,000 e portano
poco più di 2000 i Normanni; Lupo Protospatario li dice 3000. Senza esitare accetto
cotesti numeri anzichè quelli dei due cronisti normanni, cioè Guglielmo di Puglia che dà
700 cavalli e 500 fanti, e Malaterra che dice tondo 500 militi da una parte e 60,000
Greci dall’altra. Al par che nelle guerre di Sicilia, convien dividere per sei la cifra
dell’esercito nemico, e moltiplicare per sei quella del Normanno, quando si legga il
Malaterra.
Quanto alla data, la più parte degli storici, annalisti, compilatori ed eruditi editori, non
esclusi il Muratori e il De Meo, han messo l’occupazione di Melfi e la prima battaglia nel
1040. Il riscontro con fatti vicini e di data certa nella storia bizantina, ci mostra che si
debba seguire piuttosto gli Annali di Bari e il Protospatario, i quali scrivono 1041. Leone
d’Ostia ne fa anche espresso attestato, dicendo occupata Melfi anno Dominicæ
Nativitatis MLXI, quo videlicet anno dies paschalis Sabbati ipso die festivitatis Sancti
Benedicti (21 marzo) venit: e in vero la Pasqua cadde il 22 marzo nel 1041, non già nel
1040. Il Chronicon Breve Northmannicum, presso Muratori, Rerum Italicarum
Scriptores, tomo V, p. 871, porta anche nel 1041 la prima occupazione della Puglia pei
Normanni capitanati da Ardoino, e in marzo e maggio 1042 (dalla Incarnazione, ossia
1041 del conto comune) le due prime vittorie sopra i Greci.
Si riscontrino: Amato, lib. II, cap. XXI, segg.; Guglielmo di Puglia, lib. I, Audito reditu
Michælis, sino alla fine del Libro; Malaterra, cap. IX, X; Lupo Protospatario, ed Annali di
Bari, anni 1041, 1042; Leone d’Ostia, lib. II, cap. LXVI. L’ordine degli avvenimenti è
uguale in tutti; le date si trovan solo in Lupo e negli Annali di Bari. Contandosi da Lupo
gli anni dell’èra volgare, talvolta al modo salernitano dal 25 dicembre (Vedi Pertz,
Scriptores, tomo V, p. 51), ma più sovente col periodo costantinopolitano, cioè dal 1º

settembre dell’anno precedente, il settembre 1042 risponde al nostro settembre 1041,
e così fino a decembre. Che in questa epoca Lupo segua tal cronologia lo provano le
esaltazioni degli imperatori di Costantinopoli, le quali noi possiamo riscontrare con le
date di Cedreno e degli altri Bizantini.
Pro numero comitum bit sex statuere plateas,
Atque domus comitum totidem fabricantur in urbe.
Guglielmo di Puglia, Lib. I.
Cedreno dice espressamente: Italiani delle province tra il Po e le Alpi; Amato: Et li
Normant d’autre part non cessoient de querre li confin de principal pour home fort et
soffisant de combatre ec. Lib. II, cap. XXIII, p. 50.
Amato, ricordata l’occupazione di Melfi nel lib. II, cap. XIX, narra nel cap. XXX il
partaggio dei conquisti al conte d’Aversa e dodici altri capi normanni dei quali dà i nomi
ed i territorii assegnati a ciascuno, aggiugnendo: et (à) Arduyne secont lo sacrement
donnerent sa part c’est la moitié de toutes choses si come fa la covenance; il qual fatto
torna al 1043. Leone d’Ostia copia Amato nel lib. II, cap. 67, con le parole: Arduino
autem juxta quod sibi juraverant parte sua contradita. I nomi dei dodici oltre il conte
d’Aversa son tutti normanni. I territorii assegnati son quasi tutte città vescovili in un
triangolo curvilineo dal Gargano a Frigento e di lì a Monopoli, nel quale spazio rimane
in vero un’altra metà di luoghi importanti da potersi supporre assegnati ad Ardoino se si
conoscesse che i Normanni li aveano occupati in quel tempo.
Ma l’illustre capo non è nominato da nessun altro cronista dopo il patto di Melfi; non da
Amato nè da Leone dopo quel partaggio, nè alcuno dice che gli altri territorii di Puglia,
caduti poi tutti in potere dei Normanni, fossero stati tolti sia ad Ardoino sia a feudatarii
italiani della sua compagnia. Il modo più plausibile di spiegar cotesto silenzio mi par di
supporre la immatura morte di Ardoino e la incorporazione de’ suoi nelle compagnie
normanne. Guglielmo Braccio di Ferro che veniva di Sicilia con Ardoino, è il primo dei
dodici nominati nel partaggio, e nello stesso anno fu creato conte di Puglia, come or si
vedrà.
Guglielmo di Puglia, Lib. I, appone questa scelta d’uno straniero a corruzione e invidia
dei Normanni: Sed quia terrigenis, terreni semper honores, Invidiam pariunt ec.; ma
Amato, italiano ancorchè monaco, dice: Et à ce qu’il donassent ferme cuer à li colone
de la terre lo prince de Bonivent ec.
Si riscontrino: Amato, lib. II, cap. XXVII; Guglielmo di Puglia, lib. I, Nam reliqui Galli
ec.; Lupo Protospatario anno 1042. Secondo Guglielmo, vi fu un principio di divisione
tra i Normanni dopo la deposizione di Atenolfo, volendo alcuni ubbidire a Guaimario
principe di Salerno, ed altri ad Argiro. Ei narra la esaltazione di Argiro in Bari, richiesto
dal popolo, ricusante questa dignità innanzi i primarii cittadini che avea convocati nella
chiesa di Sant’Apollinare, sforzato dal comun voto ed eletto principe. Sembra che il
poeta voglia descrivere in qual modo fosse stato fatto duca di Puglia il cittadino al quale

i Normanni aggiunsero l’autorità di capo o protettore di lor banda. Ad una elezione
simultanea e comune dei Baresi e dei Normanni, ci sarebbero gravi difficoltà.
Lupo scrive: et mense februarii factus est Argyrus Barensis princeps et dux Italiæ; ma
non dice da chi. Il certo è che Bari in questo tempo era ribelle, nè tornò all’ubbidienza
dei Greci se non che il 1043.
Amato, lib. II, cap. XXVII. Secondo il Protospatario questo assedio cominciò in agosto
1042, e durò un mese.
Si riscontrino: Amato, lib. II, cap. XXVII, XXVIII; Guglielmo di Puglia in fine del primo e
in principio del secondo libro; Leone d’Ostia, lib. II, cap. 66; Lupo Protospatario, anni
1042, 1043 e 1046, nell’ultimo dei quali si nota che Argiro andò a Costantinopoli e
quella corte richiamò a Bari tutti gli esuli. Non potendo dunque strappare la Puglia ai
Normanni con la forza, gli imperatori d’Oriente cedeano ai voti dei popoli, salvo ad
aggravar di nuovo la mano quando lo potessero.
Si riscontrino: Amato, lib. II, cap. XXVII, segg.; Guglielmo di Puglia, lib. II dal principio;
Lupo Protospatario, anni 1042 a 1053; Leone d’Ostia, lib. II, cap. 66.
Guglielmo di Puglia, lib. I: Multa per hoc tempus sibi promittente Salerni, e segg.
Amato, lib. II, cap. XXVIII a XXX; Leone d’Ostia, lib. II, cap. 66. Le tredici città
assegnate, in Capitanata, Terra di Bari e Principato ulteriore, son oggi tutte vescovili, e
metà l’era anche avanti l’XI secolo. Si ricordi ciò che avvertii su questo partaggio nella
nota 2, p. 34.
Così dovea seguire necessariamente, ancorchè poche vestigia rimangano di quel primo
abbozzo della feudalità normanna. Di certo si vede che nei principii alcune terre furono
soggiogate per forza o per accordi; altre, quasi confederate, ritennero governo
municipale pagando soltanto un tributo o contribuzione federale, che forse rimase in
comune per supplire al mantenimento dell’esercito. In fatti Guglielmo di Puglia,
supponendo bene o male un partaggio avanti la occupazione di Melfi, scrive, lib. I:
...... undique terras
Divisere sibi ni sors inimica repugnet.
Singula proponunt loca quæ contingere sorte
Cuique duci debent et quæque tributa locorum.
Amato accenna in questo modo, lib. II, cap. XXVII, gli acquisti dei Normanni sotto la
condotta di Argiro, cioè nel 1042: et toutes les cités d’eluec entor constreigneient qui
estoient al lo commandement et à la rayson et statute que estoient; ensi alcun
voluntairement se soumettoient et alcun de force et alcun paioient tribut de denaviers
chascun an.
Così le concessioni del conte Unfredo a’ fratelli germani Roberto, Maugerio e Guglielmo,
e infine di Roberto a Ruggiero.

Si vegga qui sopra, p. 18.
Il luogo è determinato da Gauttier d’Arc, Histoire des conquêtes des Normands en Italie
ec., Paris 1830, lib. I, cap. IV, p. 64, segg.
Su le condizioni di Tancredi di Hauteville si riscontrino: Malaterra, lib. I, cap. IV e XL:
Cronica di Roberto Guiscardo, traduzione francese, lib. I, cap. I, p. 263; e testo latino
presso Caruso, p. 829; Cronica di San Massenzio, detta Chronicon Malleacense, nel
Recueil des Historiens des Gaules etc., tomo XI, p. 644; Guglielmo di Malmesbury, lib.
III, nella stessa raccolta, tomo IX, p. 187; Odorico Vitale, lib. V, presso Duchesne,
Historiæ Normannorum Scriptores, p. 584.
La cronica di San Massenzio dice la famiglia poco nota e povera; Guglielmo di
Malmesbury, Mediocri parentela ortus ec. Il Malaterra e la cronica di Roberto Guiscardo
rincalzano la nobiltà di Tancredi: præclari admodum generis — genere nobilis.
La parentela coi duchi di Normandia, affermata per lo primo da sbadati compilatori del
XIII e XIV secolo, non è ammessa ormai da alcun critico. Si vegga un’apposita
dissertazione di E. F. Mooyer stampata a Minden nel 1830 in-4, secondo la quale il
supposto si riduce a due fila debolissime, 1º che il padre di Tancredi fosse stato un dei
figli di Riccardo I, dei quali non si conoscono i nomi; 2º ovvero che Muriel figliuola
bastarda di esso Riccardo fosse la Moriella prima moglie di Tancredi. Questa opinione
par che corresse a corte di Palermo nel 1140, perchè la cronica di Roberto Guiscardo
scrive uxor nobilissima Muriella nomine.
Inaspettatamente ci verrebbe un lume dagli autori arabi, se potessimo fidarci a loro
scrittura ed erudizione. Ibn-Kaldûn in due luoghi della storia (Biblioteca Arabo-Sicula,
testo, p. 484 e 497) dà il nome del primo conte di Sicilia, Rogiar-ibn-Tankred-ibn-
Khaira, o secondo alcuni MSS. ibn-H»w»h, che par nome di donna e indicherebbe che
la casa di Hauteville vantasse la nobiltà della madre di Tancredi. Supponendo maschile
tal nome, com’anche si può, si leggerebbe Hugo, o anche Geir (chè la prima lettera
mutando il punto diacritico suona kh, h, ovvero g), e sarebbero nomi usati in Norvegia
e in Francia. Debbo questa conghiettura all’erudito orientalista norvegio signor Broch; il
quale crede suscettivo quel vocabolo della terza lezione Haby (o forse Habwu) che
rappresenterebbe, con errore facile a supporre, il nome del feudo Hauteville.
Wilhelm, Drogo, Humfried, e secondo la pronunzia francese Guillaume, Dreux,
Humfroy.
Amato, Malaterra e Leone d’Ostia, lo dicono condottiero della compagnia; ma parmi
errore volontario dei principi di casa Hauteville. Si vegga a questo proposito il Libro IV,
cap. X della presente opera, volume secondo, p. 380, nota 3, e 389, nota 1.
Si riscontrino Amato, Guglielmo di Puglia e gli altri contemporanei citati di sopra. M.
Gauttier d’Arc, op. cit., lib. I, cap. V, p. 141, sostiene che Drogone ebbe da Arrigo III
titol di duca; ma il passo ch’egli allega di Ermanno Contratto è dubbio, e il diploma a
nome di Drogone per lo meno è erroneo, come dato il 1053. Drogone era stato
pugnalato in agosto 1051.

Si veggano le autorità citate da Augustin Thierry, Hist. de la Conquéte d’Angleterre, lib.
III, anni 1048 a 1065.
Si riscontri Ermanno Contratto presso Pertz, Scriptores, tomo V, p. 132: Indigentes bello
premere, injustum dominatum invadere, hæredibus legitimis castella, prædia, villas,
domus, uxores etiam quibus libuit vi auferre, res ecclesiasticas diripere ec. Arnolfo,
Gesta episcoporum Mediolan., presso Pertz, Scriptores, tomo X, p. 10, 11, similmente
dice i Normanni a poco a poco ingrossati in Puglia, divenuti più crudeli dei Greci e più
feroci dei Saraceni. Anche ad Amato scappa di bocca qualche lagnanza quando si tratta
di Monte Cassino, lib. II, cap. XLI. E lo stesso Guglielmo di Puglia, accennando alle
pratiche con papa Leone, accerta che Argiro Veris commiscens fallacia mittit ec.
Tralascio tante altre testimonianze, perchè superflue, ovvero sospette, come per
esempio quella d’Anna Comnena.
Ferrari nostro, nella Histoire des Révolutions d’Italie, tomo I, p. 344, segg., crede
calunniati i Normanni dall’umor di reazione unitaria che allor si scatenò contro la
rivoluzione federale dei vescovi. Ancorchè io non osi, senza più lungo studio, negar nè
accettare le nuove spiegazioni della storia patria che vien proponendo quell’alto
ingegno, parmi pure di prestar fede alle precise affermazioni dei cronisti, che d’altronde
si accordano con lo esempio di tutti i conquistatori o dominatori stranieri. Il fatto dei
soprusi e quel della reazione non sono per altro incompatibili; e certo è che i Normanni,
se servirono una rivoluzione italiana, la voltarono ad utile e comodo proprio.
Epistola di Leone IX a Costantino Monomaco, presso Labbe, Concilia, tomo IX, p. 983.
Il papa dice a chiare note voler recuperare il patrimonio della Chiesa romana, voler
porre accordo tra i due imperatori che son le due braccia della Chiesa ec. Non
occorrono citazioni per gli altri fatti che sono notissimi, e dei miei giudizii può giudicare
il lettore senza altre autorità. Ho tolto il pretesto della difesa dei poveri da Amato, il
quale, lib. III, cap. XVI, XVII, tra le rimostranze di Leone IX ai Normanni, scrive: Et
quant cil de Bonivent oïrent tant de perfetion et de sanctitè de lo pape, chacerent lo
prince et soumistrent soi à la fidelitè soe, eaux et la citè. Come ognun sa, Leone avea
già scroccata Benevento al devoto Arrigo II, in cambio dei diritti su la Chiesa di
Bamberg.
Chronicon Breve Northmannicum, presso Muratori, Rerum Italicarum Scriptores, tomo
V, p. 278, anni 1045 a 1052.
Amato, lib. II, cap. XLV; e III, cap. VII. Si confronti con gli altri cronisti ch’è inutile
citare partitamente. Secondo Malaterra il castello fu quel di Scrible in Val di Crati.
Si confrontino: Amato, lib. III; Guglielmo di Puglia, lib. II; Lupo Protospatario, anno
1053; Malaterra, lib. I, cap. XII a XV; Leone d’Ostia, lib. II, cap. 84; Ermanno Contratto
presso Pertz, Scriptores, tomo V, p. 132.
Nè Amato, nè Guglielmo di Puglia, nè Leone d’Ostia, nè alcun altro cronista narrano
questa concessione, fuorchè il Malaterra nel quale leggiamo: Quorum (Normannorum)

legitimam benevolentiam Apostolicus gratanter suscipiens, de offensis indulgentiam et
benedictionem contulit et omnem terram quam pervaserant et quam ulterius versus
Calabriam et Siciliam lucrari possent, de Sancto Petro hæreditati feudo sibi et
hæredibus suis possidendam concessit, circa annos 1052. È anacronismo col 1059, e
sbaglio di nome di Leone IX con Niccolò II; o il conte Ruggiero, autor vero della
tradizione, sapendo dai fratelli le proposizioni che fecero allora i Normanni e qualche
vaga promessa del papa prigione, le costruiva dopo mezzo secolo, a disegno o per
incerta memoria, in espresso atto d’investitura. Si avverta che Amato, lib. III, cap.
XXXVI, fa menzione della profferta dei Normanni avanti la battaglia di ricevere
l’investitura e pagar censo: come avrebbe dunque passato sotto silenzio che il papa
prigione l’assentiva? Non fo caso qui della Cronica di Roberto Guiscardo, ch’è opera
della metà del XII secolo. E mi par che la epistola di Leone IX che citerò nella nota
seguente distrugga al tutto il racconto di Malaterra.
Epistola di Leone IX a Costantino Monomaco presso Labbe, Concilia, tomo IX, p. 981,
segg. Ancorchè non vi sia data, si dee porre tra il 18 giugno 1053 e il 19 aprile 1054,
giorno della morte del papa; perchè la battaglia di Civita vi è indicata in modo non
equivoco; nè si può ammettere l’opinione del Saint-Marc, Abrégé chronologique, tomo
III, Parte I, p. 170, segg., che riferisce questo scritto al 1051, supponendo
gratuitamente un’altra zuffa dei Normanni con soldatesche del papa. Tronca ogni
dubbio Wiberto arcidiacono di Toul, il quale nell’agiografia di Leone IX, lib. II, cap. VI,
presso i Bollandisti, 19 aprile, tomo II di quel mese, p. 663, inserisce uno squarcio della
stessa epistola per narrare, com’egli dice, con le propie parole del papa, lo scontro di
Civitatula. Aggiugne del suo i fatti che conosciamo dopo la battaglia: l’andata a
Benevento e indi a Roma, fino alla morte di Leone. Amato, lib. III, cap. XXXIX, scrive:
Et o la favor de li Normant torna à Rome à li X mois puis que avoit esté la bataille.
Amato, lib. III, cap. XLVI e XLVII. Stefano IX esaltato il 2 agosto 1057, morì il 29 marzo
1058. Amato narra ch’egli avea gettato le mani sul tesoro di Monte Cassino, per far
guerra ai Normanni.
Guglielmo di Puglia, lib. II; Malaterra, lib. I, cap. XV. Da un altro canto Amato, lib. III,
cap. XLII, segg., racconta le molestie che recavano nel principato di Salerno Unfredo, il
fratello Guglielmo e Riccardo d’Aversa.
Malaterra, l. c. Amato, che in questo periodo tocca più brevemente le cose di Puglia,
accenna verso il 1054 la venuta di Malgerio, Goffredo, Guglielmo e Ruggiero fratelli del
conte Unfredo. Questo Guglielmo era figliuolo di Tancredi per la seconda moglie
Fredesenda.

Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!
ebookgate.com

Networking Explained Second Edition Michael Gallo

More Related Content

Similar to Networking Explained Second Edition Michael Gallo (20)

Recently uploaded (20)

Networking Explained Second Edition Michael Gallo