Cs ill ch15

Networks

For many years, computers have played as important a role in

communication as they do in computation. This communication is

accomplished using computer networks. Like complex highway

systems that connect roads in various ways to allow cars to travel

from their origin to their destination, computer networks form an

infrastructure that allows data to travel from some source

computer to a destination. The computer receiving the data may

be around the corner or around the world. This chapter explores

some of the details of computer networks.

455

Chapter 15

Computer network Acollection of computingdevices connected sothat they can communi-cate and share resources

Wireless A networkconnection made withoutphysical wires

Node (or Host) Anyaddressable deviceattached to a network

Data transfer rate (alsobandwidth) The speedwith which data ismoved from one place toanother on a network

456 Chapter 15 Networks

GoalsAfter studying this chapter, you should be able to:

� describe the core issues related to computer networks.� list various types of networks and their characteristics.� explain various topologies of local-area networks.� explain why network technologies are best implemented as open systems.� compare and contrast various technologies for home Internet connections.� explain packet switching.� describe the basic roles of various network protocols.� explain the role of a firewall.� compare and contrast network hostnames and IP addresses.� explain the domain name system.

15.1 Networking

A computer network is a collection of computing devices that areconnected in various ways in order to communicate and share resources.E-mail, instant messaging, and Web pages all rely on communication thatoccurs across an underlying computer network. We use networks to shareintangible resources, such as files, as well as tangible resources, such asprinters.

Usually, the connections between computers in a network are madeusing physical wires or cables. However, some connections are wireless,using radio waves or infrared signals to convey data. Networks are notdefined only by physical connections; they are defined by the ability tocommunicate.

Computer networks contain devices other than computers. Printers, forinstance, can be connected directly to a network so that anyone on thenetwork can print to them. Networks also contain a variety of devices forhandling network traffic. We use the generic term node or host to refer toany device on a network.

A key issue related to computer networks is the data transfer rate, thespeed with which data is moved from one place on a network to another.We are constantly increasing our demand on networks as we rely on themto transfer more data in general, as well as data that is inherently morecomplex (therefore larger). Multimedia components such as audio andvideo are a large contributor to this increased traffic. Sometimes the data

Protocol A set of rulesthat defines how data isformatted and processedon a network

Client/server model

A distributed approachin which a client makesrequests of a server andthe server responds

File server A computerdedicated to storing andmanaging files fornetwork users

Web server Acomputer dedicated toresponding to requestsfor Web pages

Local-area network

(LAN) A networkconnecting a smallnumber of nodes in aclose geographic area

Ring topology A LANconfiguration in whichall nodes are connectedin a closed loop

15.1 Networking 457

Figure 15.1

Client/Server interactionResponse

Request

ServerClient

transfer rate is referred to as the bandwidth of a network. Recall that wediscussed bandwidth in Chapter 3 in the discussion of data compression.

Another key issue in computer networks is the protocols that are used.As we’ve mentioned at other points in this book, a protocol is a set of rulesdescribing how two things interact. In networking, we use well-definedprotocols to describe how transferred data is formatted and processed.

Computer networks have opened up an entire frontier in the world ofcomputing called the client/server model. No longer do you think ofcomputers solely in terms of the capabilities of the machine sitting in frontof you. Software systems are often distributed across a network, in which aclient sends a request to a server for information or action, and the serverresponds, as shown in Figure 15.1.

For example, a file server is a computer that stores and manages filesfor multiple users on a network. That way every user doesn’t need to havehis or her own copy of the files. A Web server is a computer dedicated toresponding to requests (from the browser client) for Web pages.Client/server relationships have become more complex as we rely heavilyon networks in our everyday lives. Therefore, the client/server model hasbecome increasingly important in the world of computing.

The client/server model has also grown beyond the basicrequest/response approach. Increasingly, the client/server model is used tosupport parallel processing, which is the use of multiple computers to solvea problem by breaking it into pieces as discussed in Chapter 4. Usingnetworks and the client/server model, parallel processing can be accom-plished by the client requesting that multiple machines perform a specificpart of a problem. The client gathers the responses from each to form acomplete solution to the problem.

Types of NetworksComputer networks can be classified in various ways. A local-area

network (LAN) connects a relatively small number of machines in a rela-tively close geographical area. LANs are usually confined to a single roomor building. They may sometimes span a few close buildings.

Various configurations, called topologies, have been used to administerLANs. A ring topology connects all nodes in a closed loop on whichmessages travel in one direction. The nodes of a ring network pass along

Star topology A LANconfiguration in which acentral node controls allmessage traffic

Bus topology A LANconfiguration in whichall nodes share acommon line

Ethernet The industrystandard for local-areanetworks, based on abus topology

Wide-area network

(WAN) A networkconnecting two or morelocal-area networks

Gateway A node thathandles communicationbetween its LAN andother networks

Internet A wide-areanetwork that spans theplanet

Metropolitan-area

network (MAN) Anetwork infrastructuredeveloped for a largecity


messages until they reach their destination. A star topology centers aroundone node to which all others are connected and through which allmessages are sent. A star network puts a huge burden on the central node;if it is not working, communication on the network is not possible. In abus topology, all nodes are connected to a single communication line thatcarries messages in both directions. The nodes on the bus check anymessage sent on the bus, but ignore any that are not addressed to them.These topologies are pictured in Figure 15.2. A bus technology calledEthernet has become the industry standard for local-area networks.

A wide-area network (WAN) connects two or more local-area networksover a potentially large geographic distance. A wide-area network permitscommunication among smaller networks. Often one particular node on aLAN is set up to serve as a gateway to handle all communication goingbetween that LAN and other networks. See Figure 15.3.

Communication between networks is called internetworking. TheInternet, as we know it today, is essentially the ultimate wide-areanetwork, spanning the entire globe. The Internet is a vast collection ofsmaller networks that have all agreed to communicate using the sameprotocols and to pass along messages so that they can reach their finaldestination.

Recently, the term metropolitan-area network (MAN) has been adoptedto refer to the communication infrastructures that have been developed inand around large cities. The population and needs of a metropolitan areaoften require unique attention. These networks are often implemented

Figure 15.2 Various network topologies

Ring topology Star topology Bus topology

Internet backbone Aset of high-speednetworks carryingInternet traffic

Internet service

provider (ISP) Acompany providingaccess to the Internet

15.1 Networking 459

Figure 15.3 Local-area networks connected across a distance to create a wide-area network

Gateway

Gateway

using innovative techniques such as running optical fiber cable throughsubway tunnels.

Internet ConnectionsThe Internet backbone is a term used to refer to a set of high-speednetworks that carry Internet traffic. These networks are provided bycompanies such as AT&T, GTE, and IBM. The backbone networks alloperate using connections that have high data transfer rates, ranging from1.5 megabits per second to over 600 megabits per second (using specialoptical cables).

An Internet service provider (ISP) is a company that provides othercompanies or individuals with access to the Internet. ISPs connect directlyto the Internet backbone, or they connect to a larger ISP with a connectionto the backbone. America OnLine and Prodigy are examples of Internetservice providers.

There are various technologies available that you can use to connect ahome computer to the Internet. The three most popular techniques forhome connections are a phone modem, a digital subscriber line (DSL), or acable modem. Let’s examine each of these.

Phone modem Adevice that convertscomputer data into ananalog audio signal andback again

Digital subscriber line

(DSL) An Internetconnection made using adigital signal on regularphone lines

Cable modem Adevice that allowscomputer networkcommunication using thecable TV hookup in ahome


The telephone system had already connected homes throughout theworld long before the desire for Internet connections came along. There-fore, it makes sense that the first technique for home-based networkcommunication was a phone modem. The word modem stands for modu-lator/demodulator. A phone modem converts computer data into ananalog audio signal for transfer over a telephone line, and then a modemat the destination converts it back again into data. One audio frequency isused to represent binary 0 and another to represent binary 1.

To use a phone modem, you must first establish a telephone connectionbetween your home computer and one that is permanently connected tothe Internet. That’s where your Internet service provider comes in. Youpay your ISP a monthly fee for the right to call one of several (hopefullylocal) computers that they have set up for this purpose. Once that connec-tion is made, you can transfer data via your phone lines to your ISP, whichthen sends it on its way through the Internet backbone. Incoming traffic isrouted through your ISP to your home computer.

This approach was fairly simple to implement because it does notrequire any special effort on the part of the telephone company. Since thedata is treated as if it were a voice conversation, no special translation isneeded except at either end. But that convenience comes at a price. Thedata transfer rate available with this approach is limited to that of analogvoice communication, usually 64 kilobits per second at most.

A phone line can provide a much higher transfer rate if the data istreated as digital rather than analog. A digital subscriber line (DSL) usesregular copper phone lines to transfer digital data to and from the phonecompany’s central office. Since DSL and voice communication use differentfrequencies, it is even possible to use the same phone line for both.

To set up a DSL connection, your phone company may become yourInternet service provider, or they may sell the use of their lines to a third-party ISP. To offer DSL service, the phone company must set up specialcomputers to handle the data traffic. Though not all phone companiessupport DSL yet, it is becoming an increasingly popular approach.

With DSL, there is no need to “dial in” to create the network connec-tion like there is with a phone modem. The DSL line maintains an activeconnection between your home and a computer at the ISP. However, tomake use of DSL technology, your home must be within a certain distancefrom the central office; otherwise, the digital signal degrades too muchwhile traveling between those two points.

A third option for home connections is a cable modem. In thisapproach, the data is transferred on the same line that your cable TVsignals come in on. Several leading cable TV companies in North Americahave pooled their resources to create Internet service providers for cablemodem service.

Broadband Networktechnologies that gener-ally provide data transferspeeds greater than 128bps

”Build a better mousetrap, and the worldwill beat a path to your door. Invent thecomputer mouse, and the world will all butforget your name.” This was the lead para-graph in an article celebrating the 20thbirthday of the computer mouse.1

Designed by Doug Engelbart—the namethat was forgotten—and a group of youngscientists and engineers at StanfordResearch Institute, the computer mousedebuted in 1968 at the Fall Joint Computerconference as part of a demonstration later called“The Mother of All Demos” by Andy van Dam. Thehistoric demonstration foreshadowed human-computer interaction and networking. It wasn’t until1981 that the first commercial computer with a mousewas introduced, however. In 1984 the Apple Macin-tosh brought the mouse into the mainstream. To thisday no one seems to know where the term “mouse”came from.

Engelbart grew up on a farm near Portland,Oregon, during the Depression. He served in theNavy in the Philippines during World War II as anelectronics technician. He completed his electricalengineering degree in 1948 from Oregon StateUniversity and moved to the Bay Area. In 1955 hereceived a Ph.D. from the University of California atBerkeley and joined the Stanford Research Institute.

Engelbart’s vision of the computer as an extensionof human communication capabilities and a resourcefor the augmentation of human intellect was outlinedin the seminal paper “Augmenting Human Intellect: AConceptual Framework,” published in 1962. He has

never lost this vision. Ever since, he hasbeen developing models to improve the co-evolution of computers with human organi-zations to boost collaboration, and tocreate what he calls “high performanceorganizations.” 2

During the 1970s and 1980s, Engelbartwas Senior Scientist at Tymshare, whichwas bought by McDonnell-Douglas. Whenthe program was shut down in 1989,Engelbart founded the Bootstrap Institute,

aimed at helping companies and organizations utilizehis techniques. He feels encouraged by the open-source movement, in which programmers collaborateto create advanced and complicated software. He iscurrently planning a system of open software that canbe distributed free over the Internet.

Recognition may have been long in coming, butEnglebart received 32 awards between 1987 and2001, including the Turing Award in 1997 and theNational Medal of Technology in 2000. The citationsfor these two prestigious awards read as follows:

(Turing Award) For an inspiring vision of the futureof interactive computing and the invention of keytechnologies to help realize this vision.

(National Medal of Technology) For creating thefoundations of personal computing includingcontinuous real-time interaction based on cathode-ray tube displays and the mouse, hypertext linking,text editing, online journals, shared-screen telecon-ferencing, and remote collaborative work.

Doug Engelbart

Both DSL connections and cable modems fall under the category ofbroadband connections, which generally mean speeds faster than 128 bitsper second. Debate between the DSL and cable modem communitiescontinues to rage to see who can claim the dominant market share. Bothgenerally provide data transfer speeds in the range of 1.5 to 3 megabits persecond.

461

Download Receivingdata on your homecomputer from theInternet

Upload Sending datafrom your homecomputer to a destina-tion on the Internet

Packet A unit of datasent across a network

Packet switching Theapproach to networkcommunication in whichpackets are individuallyrouted to their destina-tion, then reassembled

Router A networkdevice that directs apacket between networkstoward its final destina-tion

Repeater A networkdevice that strengthensand propagates a signalalong a long communi-cation line


For both DSL and cable modems, the speed for downloads (getting datafrom the Internet to your home computer) may not be the same as uploads

(sending data from your home computer to the Internet). Most traffic forhome Internet users are downloads: receiving Web pages to view andretrieving data (such as programs and audio and video clips) stored some-where else on the network. You perform an upload when you send an e-mailmessage, submit a Web-based form, or request a new Web page. Sincedownload traffic largely outweighs upload traffic, many DSL and cablemodem suppliers use technology that devotes more speed to downloads.

Packet SwitchingTo improve the efficiency of transferring information over a sharedcommunication line, messages are divided into fixed-sized, numberedpackets. The packets are sent over the network individually to their desti-nation, where they are collected and reassembled into the original message.This approach is referred to as packet switching.

The packets of a message may take different routes on their way to thefinal destination. Therefore, they may arrive in a different order than theway they were sent. The packets must be put into the proper order onceagain, and then combined to form the original message. This process isshown in Figure 15.4.

A packet may make several intermediate hops between computers onvarious networks before it reaches its final destination. Network devicescalled routers are used to direct packets between networks. Intermediaterouters don’t plan out the packet’s entire course; each router merely knowsthe best next step to get it closer to its destination. Eventually a messagereaches a router that knows where the destination machine is. If a path isblocked due to a down machine, or if a path currently has a lot of networktraffic, a router might send a packet along an alternative route.

If a communication line spans a long distance, such as across an ocean,a device called a repeater is installed periodically along the line to

Figure 15.4

Messages sent by packetswitching

Message is dividedinto packets

Packets are sent over the Internetby the most expedient route

Packets are reorderedand then reassembled

Packet 1

Packet 2

Packet 3

Packet 2

Packet 3

Packet 1

Sentmessage

Receivedmessage

Proprietary system Asystem that uses tech-nologies kept private bya particular commercialvendor

Interoperability Theability of software andhardware on multiplemachines and frommultiple commercialvendors to communicate

Open system A systemthat is based on acommon model ofnetwork architecture andan accompanying suiteof protocols

Open Systems Inter-

connection Reference

Model A seven-layerlogical breakdown ofnetwork interaction tofacilitate communicationstandards

15.2 Open Systems and Protocols 463

Figure 15.5

The layers of the OSI Reference Model

Application layer7

Presentation layer6

Session layer5

Transport layer4

Network layer3

Data Link layer2

Physical layer1

strengthen and propagate the signal. Recall from Chapter 3 that a digitalsignal loses information only if it is allowed to degrade too much. Arepeater keeps that from happening.

15.2 Open Systems and Protocols

Many protocols have been defined to assist in network communication.Some have gained a stronger foothold than others because of manyreasons, often historical. We focus in this section on the protocols used forgeneral Internet traffic. Before we discuss the details of particular proto-cols, however, it is important to put them in context by discussing theconcept of an open system.

Open SystemsEarly in the development of computer networks, commercial vendors cameout with a variety of technologies that they hoped businesses would adopt.The trouble was that these proprietary systems were developed with theirown particular nuances and did not permit communication betweennetworks of various types. As network technologies grew, the need forinteroperability became clear; we needed a way for computing systemsmade by different vendors to communicate.

An open system is one based on a common model of network architec-ture and a suite of protocols used in its implementation. Open-systemarchitectures maximize the opportunity for interoperability.

The International Organization for Standardization (ISO) establishedthe Open Systems Interconnection (OSI) Reference Model to facilitatethe development of network technologies. It defines a series of layers ofnetwork interaction. The seven layers of the OSI Reference Model areshown in Figure 15.5.

Protocol stack Layersof protocols that buildand rely on each other


Each layer deals with a particular aspect of network communication.The highest level deals with issues that relate most specifically to the appli-cation program in question. The lowest layer deals with the most basicelectrical and mechanical issues of the physical transmission medium (suchas types of wiring). The other layers fill in all other aspects. The networklayer, for example, deals with the routing and addressing of packets.

The details of these layers are beyond the scope of this book, but it isimportant to know that networking technology as we know it today ispossible only through the use of open-system technology and approachessuch as the OSI Reference Model.

Network ProtocolsFollowing the general concepts of the OSI Reference Model, networkprotocols are layered such that each one relies on the protocols thatunderlie it, as shown in Figure 15.6. This layering is sometimes referred toas a protocol stack. The layered approach allows new protocols to bedeveloped without abandoning fundamental aspects of lower levels. It alsoprovides more opportunity for their use in that the impact on other aspectsof network processing is minimized. Sometimes protocols at the same levelprovide the same service as another protocol at that level, but do so in adifferent way.

Keep in mind that a protocol is, in one sense, nothing more than anagreement that a particular type of data will be formatted in a particularmanner. The details of file formats and the size of data fields are importantto software developers creating networking programs, but we do notexplore those details here. The importance of these protocols is that theyprovide a standard way to interact among networked computers.

The lower two layers in Figure 15.6 form the foundation of Internetcommunication. Other protocols, sometimes referred to as high-levelprotocols, deal with specific types of network communication. These layersare essentially one particular implementation of the OSI Reference Modeland correspond in various ways to the levels described in that model. Let’sexplore these levels in more detail.

Figure 15.6

Layering of key networkprotocols

User Datagram Protocol (UDP)Transmission Control Protocol (TCP)

Internet Protocol (IP)

SMTP FTP Telnet

Transmission Control

Protocol (TCP) Thenetwork protocol thatbreaks messages intopackets, reassemblesthem at the destination,and takes care of errors

Internet Protocol (IP)

The network protocolthat deals with therouting of packetsthrough interconnectednetworks to the finaldestination

TCP/IP A suite ofprotocols and programsthat support low-levelnetwork communication

User Datagram

Protocol (UDP) Analternative to TCP thatachieves higher trans-mission speeds at thecost of reliability

Ping A program usedto test whether a partic-ular network computer isactive and reachable

Traceroute A programthat shows the route apacket takes across theInternet


TCP/IPTCP stands for Transmission Control Protocol and IP stands for Internet

Protocol. The name TCP/IP (pronounced by saying the letters T-C-P-I-P)refers to a suite of protocols and utility programs that support low-levelnetwork communication. The name TCP/IP is written to reflect the natureof their relationship—that TCP rests on top of the IP foundation.

IP software deals with the routing of packets through the maze of intercon-nected networks to their final destination. TCP software breaks messages intopackets, hands them off to the IP software for delivery, and then orders andreassembles the packets at their destination. TCP software also deals with anyerrors that occur, such as if a packet never arrives at the destination.

UDP stands for User Datagram Protocol. It is an alternative to TCP.That is, UDP software basically plays the same role as TCP software. Themain difference is that TCP is highly reliable, at the cost of decreasedperformance, while UDP is less reliable, but generally faster. Note thatUDP is part of the TCP/IP suite of protocols. Because of the heavy relianceon TCP, and for historical reasons, the entire suite is referred to as TCP/IP.

An IP program called ping can be used to test the reachability ofnetwork designations. Every computer running IP software “echoes” pingrequests, which makes ping a convenient way to test whether a particularcomputer is running and can be reached across the network. Ping officiallystands for Packet InterNet Groper, but the name was contrived to matchthe term used when submarines send out a sonar pulse and listen for thereturned echo. Since ping works at the IP level, it often responds evenwhen higher-level protocols might not. The term ping is often used as averb among network administrators: “Ping computer X to see if it isalive.”

Another TCP/IP utility program called traceroute shows the route thata packet takes to arrive at a particular destination node. The output oftraceroute is a list of the computers that serve as the intermediate stoppingpoints along the way.

High-Level ProtocolsOther protocols build on the foundation established by the TCP/IPprotocol suite. Some of the key high-level protocols are:

� Simple Mail Transfer Protocol (SMTP)—A protocol used to specifythe transfer of electronic mail

� File Transfer Protocol (FTP)—A protocol that allows a user on onecomputer to transfer files to and from another computer

� Telnet—A protocol used to log into a computer system from aremote computer. If you have an account on a particular computer

MIME type A standardfor defining the format offiles that are included ase-mail attachments or onWeb sites

Port A numeric desig-nation corresponding toa particular high-levelprotocol


Figure 15.7

Some protocols and theports they use

Protocol Port

EchoFile Transfer Protocol (FTP)TelnetSimple Mail Transfer Protocol (SMTP)Domain Name Service (DNS)GopherFingerHyper Text Transfer Protocol (HTTP)Post Office Protocol (POP3)Network News Transfer Protocol (NNTP)Internet Relay Chat (IRC)

721232553707980

110119

6667

that allows telnet connections, you can run a program that uses thetelnet protocol to connect and log in to that computer as if you wereseated in front of it.

� Hyper Text Transfer Protocol (HTTP)—A protocol defining theexchange of World Wide Web documents, which are typicallywritten using the Hyper Text Markup Language (HTML). HTML isdiscussed in more detail in the next chapter.

These protocols all build on TCP. Some high-level protocols have alsobeen defined that build on top of UDP in order to capitalize on the speed itprovides. But because UDP does not provide the reliability that TCP does,UDP protocols are less popular.

Several high-level protocols have been assigned a particular portnumber. A port is a numeric designation that corresponds to a particularhigh-level protocol. Servers and routers use the port number to helpcontrol and process network traffic. Common protocols and their ports arelisted in Figure 15.7. Some protocols, such as HTTP have default ports butcan use other ports as well.

MIME TypesRelated to the idea of network protocols and standardization is theconcept of a file’s MIME type. MIME stands for Multipurpose InternetMail Extension. Although MIME types do not define a network protocol,they define a standard for attaching or including multimedia or otherwisespecially formatted data with other documents, such as e-mail.

Based on a document’s MIME type, an application program can decidehow to deal with the data it is given. For example, the program you use to

Firewall A gatewaymachine and its softwarethat protects a networkby filtering the traffic itallows

Access control policy

A set of rules establishedby an organization thatspecify what types ofnetwork communicationare permitted anddenied


Figure 15.8 A firewall protecting a LAN

Protected LAN

Internetusers

Firewall

read e-mail may examine the MIME type of an e-mail attachment to deter-mine how to display it (if it can).

MIME types have been defined for the documents created by manycommon application programs, as well as for data from particular contentareas. Chemists and chemical engineers, for example, have defined a largeset of MIME types for various types of chemical-related data.

FirewallsA firewall is a machine and its software that serve as a special gateway to anetwork, protecting it from inappropriate access. A firewall filters thenetwork traffic that comes in, checking the validity of the messages asmuch as possible and perhaps denying some messages altogether. The maingoal of a firewall is to protect (and to some extent hide) a set of moreloosely administered machines that reside “behind” it. This process ispictured in Figure 15.8.

A firewall enforces an organization’s access control policy. Forexample, a particular organization may allow network communicationonly between its users and the “outside world” via e-mail, and deny othertypes of communication, such as accessing Web sites. Another organiza-tion may want to allow its users to freely access the resources of theInternet, but may not want general Internet users to be able to infiltrate itssystems or gain access to its data.

Hostname A namemade up of words sepa-rated by dots thatuniquely identifies acomputer on the Internet;each hostname corre-sponds to a particular IPaddress

IP address An addressmade up of four numericvalues separated by dotsthat uniquely identifies acomputer on the Internet


The system administrators of an organization set up a firewall for theirLAN that permits “acceptable” types of communication and denies othertypes. There are various ways in which this can be accomplished, thoughthe most straightforward is to deny traffic on particular ports. Forexample, a firewall could be set up to deny the ability for a user outsidethe LAN to create a telnet connection to any machine inside the LAN bydenying all traffic that comes in on port 23.

More sophisticated firewall systems may maintain internal informationabout the state of the traffic passing through them and/or the content of thedata itself. The more a firewall can determine about the traffic, the more ableit is to protect its users. Of course, this security comes at a price. Some sophis-ticated firewall approaches might create a noticeable delay in network traffic.

15.3 Network Addresses

When you communicate across a computer network, you ultimatelycommunicate with one particular computer out of all possible computersin the world. There is a fairly sophisticated mechanism for identifyingspecific machines to establish that communication.

A hostname is a unique identification that specifies a particularcomputer on the Internet. Hostnames are generally readable words sepa-rated by dots. For example:

X=\"——elo—olE"<<=>!E=leO“

o!>O!FlOeEe<!o!Fzlo!X

We humans prefer to use the hostnames when dealing with e-mailaddresses and Web sites because they are easy to use and remember.Behind the scenes, however, network software translates a hostname intoits corresponding IP address, which is easier for a computer to use. An IPaddress is usually represented as a series of four decimal numbers sepa-rated by dots. For example:

�9�lh”l�,�l��

�”hl�hhl�9l,

An IP address is stored in 32 bits. Each number in an IP address correspondsto one byte in the IP address. Since one byte (8 bits) can represent 256 things,each number in an IP address is in the range 0 to 255. See Figure 15.9.

It’s tempting to assume that since both hostnames and IP addresses areseparated into sections by dots, there is a correspondence between thesections. That is not true. First of all, an IP address always has four values,but hostnames can have a variety of sections.

Network address Thepart of an IP address thatspecifies a specificnetwork

Host number The partof an IP address thatspecifies a particularhost on the network

Domain name The partof a hostname that speci-fies a specific organiza-tion or group

15.3 Network Addresses 469

Figure 15.9

An IP address is stored infour bytes

�99�9�99

148

9�99��9

78

��9�9

250. . .

9999��99

12

An IP address can be split into a network address, which specifies aspecific network, and a host number, which specifies a particular machinein that network. How the IP address is split up depends on what network“class” it represents. The classes of networks (A, B, and C) provide fornetworks of various sizes.

Class A networks use the first byte for the network address and theremaining three bytes for the host number. Class B networks use the firstand second bytes for the network address and the last two bytes for thehost number. Class C networks use the first three bytes for the networknumber and the last byte for the host number.

Think about the range of values this addressing approach allows for thevarious network classes. There are relatively few class A networks, withpotentially many hosts on each. On the other hand, there are many class Cnetworks, but only a few (maximum 256) hosts on each. Class C networkaddresses are assigned to most organizations, whereas class A and B networksare reserved for very large organizations and Internet service providers.

The entire Internet protocol is based on a 32-bit IP address. If the use ofInternet-ready devices continues to grow, we will eventually run out ofreasonable address space to use. Debate continues to rage in networkingcircles about how to handle this dilemma.

Domain Name SystemA hostname consists of the computer name followed by the domain name.For example, in the hostname

X=\"——elo—olE"<<=>!E=leO“

X=\"——e is the name of a particular computer, and o—olE"<<=>!E=leO“ isthe domain name. A domain name is separated into two or more sectionsthat specify the organization, and possibly a subset of an organization, ofwhich the computer is a part. In this example, X=\"——e is a computer inthe Department of Computing Sciences at Villanova University.

The domain names narrow in on a particular set of networks controlledby a particular organization. Note that two organizations (or even subor-ganizations) can have a computer named the same thing because thedomain name makes it clear which one is being referred to.

Top-level domain (TLD)

The last section of adomain name, specifyingthe type of organizationor its country of origin


Figure 15.10

Top-level domains,including some relativelynew ones

Top-Level Domain General Purpose New TLDs General Purpose

.biz

.info

.pro

.museum

.aero

.coop

Business

Information

Professional

Museums

Aerospace industry

Cooperative

.com

.net

.org

.edu

.int

.mil

.gov

U.S. Commercial

Network

Nonprofit organization

U.S. Educational

International

U.S. Military

U.S. Government

Figure 15.11

Some of the top-leveldomain names based oncountry codes

Country Code TLD Country

.au

.br

.ca

.gr

.in

.ru

.uk

Australia

Brazil

Canada

Greece

India

Russian Federation

United Kingdom

The very last section of the domain is called its top-level domain (TLD)name. The primary top-level domains are listed in Figure 15.10.

The first column of Figure 15.10 shows the top-level domains that havebeen around since the Internet first evolved. Each one is used for organiza-tions of a particular type, such as lo!X for commercial businesses and leO“for colleges and universities. Organizations based in countries other thanthe United States use a top-level domain that corresponds to their two-letter country codes. Some of these codes (there are hundreds of them) arelisted in Figure 15.11.

Initially, anyone or any organization could register a domain name fortheir own use as long as that name hadn’t already been taken. As theInternet expanded, with new domain names being claimed regularly, itquickly became clear that there was a problem. A common lament amongnewcomers to the Internet was that the best domain names had already beentaken. Sometimes a name had already been claimed by another similarorganization, but in other cases people were trying to claim as many popularnames as possible, hoping to sell (ransom) them to large corporations (seethe discussion on domain name squatting at the end of this chapter).

Domain name system

A distributed system formanaging hostnameresolution

Domain name server

A computer that attemptsto translate a hostnameinto an IP address

Summary 471

To alleviate the problem of domain name use, a new set of top-leveldomains have been approved and are slowly being made available. The righthalf of Figure 15.10 shows the new TLDs. This time the ability to register adomain name using one of the new TLDs is being controlled, giving prefer-ence to organizations that hold trademarks on particular names.

The domain name system (DNS) is chiefly used to translate hostnamesinto numeric IP addresses. Before the DNS system was established, a Stan-ford research group maintained a single file known as the host table. Asnew host names were established, the Stanford groupwould add them to the table (usually twice a week).System administrators would retrieve the revised hosttable occasionally to update their domain name

servers, which are computers that translate (resolve) ahostname into its IP address.

As the number of hostnames grew, the single tableapproach became unreasonable. It simply wasn’t apractical way to update and distribute the information.In 1984, network engineers designed the more sophisti-cated domain name system that is in use today. DNS isan example of a distributed database (as discussed inChapter 12); no one organization is responsible forupdating the hostname/IP mappings.

When you specify a hostname in a browser windowor e-mail address, the browser or e-mail software sendsa request to a nearby domain name server. If that servercan resolve the hostname, it does so. If not, that serverasks another domain name server. If that second servercan’t resolve it, the request continues to propagate.Ultimately, the request finally reaches a server that canresolve the name, or the request expires because it tooktoo much time to resolve.

Summary

A network is a collection of computers connected to share resources anddata. Network technologies must concern themselves with underlyingprotocols and data transfer speeds. The client-server model has emerged asan important software technology given our ever-increasing reliance onnetworks.

Networks are often classified by their scope. A local-area network(LAN) covers a small geographic area and a relatively small number ofconnected devices. A wide-area network (WAN) embraces internet-working, connecting one network to another, and covers a large

Terrorist attacks spark run ondomain names

After the suicide jetliner attacks against theUnited States on September 11, 2001, hundredsof related Internet domain names were regis-tered—some for tributes and others for profit.Some legitimate sites included WTCStories.com,a collection of quotes and information on chari-ties, and AirTragedy.com, which contained newsand resources for victims and survivors.

On the other hand, some domain names wereclaimed in the hopes of selling them to interestedparties. WTCNot.com advertised that it was forsale for $500,000, and WTCdestruction.net wasavailable for $75,000. One major domainname reseller halted auctions for tasteless namessuch as NewYorkCarnage.com. The CEO of aname registration company said spectators ofattack-related names were misguided about theirvalue. His advice was to donate the $30 cost ofdomain-name registration to an appropriatecharity instead.3


geographic area. A municipal-area network (MAN) is specially designedfor large cities. LAN topologies include ring, star, and bus networks.Ethernet has become a standard for local-area networks.

Open systems are based on a common model of network architecture andprotocols, allowing for interoperability. The OSI Reference Model is a seven-layer description of network processing based on open-system principles.

The Internet backbone is a set of high-speed networks provided byvarious companies. Internet service providers (ISP) connect to the back-bone or to other ISPs and provide connections for both home and businesscomputing. Popular home connection technologies include phone modems,digital subscriber lines (DSL), and cable modems. Phone modems transferdata as audio signals and are therefore quite slow. DSL uses the samephone lines but transfers data digitally. Cable modems are also digital anduse the cable TV wiring to transfer data.

Messages are transferred over the Internet by breaking them up intopackets and sending those packets separately to their destination wherethey are reformed into the original message. Packets may make severalintermediate hops between networks before arriving at their destination.Routers are network devices that guide a packet between networks.Repeaters strengthen digital signals before they degrade too much.

Network protocols are layered so that a high-level protocol relies onlower-level protocols that support it. The key lower-level protocol suite forInternet traffic is TCP/IP. IP protocols and software deal with the routingof packets. TCP protocols and software divide messages into packets,reassemble them at the destination, and take care of errors that occur.High-level protocols include SMTP for e-mail traffic, FTP for file transfers,telnet for remote login sessions, and HTTP for Web traffic. Several high-level protocols have been assigned port numbers, which are used to helpcontrol and process network traffic. MIME types have been defined formany types of documents and special data formats.

A firewall protects a network from inappropriate access and enforces anorganization’s access control policy. Some firewalls simply block traffic onspecific ports, while more sophisticated firewalls analyze the content ofnetwork traffic.

An Internet network address must pinpoint a particular machine among allpossible ones in the world. A hostname uses readable words separated by dots.A hostname gets translated into an IP address, which is a numeric addressseparated into four sections. Part of the IP address identifies the network andpart identifies the specific host on that network. How the IP address is brokendown depends on the network class (A, B, or C) that the address references.

The domain name system (DNS) translates hostnames into IP addresses.DNS has evolved from using a single file containing all of the informationinto a distributed system dividing the responsibility among millions of

Ethical Issues 473

domain name servers. Top-level domains, such as lo!X and leO“, havebecome crowded, so some new top-level domains, such as l">D! andlY"i, have been approved.

CybersquattingCybersquatting refers to registering an Internet domain name (also

called “dot com” name) for the purpose of selling it later. How can

domain names become theft for resale? Why are they important

enough for someone to want to buy them? A company with a well-

known trademark tries to register the trademark as a domain name

only to find that someone else has already registered it. It may be that

the business that registered the name has a similar name or, more

likely, the name has been registered with the intention of selling it to

the company with the same trademark. Common names are also

subject to cybersquatting. For example, drugstore.com, furniture.com,

gardening.com, and Internet.com were sold by cybersquatters.

Names of famous people are targets for cybersquatters as well. For

example, in the 2000 National Football League draft, of the 120

players expected to be drafted only a few didn’t have domain sites

registered with their names—and very few of these sites were regis-

tered by the players themselves. One fan collects such sites as a piece

of history, but most of the people registering the names expected to

sell them after the draft.

In the late 1990s, domain-name auction houses arose. For a fee, the

house would appraise a registered domain name and offer it for

auction. Names such as 411.com, 611.com, and 911.com, all of which

were listed on one site, were expected to go for as much as $10 million

each. In addition, FastRefill.com was listed for $90,000, and Roast-

Beef.com was listed for $350,000.4

A different but related issue is the registering of domain names that

are clearly related to a famous person or brand name. For example, the

satirical site http://www.gwbush.com/ was set up during the 2000

presidential election campaign to poke fun at candidate George Bush.

(The official site was -\\zSTT lPe!FPeY“—-lo!XT.) Another site,

\!"—IFI“—lo!l“L, was set up to air a customer’s grievances with

the Toys ‘R’ Us company.

WWW


In November 1999, the Anti-cyber Piracy Act was passed by

Congress and signed by President Clinton. The Act establishes that

someone registering a domain name may be liable to the owner of a

trademark or to others that may be affected by the “bad faith” of the

domain name registrant. In August of 2000, Governor Davis of Cali-

fornia signed into law a bill that closed gaps in the federal legislation

by including protection for names that are not trademarked or suffi-

ciently famous to meet the federal standards.

In 1998 the Internet Corporation for Assigned Names and Numbers

(ICANN), a technical coordination body for the Internet, was created in

the private sector. ICANN issued the Uniform Domain-Name Dispute-

Resolution Policy (often referred to as the “UDRP”). As the ICANN Web

site states,

“Under the policy, most types of trademark-based domain-name

disputes must be resolved by agreement, court action, or arbitration

before a registrar will cancel, suspend, or transfer a domain name.

Disputes alleged to arise from abusive registrations of domain names (for

example, cybersquatting) may be addressed by expedited administrative

proceedings that the holder of trademark rights initiates by filing a

complaint with an approved dispute-resolution service provider.”5

These laws and policies have cut down on the cases of cybersquat-

ting, but some people are concerned that they go too far. These people

fear that in curbing domain-name abuses, individual rights to free

speech have been abridged.

Key TermsAccess control policy pg. 467

Broadband pg. 461

Bus topology pg. 458

Cable modem pg. 460

Client/Server model pg. 457

Computer network pg. 456

Data transfer rate (also band-width) pg. 456

Digital subscriber line (DSL) pg. 460

Domain name pg. 469

Domain name server pg. 471

Domain name system pg. 471

Download pg. 462

Ethernet pg. 458

File server pg. 457

Firewall pg. 467

Gateway pg. 458

Host number pg. 469

Hostname pg. 468

Internet pg. 458

Internet backbone pg. 459

Exercises 475

Internet Protocol (IP) pg. 468

Internet service provider (ISP) pg. 459

Interoperability pg. 463

IP address pg. 468

Local-area network (LAN) pg. 457

Metropolitan-area network(MAN) pg. 458

MIME type pg. 466

Network address pg. 469

Node (or Host) pg. 456

Open system pg. 463

Open Systems InterconnectionReference Model pg. 463

Packet pg. 462

Packet switching pg. 462

Phone modem pg. 460

Ping pg. 465

Port pg. 466

Proprietary system pg. 463

Protocol pg. 457

Protocol stack pg. 464

Repeater pg. 462

Ring topology pg. 457

Router pg. 462

Star topology pg. 458

TCP/IP pg. 465

Top-level domain (TLD) pg. 470

Traceroute pg. 465

Transmission Control Protocol(TCP) pg. 465

Upload pg. 462

User Datagram Protocol (UDP)pg. 465

Web server pg. 457

Wide-area network (WAN) pg. 458

Wireless pg. 456

Exercises1. What is a computer network?

2. How are computers connected together?

3. To what does the word node (host) refer?

4. Name and describe two key issues related to computer networks.

5. What is a synonym for data transfer rate?

6. Describe the client/server model and discuss how it has changed howwe think about computing.

7. Just how local is a local-area network?

8. Distinguish between the following LAN topologies: ring, star, and bus.

9. How does the shape of the topology influence message flow through aLAN?

10. What is Ethernet?

11. What is a WAN?

12. What is a gateway and what is its purpose?


13. What is the Internet?

14. What is a MAN and what makes it different from a LAN and a WAN?

15. Distinguish between the Internet backbone and an Internet serviceprovider (ISP).

16. Name at least two national ISPs.

17. Name and describe three technologies for connecting a home com-puter to the Internet.

18. What role do ISPs play with the three technologies in Exercise 17?

19. What are the advantages and disadvantages of each of the technolo-gies in Exercise 17?

20. Phone modems and digital subscriber lines (DSLs) use the same kindof phone line to transfer data. Why is DSL so much faster than phonemodems?

21. Why do DSL and cable modem suppliers use technology that devotesmore speed to downloads than to uploads?

22. Messages sent across the Internet are divided into packets. What is apacket and why are messages divided into them?

23. Explain the term packet switching.

24. What is a router?

25. What is a repeater?

26. What problems arise due to packet switching?

27. What are proprietary systems and why do they cause a problem?

28. What do we call the ability of software and hardware on multipleplatforms from multiple commercial vendors to communicate?

29. What is an open system and how does it foster interoperability?

30. Compare and contrast proprietary and open systems.

31. What is the seven-layer logical breakdown of network interactioncalled?

32. What is a protocol stack and why is it layered?

33. What constitutes the foundation of Internet communication?

34. What is the role of the IP protocol?

35. What is the role of the TCP protocol?

36. Define TCP/IP.

37. Compare TCP and UDP.

38. What is the functionality of the utility program ping?

39. What is the functionality of the utility program Traceroute?

Thought Questions 477

40. List four high-level protocols and what they specify.

41. What do we call a numeric designation corresponding to a particularhigh-level protocol?

42. Define MIME type.

43. What is a firewall, what does it accomplish, and how does it accom-plish it?

44. What is a host name and how is it composed?

45. What is an IP address and how is it composed?

46. What is the relationship between a hostname and an IP address?

47. Into what parts can an IP address be split?

48. What are the relative sizes of Class A networks, Class B networks, andClass C networks?

49. How many hosts are possible in Class C networks, in Class B net-works, and in Class A networks?

50. What is a domain name?

51. What is a top-level domain name?

52. How does the current domain name system try to resolve a hostname?

Thought Questions1. What is the computer system in your school like? Are all the com-

puters networked? Is there more than one network? Are the dormito-ries networked?

2. If you wanted to register a domain name, how would you go about it?lY"i, l">D!, lzF!, lX“—e“X, l=eF!, and lo!!z are new top-leveldomain names. Are there any current restrictions on the use of thesenew top-level domain names?

3. Do you think that the name Internet is appropriate? Would Intranetbe a better name?

4. Do you think the government or the private sector should monitordomain-name abuse?

5. Go to the ICANN Web site and read the UDRP. Do you think theirdefinition of “bad faith” is reasonable? Is it adequate to solve theabuses described here?

6. Should a person be allowed to create a Web site with the intention ofbroadcasting unsubstantiated claims against a product or company?

7. Should a person be allowed to create a Web site to parody a person,company, or political institution?

?