Telecommunications - Rutgers Universitycrab.rutgers.edu/~sundares/MIS334Sec40.Sp08/protected/ref_resour… · of telecommunications and is achieved through the use of telecommunication

T1 HardwareT2 SoftwareT3 Data and DatabasesT4 TelecommunicationsT5 The Internet and the WebT6 Technical View of System Analysis and Design

�

T4.1

T E C H N O L O G Y G U I D E

4Telecommunications

T4.1Telecommunications Concepts

T4.2Communications Media

(Channels)

T4.3Network Systems: Protocols,

Standards, Interfaces, andTopologies

T4.4Network Architecture

Technology Guides

t04.qxd 01/27/2005 04:43PM Page T4.1 EQA

T4.2 TECHNOLOGY GUIDES TELECOMMUNICATIONS

The term telecommunications generally refers to all types of long-distancecommunication that uses common carriers, including telephone, television, andradio. Data communications is the electronic collection, exchange, and pro-cessing of data or information, including text, pictures, and voice, that is digitallycoded and intelligible to a variety of electronic devices. Today’s computing envi-ronment is dispersed both geographically and organizationally, placing data com-munications in a strategic organizational role. Data communications is a subsetof telecommunications and is achieved through the use of telecommunicationtechnologies.

In modern organizations, communications technologies are integrated. Busi-nesses are finding electronic communications essential for minimizing time anddistance limitations. Telecommunications plays a special role when customers,suppliers, vendors, and regulators are part of a multinational organizationin a world that is continuously awake and doing business somewhere 24 hoursa day, 7 days a week (“24�7”). Figure T4.1 represents a model of an inte-grated computer and telecommunications system common in today’s businessenvironment.

A telecommunications system is a collection of compatible hardware andsoftware arranged to communicate information from one location to another.These systems can transmit text, data, graphics, voice, documents, or full-motionvideo information.

A typical telecommunications system is shown in Figure T4.2. Such systemshave two sides: the transmitter and the receiver.

The major components are:

1. Hardware—all types of computers (e.g., desktop, server, mainframe) andcommunications processors (such as a modems or small computers dedicatedsolely to communications).

2. Communications media—the physical media through which electronicsignals are transferred; includes both wireline and wireless media.

3. Communications networks—the linkages among computers and communi-cations devices.

4. Communications processors—devices that perform specialized data commu-nication functions; includes front-end processors, controllers, multiplexors,and modems.

5. Communications software—software that controls the telecommunicationssystem and the entire transmission process.

6. Data communications providers—regulated utilities or private firms thatprovide data communications services.

T4.1 TELECOMMUNICATIONS CONCEPTS

FIGURE T4.1 An inte-grated computer andtelecommunicationssystem.

TelecommunicationsSystem

Computer ComputerCommunications

ProcessorCommunications

channels andmedia

CommunicationsProcessor

NetworkSoftware

t04.qxd 01/27/2005 04:43PM Page T4.2 EQA

T4.1 TELECOMMUNICATIONS CONCEPTS T4.3

7. Communications protocols—the rules for transferring information across thesystem.

8. Communications applications—electronic data interchange (EDI), telecon-ferencing, videoconferencing, e-mail, facsimile, electronic funds transfer, andothers.

To transmit and receive information, a telecommunications system mustperform the following separate functions that are transparent to the user:

● Transmit information.

● Establish the interface between the sender and the receiver.

● Route messages along the most efficient paths.

● Process the information to ensure that the right message gets to the rightreceiver.

● Check the message for errors and rearrange the format if necessary.

● Convert messages from one speed to that of another communications line orfrom one format to another.

● Control the flow of information by routing messages, polling receivers, andmaintaining information about the network.

● Secure the information at all times.

Telecommunications media can carry two basic types of signals, analog and digi-tal (see Figure T4.3). Analog signals are continuous waves that “carry” infor-mation by altering the amplitude and frequency of the waves. For example, soundis analog and travels to our ears in the form of waves—the greater the height(amplitude) of the waves, the louder the sound; the more closely packed thewaves (higher frequency), the higher the pitch. Radio, telephones, and record-ing equipment historically transmitted and received analog signals, but they arerapidly changing to digital signals.

Digital signals are discrete on-off pulses that convey information in termsof 1’s and 0’s, just like the central processing unit in computers. Digital signalshave several advantages over analog signals. First, digital signals tend to be less

FIGURE T4.2 A telecom-munications system.

Electronic Signals

FIGURE T4.3 Analog vs.digital signals.

HostComputer

HostComputer

PC orTerminal

Front endprocessor

Receiver

Hardware

Multiplex Modem MultiplexModem Front EndProcessor

Telecommunication media(channels)

Analog data transmission(wave signals)

Digital data transmission(pulse signals)

0 1 0 0 0 01 1 1

t04.qxd 01/27/2005 04:43PM Page T4.3 EQA


affected by interference or “noise.” Noise (e.g., “static”) can seriously alter theinformation-carrying characteristics of analog signals, whereas it is generally eas-ier, in spite of noise, to distinguish between an “on” and an “off.” Consequently,digital signals can be repeatedly strengthened over long distances, minimizingthe effect of any noise. Second, because computer-based systems process digi-tally, digital communications among computers require no conversion fromdigital to analog to digital.

Communications processors are hardware devices that support data trans-mission and reception across a telecommunications system. These devicesinclude modems, multiplexers, front-end processors, and concentrators.

MODEM. A modem is a communications device that converts a computer’s dig-ital signals to analog signals before they are transmitted over standard telephonelines. The public telephone system (called POTS for “Plain Old Telephone Ser-vice”) was designed as an analog network to carry voice signals or sounds inan analog wave format. In order for this type of circuit to carry digital infor-mation, that information must be converted into an analog wave pattern. Theconversion from digital to analog is called modulation, and the reverse isdemodulation. The device that performs these two processes is called amodem, a contraction of the terms modulate/demodulate (see Figure T4.4).Modems are always used in pairs. The unit at the sending end converts digitalinformation from a computer into analog signals for transmission over analoglines; at the receiving end, another modem converts the analog signal back intodigital signals for the receiving computer. Like most communications equipment,a modem’s transmission speed is measured in bits per second (bps). Today,typical modem speeds range from 38,400 to 57,600 bps.

There are various types of modems, as described below:

● External modem—a stand-alone device, attaches to a special serial port ona computer, and a standard telephone cord connects to a telephone outlet.

● Internal modem—a card that you can insert into an expansion slot on acomputer’s motherboard. The internal moderm has the same functions asthose of an external modem.

● Digital modem—one that sends and receives data and information to andfrom a digital telephone line such as IDSN or DSL (see below).

● Cable modem—a modem that sends and receives data over the cable tele-vision (CATV) network.

CommunicationsProcessors

FIGURE T4.4 A modemconverts digital to analogsignals and vice versa.(Source: Computing in theInformation Age, Stern andStern, © 1993 John Wiley &Sons, Inc.)

TerminalorPC

Digitalsignal

Modemmodulates

signal

Analogtransmission

Modemdemodulates

signal

Digitalsignal

Centralcomputer

Company A

Direction of message

Company B

t04.qxd 01/27/2005 04:43PM Page T4.4 EQA

T4.1 TELECOMMUNICATIONS CONCEPTS T4.5

The amount of data actually transferred from one system to another in afixed length of time is only partially dependent on the transmission speed.Actual throughput speed, or the effective throughput speed (usually measuredin characters per second), varies with factors such as the use of data compressionor electrical noise interference.

NEWER ALTERNATIVES TO ANALOG MODEMS. Digital subscriber line (DSL)service allows the installed base of twisted-pair wiring in the telecommunica-tions system (see section T4.2) to be used for high-volume data transmission.DSL uses digital transmission techniques over copper wires to connect the sub-scribers to network equipment located at the telephone company central office.Asymmetric DSL (ADSL) is a variety of DSL that enables a person connecting fromhome to upload data at speeds from 16 to 640 Kbps and download data at 1.5to 8 Mbps. Clearly, this is many times faster than an analog modem. However,where it is available, ADSL service currently costs about $50 per month (whichusually includes Internet service). Voice-over-DSL (VoDSL) is a kind of telecom-munication service replacing the traditional telephone system. It provides voicephone functions using DSL services.

An ADSL circuit connects an ADSL modem on each end of a twisted-pairtelephone line, creating three information channels—a high-speed downstreamchannel; a medium-speed duplex channel, depending on the implementation ofthe ADSL architecture; and a POTS (Plain Old Telephone Service) or an ISDNchannel. The POTS/ISDN channel is split off from the digital modem by filters,thus guaranteeing uninterrupted POTS/ISDN, even if ADSL fails.

ADSL depends on advanced digital signal processing and creative algo-rithms to squeeze so much information through twisted-pair telephone lines. Tocreate multiple channels, ADSL modems divide the available bandwidth of atelephone line in one of two ways—frequency division multiplexing (FDM) orecho cancellation. FDM assigns one band for upstream data and another bandfor downstream data. Echo cancellation assigns the upstream band to overlapthe downstream and separates the two by means of local echo cancellation, atechnique well known in V.32 and V.34 modems. With either technique, ADSLsplits off a 4-kHz region for POTS at the DC end of the band.

An ADSL modem organizes the aggregate data stream created by multiplex-ing downstream channels, duplex channels, and maintenance channels togetherinto blocks, and it attaches an error-correction code to each block. The receiverthen corrects errors that occur during transmission up to the limits implied bythe code and the block length. The unit may, at the user’s option, also createsuperblocks by interleaving data within subblocks; this allows the receiver to cor-rect any combination of errors within a specific span of bits. This techniqueallows for effective transmission of both data and video signals alike.

As noted above, cable modems are offered by cable television companiesin many areas as a high-speed way to access a telecommunications network.These modems operate on one channel of the TV coaxial cable. Cost andtransmission speed are comparable to that of an ADSL. A cable modem givesusers high-speed Internet access through a cable TV network at more than 1mbps (1 million bits per second), or about 20 times faster than a traditional dial-up modem. When a cable modem unit is installed next to the computer, a split-ter is placed on the side of the household. It separates the coaxial cable line serv-ing the cable modem from the line that serves the TV sets. A separate coaxial

t04.qxd 2/16/05 4:18 PM Page T4.5 EQA


cable line is then run from the splitter to the cable modem. Cable modems typ-ically connect to computers through a standard 10Base-T Ethernet interface. Dataare transmitted between the cable modem and computer at 10 mbps. For details,see cable-modem.net/tt/primer.html.

MULTIPLEXER. A multiplexer is an electronic device that allows a singlecommunications channel (e.g., a telephone circuit) to carry data transmissionssimultaneously from many sources. The objective of a multiplexer is to reducecommunication costs by maximizing the use of a circuit by sharing it. A multi-plexer merges the transmissions of several terminals at one end of the channel,while a similar unit separates the individual transmissions at the receiving end.This process is accomplished through frequency division multiplexing (FDM), timedivision multiplexing (TDM), or statistical time division multiplexing (STDM). FDMassigns each transmission a different frequency. TDM and STDM merge togethermany short time segments of transmissions from different sending devices.

FRONT-END PROCESSOR. With most computers, the central processing unit(CPU) has to communicate with several devices or terminals at the same time.Routine communication tasks can absorb a large proportion of the CPU’s pro-cessing time, leading to degraded performance on more important jobs. In ordernot to waste valuable CPU time, many computer systems have a small second-ary computer dedicated solely to communication. Known as a front-endprocessor, this specialized computer manages all routing communications withperipheral devices.

The functions of a front-end processor include coding and decoding data,error detection, recovery, recording, interpreting, and processing the controlinformation that is transmitted. It can also poll remote terminals to deter-mine if they have messages to send or are ready to receive a message. Inaddition, a front-end processor has the responsibility of controlling access tothe network, assigning priorities to messages, logging all data communicationsactivity, computing statistics on network activity, and routing and reroutingmessages among alternative communication links and channels.

Cable Media

T4.2 COMMUNICATIONS MEDIA (CHANNELS)For data to be communicated from one location to another, a physical pathwayor medium must be used. These pathways are called communications media(channels) and can be either physical or wireless. The physical transmissionuse wire, cable, and other tangible materials; wireless transmission media sendcommunications signals through the air or space. The physical transmissionmedia are generally referred to as cable media (e.g., twisted pair wire, coax-ial cable, and fiber optic cable). Wireless media include cellular radio,microwave transmission, satellite transmission, radio and infrared media.

The advantages and disadvantages of various media are highlighted in TableT4.1. The essentials of these communications media are described below.

Cable media (also called wireline media) use physical wires or cables to transmitdata and information. Twisted-pair wire and coaxial cable are made of copper,and fiber-optic cable is made of glass. However, with the exception of fiber-optic

t04.qxd 2/16/05 12:15 PM Page T4.6 EQA

T4.2 COMMUNICATIONS MEDIA (CHANNELS) T4.7

cables, cables present several problems, notably the expense of installation andchange, as well as a fairly limited capacity.

Several cable media exist, and in many systems a mix of media (e.g., fiber-coax) can be found. The major cable media are as follows.

TWISTED-PAIR WIRE. Twisted-pair wire is the most prevalent form of com-munications wiring, because it is used for almost all business telephone wiring.Twisted-pair wire consists of strands of insulated copper wire twisted in pairs toreduce the effect of electrical noise.

Twisted-pair wire is relatively inexpensive, widely available, easy to workwith, and can be made relatively unobtrusive by running it inside walls, floors,and ceilings. However, twisted-pair wire has some important disadvantages. Itemits electromagnetic interference, is relatively slow for transmitting data, issubject to interference from other electrical sources, and can be easily “tapped”to gain unauthorized access to data.

Twisted-pair cabling comes in two varieties: shielded and unshielded.Unshielded twisted pair (UTP) is more popular and is generally the better optionfor small networks. The cable has four pairs of wires inside the jacket. Each pairis twisted with a different number of twists per inch to help eliminate interfer-ence from adjacent pairs and other electrical devices. The support for transmis-sion rate is higher depending on how tight the wires are twisted. The EIA/TIA(Electronic Industry Association/Telecommunication Industry Association) hasestablished standards of UTP and rated five categories of wire. A disadvantage

TABLE T4.1 Advantages and Disadvantages of Communications Channels

Channel Advantages Disadvantages

Twisted-pair

Coaxial cable

Fiber-optic cable

Microwave

Satellite

Radio

Cellular Radio

Infrared

InexpensiveWidely availableEasy to work withUnobtrusiveHigher bandwidth than twisted pairLess susceptible to electromagnetic

interferenceVery high bandwidthRelatively inexpensiveDifficult to tap (good security)High bandwidthRelatively inexpensiveHigh bandwidthLarge coverage area

High bandwidthNo wires neededSignals pass through wallsInexpensive and easy to installLow-to-medium bandwidthSignals pass through wallsLow-to-medium bandwidth

Slow (low bandwidth)Subject to interferenceEasily tapped (low security)

Relatively expensive and inflexibleEasily tapped (low-to-medium security)Somewhat difficult to work withDifficult to work with (difficult to splice)

Must have unobstructed line of sightSusceptible to environmental interferenceExpensiveMust have unobstructed line of sightSignals experience propagation delayMust use encryption for securityCreate electrical interference problemsSusceptible to snooping unless encrypted

Require construction of towersSusceptible to snooping unless encryptedMust have unobstructed line of sightUsed only for short distances

t04.qxd 2/16/05 12:15 PM Page T4.7 EQA


of untwisted pair cable is that it may be susceptible to radio and electrical fre-quency interference. Shielded twisted pair is suitable for environments with elec-trical interference. Shielded twisted pair is often used on networks using IBM’sToken Ring topology.

COAXIAL CABLE. Coaxial cable consists of insulated copper wire surroundedby a solid or braided metallic shield and wrapped in a plastic cover. It is muchless susceptible to electrical interference and can carry much more data thantwisted-pair wire. For these reasons, it is commonly used to carry high-speeddata traffic as well as television signals (i.e., in cable television). However, coax-ial cable is 10 to 20 times more expensive, more difficult to work with, and rela-tively inflexible. Because of its inflexibility, it can increase the cost of installationor recabling when equipment must be moved.

Data transmission over coaxial cable is divided into two basic types:

● Baseband. Transmission is analog, and each wire carries only one signal at atime.

● Broadband. Transmission is digital, and each wire can carry multiple signalssimultaneously.

Because broadband media can transmit multiple signals simultaneously, it isfaster and better for high-volume use. Therefore, it is the most popular Internet-access method.

Broadband needs a network interface card (NIC), also called a LANadapter, in order to run. An NIC is a card that is inserted into an expansion slotof computer or other device, enabling the device top connect to a network.Today, it can be in a form of USB type or PCMCIA type.

FIBER OPTICS. Fiber-optic technology, combined with the invention of thesemiconductor laser, provides the means to transmit information through clearglass fibers in the form of light waves, instead of electric current. Fiber-opticcables contain a core of dozen or thin strands of glass or plastic. Each strandis called an optical fiber and is thin as hair. These fibers can conduct light pulsesgenerated by lasers at transmission frequencies that approach the speed of light.Advantages are: able to carry significantly more signals than wire, faster datatransmission, less susceptible to noise from other devices, better security for sig-nals during transmission, smaller size. Disadvantages are: costs more than wire,can be difficult to install and modify.

Besides significant size and weight reductions over traditional cable media,fiber-optic cables provide increased speed, greater data-carrying capacity, andgreater security from interferences and tapping. A single hairlike glass fiber cancarry up to 50,000 simultaneous telephone calls, compared to about 5,500 callson a standard copper coaxial cable. The capacity of fiber is doubling every 6 to12 months. Optical fiber has reached data transmission rates of six trillion bits(terabits) per second in laboratories and, theoretically, fiber can carry up to 25terabits per second. Until recently, the costs of fiber and difficulties in installingfiber-optic cable slowed its growth.

The technology of generating and harnessing light and other forms of radi-ant energy whose quantum unit is the photon is called photonics. This scienceincludes light emission, transmission, deflection, amplification, and detection byoptical components and instruments, lasers and other light sources, fiber optics,

t04.qxd 2/16/05 12:15 PM Page T4.8 EQA


electro-optical instrumentation, related hardware and electronics, and sophisti-cated systems. The range of applications of photonics extends from energygeneration to detection to communications and information processing. Photonsare used to move data at gigabit-per-second speeds across more than 1,000wavelengths per fiber strand.

Since the most common method of increasing cable capacity is to send morewavelengths through each fiber, attenuation is a problem for fiber transmission.Attenuation is the reduction in the strength of a signal, whether analog or digi-tal. Attenuation requires manufacturers to install equipment to receive the dis-torted or “dirty” signals and send them out “clean.” These signal regeneratorscan cost tens of thousands of dollars to install on land; those under water cancost one million dollars each.

In addition, fiber absorbs some of the light passing through it, making itnecessary to amplify optical signals every 50 to 75 miles or so along the route.(Boosting the strength of the light so it can travel farther without amplificationincreases interference, causing distortion.) Eliminating optical amplifiers couldsave millions of dollars.

A recent advance has dramatically increased the capacity of fiber-opticcables. Scientists have replaced solid glass fibers with hollow glass tubes con-taining a vacuum. These tubes are lined with mirrors that reflect virtually 100percent of the light beaming through the tube. This advance multiplies fibercapacity and reduces the need for expensive amplification equipment.

A metropolitan area network (MAN) is a data network designed usu-ally for a town or a city. The fiber optic and associated equipment that makeup the MAN can be connected to the national communications backbone.Because the demand for high-speed data service is growing fast (even duringan economic downturn), it is estimated that roughly 43.5 million high-speedaccess devices will be in use by 2005. The existing legacy system is called Syn-chronous Optical Networks (SONETs), and these networks are best suited forvoice traffic. The MAN can be divided loosely into the metro core and the edge:The core connects to long-haul points-of-presence; the edge is the aggregatoror collector networks, which interface with large customers.

The approaches for building the MAN are either to improve exising SONETsby making “next-generation” SONET boxes or to develop multi-protocol DenseWavelength Division Multiplexing (DWDM) devices. The DWDM devices sit onthe fiber rings and allow each wavelength to act as a separate pipeline, whileadding intelligence to make transport more efficient. Both approaches are beingused.

Cable media (with the exception of fiber-optic cables) present several problems,notably the expense of installation and change, as well as a fairly limited capac-ity. The alternative is wireless communication. Common uses of wireless datatransmission include pagers, cellular telephones, microwave transmissions,communications satellites, mobile data networks, personal communicationsservices, and personal digital assistants (PDAs). Table T4.2 shows comparisonsamong various communications mediums.

MICROWAVE. Microwave systems are widely used for high-volume, long-distance, point-to-point communication. These systems were first used exten-sively to transmit very-high-frequency (up to 500 GHz) radio signals at the

Wireless Media

t04.qxd 2/16/05 12:15 PM Page T4.9 EQA


speed of light in a line-of-sight path between relay stations spaced approximately30 miles apart (due to the earth’s curvature). To minimize line-of-sight prob-lems, microwave antennas were usually placed on top of buildings, towers, andmountain peaks. Long-distance telephone carriers adopted microwave systemsbecause they generally provide about 10 times the data-carrying capacity of awire without the significant efforts necessary to string or bury wire. Comparedto 30 miles of wire, microwave communications can be set up much morequickly (within a day) and at much lower cost.

However, the fact that microwave requires line-of-sight transmissionseverely limits its usefulness as a practical large-scale solution to data commu-nication needs, especially over very long distances. Additionally, microwavetransmissions are susceptible to environmental interference during severeweather such as heavy rain or snowstorms. Although still fairly widely used,long distance microwave data communications systems have been largelyreplaced by satellite communications systems.

SATELLITE. A satellite is a space station that receives microwave signals froman earth-based station, amplifies the signals, and broadcasts the signals back overa wide area to any number of earth-based stations. Transmission to a satellite isan uplink, whereas transmission from a satellite to an earth-based station is adownlink.

A major advance in communications in recent years is the use of communi-cations satellites for digital transmissions. Although the radio frequencies used bysatellite data communication transponders are also line-of-sight, the enormous“footprint” of a satellite’s coverage area from high altitudes overcomes the lim-itations of microwave data relay stations. For example, a network of just threeevenly spaced communications satellites in stationary “geosynchronous” orbit22,241 miles above the equator is sufficient to provide global coverage.

The advantages of satellites include the following: The cost of transmission isthe same regardless of the distance between the sending and receiving stations

TABLE T4.2 Comparisons among Various Communications Media

Technology Capacity (Mbps) Advantage Limitations

Fiber to Several hundred, Highest speed costHome up to 1000

DSL Downstream: Uses existing Speed decreases6–8; upstream: phone lines with distance, noup to 1.5 service past 18,000 ft

Wireless Comparable to No cables Multipath interference,(terrestrial) DSL required weather and terrain

problems, limiteddistance

Wireless Varied No cable, no Limited data (satellite) antennas, best rates likely

suited tobroadcasts

Cable Downstream: typical Uses existing Data rate drops with�1; upstream: coaxial cable number of users,0.1–0.5 poor security, requires

major upgrade

t04.qxd 2/16/05 12:15 PM Page T4.10 EQA


within the footprint of a satellite, and cost remains the same regardless of thenumber of stations receiving that transmission (simultaneous reception). Satelliteshave the ability to carry very large amounts of data. They can easily cross or spanpolitical borders, often with minimal government regulation. Transmission errorsin a digital satellite signal occur almost completely at random; thus, statisticalmethods for error detection and correction can be applied efficiently and reliably.Finally, users can be highly mobile while sending and receiving signals.

The disadvantages of satellites include the following: Any one-way trans-mission over a satellite link has an inherent propagation delay (approximatelyone-quarter of a second), which makes the use of satellite links inefficient forsome data communications needs (voice communication and “stepping-on” eachother’s speech). Due to launch-weight limitations, satellites carry or generatevery little electrical power, and this low power, coupled with distance, can resultin extremely weak signals at the receiving earth station. Signals are inherentlynot secure because they are available to all receivers within the footprint—intended or not. Some frequencies used are susceptible to interference from badweather or ground-based microwave signals.

Types of Orbits. Currently, there are three types of orbits in which satellitesare placed: geostationary earth orbit, medium earth orbit, and low earth orbit.

Geostationary earth orbit (GEO) satellites orbit 22,241 miles directlyabove the equator and maintain a fixed position above the earth’s surface. Thesesatellites are excellent for sending television programs to cable operators andbroadcasting directly to homes. However, transmissions from GEO satellites takea quarter of a second to send and return (called propagation delay), making two-way telephone conversations difficult. Also, GEO satellites are large and expen-sive, and the equatorial orbit cannot hold many more GEO satellites than the160 that now orbit there. In 2000, a system of eight GEO satellites was launchedby Hughes Electronics at a cost of $3 billion.

Medium earth orbit (MEO), also called intermediate circuit orbit (ICO), satel-lites are located 6,250 to 13,000 miles above the earth’s surface, in orbits inclinedto the equator. While fewer satellites are needed to cover the earth than in LEOorbits, telephones need more power to reach MEO satellites than to reach LEOsatellites.

Low earth orbit (LEO) satellites are located 500 to 1,500 miles above theearth’s surface. Their closer proximity to the earth reduces or eliminates appar-ent signal delay. They can pick up signals from weak transmitters, meaning thatcellular (wireless) telephones need less power and can use smaller batteries. LEOsatellites consume less power and cost less to launch than GEO and MEO satel-lites. However, the footprints of LEO satellites are small, requiring many of themin order to cover the earth. Table T4.3 shows the differences among the threetypes of satellites.

Satellite Networks. Multiple LEO satellites from one organization arereferred to as constellations. Many companies are in the process of buildingconstellations of satellites for commercial service. SkyBridge (skybridgesatellite.com)uses two constellations of 40 LEO satellites each, orbiting at an altitude of 1,469kilometers, to cover the entire earth, except for the polar regions.

Satellite networking technology has taken a great stride forward. Now, thesatellite dish (which is a major part of the hardware) is just half the previous size,with uplink speeds increased 533 times (from 19.2 Kbps to 10 Mbps) as a resultof better compression, higher-powered satellites, and improvements in satellite

t04.qxd 2/16/05 12:15 PM Page T4.11 EQA


modems. Downstream rates have been improved from 1 Mbps to 60 Mbps. Oneunique feature of satellite networks is that they can broadcast massive chunks ofdata to multiple points. For more on satellites, see farsite.co.uk/Satellite_Routers/FarLinX_satellite_router.htm.

GLOBAL POSITIONING SYSTEMS. A global positioning system (GPS) is awireless system that uses satellites to enable users to determine their positionanywhere on the earth. GPS is supported by 24 U.S. government satellites thatare shared worldwide. Each satellite orbits the earth once in 12 hours, on a pre-cise path at an altitude of 10,900 miles. At any point in time, the exact posi-tion of each satellite is known, because the satellite broadcasts its position anda time signal from its on-board atomic clock, accurate to 1-billionth of a sec-ond. Receivers also have accurate clocks that are synchronized with those ofthe satellites. Knowing the speed of signals (186,272 miles per second), it is pos-sible to find the location of any receiving station (latitude and longitude) withinan accuracy of 50 feet by triangulation, using the distance of three satellitesfor the computation. GPS software computes the latitude and longitude andconverts it to an electronic map.

Other countries, troubled that the Global Positioning System is run by theU.S. military and controlled by the U.S. government, are building independentsatellite navigation networks. As a result, Europe is building a civil satellite sys-tem called Galileo, scheduled to be in operation by 2008. Mainland China andRussia are also constructing satellite systems for GPS uses.

GPS equipment has been used extensively for navigation by commercialairlines and ships and for locating trucks. GPS is now also being added to many

TABLE T4.3 Three Basic Types of Telecommunications Satellites

Type Considerations Orbit Number

GEO ● Stellites remain stationary relative to point 22,282 8on Earth miles

● Few satellites needed for global coverage● Transmission delay (approximately .25

second)● Most expensive to build and launch● Longest orbital life (12+ years)

MEO ● Satellites move relative to point on Earth 6,250 to 13,000 10–12● Moderate number needed for global coverage miles● Require medium-powered transmitters● Negligible transmission delay● Less expensive to build and launch● Moderate orbital life (6–12 years)

LEO ● Satellites move rapidly relative to point 500 to 1,500 manyon Earth miles

● Large number needed for global coverage● Require only low-power transmitters● Negligible transmission delay● Least expensive to build and launch● Shortest orbital life (as low as 5 years)

t04.qxd 2/16/05 12:15 PM Page T4.12 EQA


consumer-oriented electronic devices. The first dramatic use of GPS came dur-ing the Persian Gulf War, when troops relied on the technology to find theirway in the Iraqi desert. GPS also played the key role in targeting for smartbombs. Since then, commercial use has become widespread, including navi-gation, mapping, and surveying, particularly in remote areas. For example,several car manufacturers (e.g., Toyota, Cadillac) provide built-in GPS naviga-tion systems in their cars. GPSs are also available on cell phones, so you canknow where the caller is located. As of October 2001, cell phones in theUnited States must have a GPS embedded in them so that the location of acaller to 911 can be detected immediately. GPSs are now available to hikers inthe form of handheld devices costing less than $100. GPSs are also embeddedin some PDAs.

RADIO. Radio electromagnetic data communications do not have to dependon microwave or satellite links, especially for short ranges such as within anoffice setting. Broadcast radio is a wireless transmission medium that distributesradio signals through the air over both long distances and short distances. Radiois being used increasingly to connect computers and peripheral equipment orcomputers and local area networks. The greatest advantage of radio for datacommunications is that no wires need be installed. Radio waves tend to prop-agate easily through normal office walls. The devices are fairly inexpensive andeasy to install. Radio also allows for high data transmission speeds.

However, radio can create reciprocal electrical interference problems—withother office electrical equipment, and from that equipment to the radio com-munication devices. Also, radio transmissions are susceptible to snooping byanyone similarly equipped and on the same frequency. (This limitation can belargely overcome by encrypting the data being transmitted.)

INFRARED. Infrared (IR) light is light not visible to human eyes that can bemodulated or pulsed for conveying information. IR requires a line-of-sighttransmission. Many computers and devices have an IrDA (Infrared Data Associ-ation) port that enables the transfer of data using infrared light rays. IrDA is astandard defined by the IrDA Consortium. It specifies a way to transfer data wire-lessly via infrared radiation. The most common application of infrared light iswith television or videocassette recorder remote control units. With computers,infrared transmitters and receivers (or “transceivers”) are being used for short-distance connection between computers and peripheral equipment, or betweencomputers and local area networks. Many mobile phones have a built-in infrared(IrDA) port that supports data transfer.

Advantages of infrared light include no need to lay wire, equipment ishighly mobile, no electrical interference problems, no Federal CommunicationsCommission (FCC) permission required to operate an infrared transmitter, nocertification needed before selling an infrared device, and fairly inexpensivedevices with very high data rates. Disadvantages of infrared media include sus-ceptibility to fog, smog, smoke, dust, rain, and air temperature fluctuations. Fordetails, see hw.cz/english/docs/irda/irda.html.

CELLULAR RADIO TECHNOLOGY. Mobile telephones, which are being usedincreasingly for data communications, are based on cellular radio technology,which is a form of broadcast radio that is widely used for mobile communications.

t04.qxd 01/27/2005 04:43PM Page T4.13 EQA


The basic concept behind this technology is relatively simple: The Federal Com-munication Commission (FCC) has defined geographic cellular service areas;each area is subdivided into hexagonal cells that fit together like a honeycombto form the backbone of that area’s cellular radio system. Located at the centerof each cell is a radio transceiver and a computerized cell-site controller thathandles all cell-site control functions. All the cell sites are connected to a mobiletelephone switching office that provides the connections from the cellular sys-tem to a wired telephone network and transfers calls from one cell to anotheras a user travels out of the cell serving one area and into another.

The cellular telephone infrastructure has primarily been used for voicetransmission, but recent development of a transmission standard called cellulardigital packet data (CDPD) has made it possible for the infrastructure to supporttwo-way digital transmission. The evolution of cellular transmission from analogto digital is described below.

First-generation (1G) and second-generation (2G) cellular data transmission.1G technology was characterized by bulky handsets and adjustable antenna, andwas based on analog technology. 1G allowed only limited roaming.

Second-generation (2G) cellular data transmission. 2G technology providesdigital wireless transmission. 2G increases the voice capacity of earlier analogsystems, and provides greater security, voice clarity, and global roaming.

2.5-generation (2.5G) cellular data transmission. 2.5G technology, usuallyassociated with General Pocket Radio Service (GPRS), extends the 2G digitalcellular standard and is installed as an upgrade to an existing 2G network.

Third-generation (3G) technologies. 3G technology offers increased effi-ciency and capacity; new services, such as wide-area networks for PCs andmultimedia; seamless roaming across dissimilar networks; integration ofsatellite and fixed wireless access services into cellular networks; and greaterbandwidth.

MOBILE COMPUTING. Mobile computing refers to the use of portable com-puter devices in multiple locations. It occurs on radio-based networks that trans-mit data to and from mobile computers. Computers can be connected to thenetwork through wired ports or through wireless connections. Mobile computingprovides for many applications, including m-commerce (see Chapter 6).

Another type of mobile data network is based on a series of radio towers con-structed specifically to transmit text and data. BellSouth Mobile Data and Ardis(formerly owned by IBM and Motorola) are two privately owned networks thatuse these media for national two-way data transmission.

PERSONAL COMMUNICATION SERVICE. Personal communication service(PCS) uses lower-power, higher-frequency radio waves than does cellular tech-nology. It is a set of technologies used for completely digital cellular devices,including handheld computers, cellular telephones, pagers, and fax machines.The cellular devices have wireless modems, allowing you Internet access ande-mail capabilities. The lower power means that PCS cells are smaller and mustbe more numerous and closer together. The higher frequency means that PCSdevices are effective in many places where cellular telephones are not, such as intunnels and inside office buildings. PCS telephones need less power, are smaller,and are less expensive than cellular telephones. They also operate at higher,

t04.qxd 01/27/2005 04:43PM Page T4.14 EQA


less-crowded frequencies than cellular telephones, meaning that they will havethe bandwidth necessary to provide video and multimedia communications.

PERSONAL DIGITAL ASSISTANTS. Personal digital assistants (PDAs) aresmall, handheld computers capable of entirely digital communications trans-mission (see discussion in Technology Guide 1). They have built-in wirelesstelecommunications capabilities. Applications include Internet access, e-mail,fax, electronic scheduler, calendar, and notepad software.

UPS’s Delivery Information Acquisition Device (DIAD) is a hand-held elec-tronic data collector that UPS drivers use to record and store information, thushelping UPS to keep track of packages and gather delivery information withinUPS’s nationwide, mobile cellular network. It digitally captures customers’ pack-age information, thus enabling UPS to keep accurate delivery records. Driversinsert the DIAD into a DIAD vehicle adapter (DVA) in their delivery vehicles totransmit over UPS’s nationwide cellular network for immediate customer use.It contains 1.5MB RAM, can consolidate multiple functions into single keys,accepts digital signatures, and has a built-in acoustical modem. Its laser scannerreads package labels quickly and accurately, “smart” software knows the driver’snext street, and the device interacts with UPS cellular service.

WIRELESS APPLICATION PROTOCOL. Wireless Application Protocol (WAP)is a technology that enable wireless trasmissions. For example, one popularapplication that utilizes WAP is i-mode, a wireless portal that enables users toconnect to the Internet. Developed by NTT DoCoMo, i-mode provides analways-on connection to the Internet and content sites from popular media out-lets, all accessible via color-screen handsets with polyphonic sound. It is chargedat actual usage instead of on a pre-paid basis. WAP is criticized for browsingwith small screens, little compelling content, and bad connections at great costthrough a browser. Despite these drawbacks, it offers users the ability to makewireless connections to the Internet, which has enormous commercial appeal.

NEWER WIRELESS TECHNOLOGIES. Because of the requirements of fasterspeed and strict security requirements that existing WAP cannot fulfill, newerwireless technologies are being created for future purposes. Listed below aresome of the major new wireless technologies.

Bluetooth. A relatively new technology for wireless connectivity is calledBluetooth. It is the term used to describe the protocol of a short-range (10-meter), frequency-hopping radio link between devices. Bluetooth allows wire-less communication between mobile phones, laptops, and other portabledevices. Bluetooth technology is currently being built into mobile PCs, mobiletelephones, and PDAs.

Bluetooth is the code name for a technology designed to provide an openspecification for wireless communication of data and voice. It is based on alow-cost, short-range radio link built into a 9 � 9 mm microchip, providingprotected ad hoc connections for stationary and mobile communication envi-ronments. It allows for the replacement of the many existing proprietary cablesthat connect one device to another with one universal short-range radio link.Designed to operate in a noisy radio-frequency environment, the Bluetoothradio uses a fast acknowledgement and frequency-hopping system scheme to

t04.qxd 2/16/05 12:15 PM Page T4.15 EQA


make the link robust. Its modules avoid interference from other signals by hop-ping to a new frequency after transmitting or receiving a packet. It operates inISM band at 2.4 GHz. Bluetooth is also used in WPAN (described below) in anew standard 802.15.

Fiber optics without the fiber. Another new technology is “fiber optics with-out the fiber.” With this technology, laser beams are transmitted through theair between two buildings or other points. Terabeam Corporation recently intro-duced such a service. Like other wireless media, the chief advantage of thistechnology is that there is no need to gain rights of way to lay cabling. However,weather can have a negative impact on transmission quality.

Ultrawideband. Ultrawideband (UWB) is a superfast, short distance wire-less technology that will have data speeds 10 times faster than Wi-Fi, which isactually a nickname for the 802.11b protocol. (See Chapter 6 for more on Wi-Fi.) It works by transmitting its signal over a wide swath of frequencies at a lowpower that does not interfere with the other occupants of the spectrum. It cangive rise to a new generation of portable and home entertainment products withquality equal to a hardwired system that is perfect for home networking.

Software-Defined Radio. Software-defined radio is a concept of a recon-figurable device that can automatically recognize and communicate with otherdevices. This concept could impose wireless standards that compete with theexisting ones like CDMA, GSM, TDMA to transform today’s rigid networks intoan open system. The benefits are: improved system performance, cheaper servicecost, seamless roaming (i.e., you could carry a single device for multiplepurposes). Intel is now developing a new CPU that will include a type ofsoftware-defined radio that can adapt to different wireless LAN standards.

Mesh Networks. A mesh network is created by a device that can turnnearly any wireless device into a router, creating an ad hoc network. Membersof a network no longer rely on a central routing hub to distribute data; rather,the information hops from one user’s device to another until it gets where it’sgoing. Benefits are cheaper service, wider coverage areas, and speed (the meshnetwork can send data at speeds above 6 Mbps). Drawbacks are a security prob-lem because of numerous pass-throughs and billing problems due to changes inconnectivity.

Wireless Personal Area Networks. Defined by IEEE as wireless networksthat cover an area of at least 10 meters around a person, wireless personalarea networks (WPANs) could eliminate cable and wire networks. Comput-ing devices within a WPAN create a flow of machine-to-machine communica-tion that personalizes services spontaneously. Possible problems are managingdevice interoperability, maintaining always-on connectivity between devices,and leakage of privacy information. (See Chapter 6 for more.)

Adaptive Radio. Adaptive radio is a technology that lets wireless devicesscout out the spectrum wherever they are, avoiding interference by tuning theirtransmissions to the available gaps. The primary benefit of this technology is itenables wireless devices to modify their power, frequencies, or timing to suitthe environment they find themselves in, making such adjustments at occa-sional intervals or constantly checking and changing as airwave traffic shiftsaround them.

HomePlug. A product called HomePlug makes it easy to use existing in-wall electrical wiring for fast home networks. It lets you network devices byplugging an external adapter into a standard wall outlet, and delivers performance

t04.qxd 2/16/05 12:15 PM Page T4.16 EQA


superior to that of 802.11b wireless networks at only a small price premium. Itssecurity protection is more robust than 802.11b because it uses DES encryptionwhile 802.11b uses RC4 algorithm. It has a maximum speed of 14 mbps whichis slightly faster than 802.11b’s 11 mbps. Moreover, it is not subjected to otherwireless traffic or to interference from walls and doors like that of 802.11b.Table T4.4 shows 802.11 wireless networking standard.

Communications media have several characteristics that determine their effi-ciency and capabilities. These characteristics include the speed, direction, mode,and accuracy of transmission.

TRANSMISSION SPEED. Bandwidth refers to the range of frequencies thatcan be sent over a communications channel. Frequencies are measured in thenumber of cycles per second (or Hertz, abbreviated Hz). Bandwidth is animportant concept in communications because the transmission capacity of achannel is largely dependent on its bandwidth. Capacity is stated in bits persecond (bps), thousands of bps (Kbps), millions of bps (Mbps), and billions ofbps (Gbps). In general, the greater the bandwidth of a channel, the greater thechannel capacity.

A baud is a detectable change in a signal (i.e., a change from a positive toa negative voltage in a wire). The amount of data that can be transmittedthrough a channel is known as its baud rate, measured in bits per second (bps).A baud represents a signal change from positive to negative, or vice versa. Thebaud rate is not always the same as the bit rate. At higher transmission speeds,a single signal change can transmit more than one bit at a time, so the bit ratecan be greater than the baud rate.

For many data communications applications (i.e, those that involve textualdata), a low bandwidth (2400 to 14,400 bps) is adequate. On the other hand,acceptable performance for transmission of graphical information requiresbandwidth in the Mbps range.

Channel capacity is usually divided into three bandwidths: narrowband, voice-band, and broadband channels. Slow, low-capacity transmissions, such as thosetransmitted over telegraph lines, make use of narrowband channels, while tele-phone lines utilize voiceband channels. The channel bandwidth with the highestcapacity is broadband, used by microwave, cable, and fiber-optic media.

Communications channels have a wide range of speeds based on the tech-nology used. Transfer rates measure the speed with which a line carries data. Thefaster the transfer rate, the faster you can send and receive data and information.

TABLE T4.4 802.11 Wireless Networking Standards

802.11 standard Functions

802.11a 54-mbps top speed; incompatible with 802.11b802.11b 11-mbps top speed; popular in home and small-business

networks802.11e Enhances audio and video transmission on 802.11a, b, or g802.11g New standard with 54-mbps top speed; compatible

with 802.11b802.11i Adds enhanced 128-bit encryption to 802.11a, b, or g

Characteristics ofCommunications

Media

t04.qxd 2/16/05 12:15 PM Page T4.17 EQA


The unit is bits per second (bps). Transfer rates for various types of transmis-sion media are shown in Table T4.5.

The amount of data actually transferred from one system to another in a fixedlength of time is only partially dependent on the transmission speed. Actual oreffective throughput speed (usually measured in characters per second) varies withfactors such as the use of data compression or electrical noise interference.

OPTICAL NETWORKING. Wave division multiplexing (WDM) is a techniquewhereby different colors of light are transmitted on an optical fiber so that morethan one message can be transmitted at a time. (This is actually a form of FDM.)Recent innovations have led to dense wave division multiplexing (DWDM), wherehundreds of messages can be sent simultaneously. DWDM dramatically increasesthe capacity of existing optical fiber networks without laying any new cable.Eventually, an all-optical network, where signals coming into the network areimmediately converted to colors of light and managed at an optical layer, willbecome economically feasible. The increased capacity will be very significant.

Optical networks use beams of light, converted into light pulses, to carrydata. Lasers send billions of such pulses per second, ion patterns representingdigital ones and zeros, through tiny strands of glass fiber that can extend forhundreds or thousands of miles. There are two types of optical switches toincrease the network’s flexibility and efficiency: The first one is called optical-electrical-optical (OEO) switches. They take light pulses from the incoming opticalfiber, convert them to electrical signals, redirect or process them electronically,and use lasers to convert them back into light for transmission over the outgo-ing fiber. Currently, most of the networks use optoelectrical switches to directnetwork traffic, but the conversion process that is involved is very slow.

The second type of optical switch is called an all-optical switch. It accepts astream of light pulses from an incoming optical fiber and merely redirects it toan output port via miniature movable mirrors, without ever converting the opti-cal signals into electronic ones. Since they use mirrors instead of electronics,they eliminate the conversion step and can redirect light streams no matter howmany bits they carry. Indeed, since they cannot break down data streams elec-tronically, they can switch nothing smaller than full-scale optical data channels.The benefits are faster performance and higher network capacity. Thus, the use

TABLE T4.5 Transmission Rates in Different Media

Medium Capacity

Twisted pair Up to 128 MbpsCoaxial cable Up to 200 MbpsFiber-optic cable 100 Mbps to 2 GbpsBroadcast radio Up to 2 MbpsMicrowave 45 MbpsSatellite 50 MbpsCellular radio, 2 G cell phone 9600 bps to 14.4 KbpsCell phone, 3 G Up to 2 MbpsCell phone, 2.5 G GPRS, up to 115 Kbps

EDGE, up to 384 KbpsInfrared 1 to 4 Mbps

t04.qxd 01/27/2005 04:43PM Page T4.18 EQA


of such switches will encourage the construction of networks containing morehigh-capacity circuits than would the use of OEO switches. The drawback is thatthey rely on a series of microscopic mechanical mirrors to do their work, andthus are not as reliable as their old optoelectronic counterparts.

Data transmissions can be described in terms of their direction and their tim-ing. Direction of data transmission can be a simplex, half-duplex, or full-duplex.Timing of data transmissions can be either asynchronous or synchronous.

SIMPLEX TRANSMISSION. Simplex data transmission uses one circuit in onedirection only—similar to a doorbell, a public announcement system, or broadcasttelevision and radio. Simplex transmission is simple and relatively inexpensivebut very constraining, since communication is one way only.

HALF-DUPLEX TRANSMISSION. Like simplex transmission, half-duplex datatransmission uses only one circuit, but it is used in both directions—one direc-tion at a time. Examples include an intercom or a citizen’s band radio whereusers can receive or transmit, but cannot do both simultaneously. Two-way datatransfer makes half duplex much more useful than simplex, but coordinationof half-duplex communications could be difficult.

FULL-DUPLEX TRANSMISSION. Full-duplex data transmission uses two cir-cuits for communications—one for each direction simultaneously (for example,a common telephone). Full-duplex is easier to use than half-duplex, but thecost of two circuits can be significant, especially over long distances. Most datadevices can operate in both half- and full-duplex directions.

ASYNCHRONOUS TRANSMISSION. In asynchronous transmission, only onecharacter is transmitted or received at a time. During transmission, the charac-ter is preceded by a start bit and followed by a stop bit that lets the receivingdevice know where a character begins and ends. There is typically idle timebetween transmission of characters, so synchronization is maintained on a char-acter-by-character basis. Asynchronous transmission is inherently inefficient dueto the additional overhead required for start and stop bits, and the idle timebetween transmissions. It is generally used only for relatively low speed datatransmission (up to 56 Kbps).

SYNCHRONOUS TRANSMISSION. With synchronous transmission, a groupof characters is sent over a communications link in a continuous bit streamwhile data transfer is controlled by a timing signal initiated by the sendingdevice. The sender and receiver must be in perfect synchronization to avoid theloss or gain of bits; therefore, data blocks are preceded by unique charac-ters called sync bits that are encoded into the information being transmitted.The receiving device recognizes and synchronizes itself with a stream of thesecharacters. Synchronous transmission is generally used for transmitting largevolumes of data at high speeds.

An electrical communications line can be subject to interference from storms,signals from other lines, and other phenomena that introduce errors into atransmission. Telephone line cables may be mishandled by repair personnel,

TransmissionDirection and Mode

TransmissionAccuracy

t04.qxd 2/16/05 12:15 PM Page T4.19 EQA


accidentally cut by construction workers, or subjected to power surges whiledata are being transmitted. These events might cause one bit or several bits tobe “dropped” during transmission, thus corrupting the integrity of the informa-tion. Because the loss of even one bit could alter a character or control code,data transmission requires accuracy controls. These controls consist of bits, calledparity bits, that are like checksums added to characters and/or blocks of char-acters at the sending end of the line. Parity bits are checked and verified at thereceiving end of the line to determine whether bits were lost during transmission.

If transmission errors are detected, there are two general types of actionstaken—backward error correction and forward error correction. Backward errorcorrection (BEC) entails going back to the sender and requesting retransmissionof the entire data stream or of a particular part, if it can be identified. Forwarderror correction (FEC) uses knowledge about the message stream and mathe-matical algorithms to allow the receiver to correct the received data stream with-out having to go back to the sender. BEC is much simpler and less expensive touse when there are few errors or when time delays are not crucial. FEC is morecomplex but may be necessary over long distances when retransmission is costly.

Telecommunications carriers are companies that provide the communica-tions technology (e.g., telephone lines, satellites, and communications software)and services needed for data communications. These carriers include commoncarriers, other special-purpose carriers, and value-added carriers. The commoncarriers are the long-distance telephone companies. For example, AT&T, MCI,and Sprint are common carriers for long-distance service. Special-purposecarriers typically provide other services such as WATS (wide area telephoneservice) lines. Value-added carriers are companies that have developed pri-vate telecommunications systems and provide services for a fee (such asmicrowave or satellite transmission). An example is Tymnet by MCI WorldCom.Global Crossing is a new kind of telecommunications carrier. It is based on lasertechnology and its U.S. network connects 25 cities with just six soft switches,each able to connect 1600 calls per second.

Telecommunications carriers use the following technologies:

SWITCHED AND DEDICATED LINES. Switched lines are telephone lines, pro-vided by common carriers, that a person can access from his or her computerto transmit data to another computer; the transmission is routed or switchedthrough paths to its destination. A switch is a special-purpose circuit that routesmessages along specific paths in a telecommunications system.

Dedicated lines, also called leased lines, provide a constant connectionbetween two devices and require no switching or dialing. These lines are con-tinuously available for transmission, and the lessee typically pays a flat rate fortotal access to the line. The lines can be leased or purchased from common car-riers or value-added carriers. These lines typically operate at higher speed thanswitched lines and are used for high-volume transactions.

A dedicated line may handle digital data only, or it may be capable of han-dling both voice and digital data, just as a standard telephone line does. Whendedicated lines have been designed specifically for data transmission, they pro-duce less static and fewer transmission errors than regular telephone lines, andthey are more secure from wiretapping and other security risks. Most impor-tantly, the central processor is always accessible through the dedicated line.

TelecommunicationsCarriers

t04.qxd 2/16/05 12:15 PM Page T4.20 EQA

T4.3 NETWORK SYSTEMS: PROTOCOLS, STANDARDS, INTERFACES, AND TOPOLOGIES T4.21

Network architectures facilitate the operation, maintenance, and growth of thenetwork by isolating the user and the application from the physical details ofthe network. Network architectures include protocols, standards, interfaces, andtopologies.

Devices that are nodes in a network must access and share the network totransmit and receive data. These components work together by adhering to acommon set of rules that enable them to communicate with each other. This setof rules and procedures governing transmission across a network is a protocol.

The principal functions of protocols in a network are line access and collisionavoidance. Line access concerns how the sending device gains access to the net-work to send a message. Collision avoidance refers to managing message trans-mission so that two messages do not collide with each other on the network.Other functions of protocols are to identify each device in the communicationpath, to secure the attention of the other device, to verify correct receipt of thetransmitted message, to verify that a message requires retransmission because itcannot be correctly interpreted, and to perform recovery when errors occur.

The simplest protocol is polling, where a master device (computer or com-munications processor) polls, or contacts, each node. Polling can be effectivebecause the speed of mainframe and communications processors allows themto poll and control transmissions by many nodes sharing the same channel,particularly if the typical communications are short.

In the token-passing approach, a small data packet, called a token, is sentaround the network. If a device wants to transmit a message, it must wait forthe token to pass, examine it to see if it is in use and pass it on, or use thetoken to help route its message to its destination on the network. After trans-mission is completed, the token is returned to the network by the receivingterminal if it is not needed. IBM token ring networks use this access method.

In another approach, called contention, which is part of the Ethernet pro-tocol, a device that wants to send a message checks the communicationsmedium (e.g., a twisted pair wire) to see if it is in use. If one device (e.g., a PC)detects that another device (e.g., a printer) is using the channel (i.e., a collisionoccurs), it waits a random time interval and retries its transmission.

The most common protocol is Ethernet 10BaseT. Over three-fourths of allnetworks use the Ethernet protocol. The 10BaseT means that the network has aspeed of 10 Mbps. Fast Ethernet is 100BaseT, meaning that the network has aspeed of 100 Mbps. The most common protocol in large corporations is theGigabit Ethernet. That is, the network provides data transmission speeds ofone billion bits per second (666 times faster than a T1 line). However,100-gigabit Ethernet is becoming the standard (100 billion bits per second). (See307.ibm.com/pc/support/site.wss/document.do?Indocid=MIGR-43949.)

The Transmission Control Protocol/Internet Protocol (TCP/IP) is aprotocol for sending information across sometimes-unreliable networks with theassurance that it will arrive in uncorrupted form. TCP/IP allows efficient andreasonably error-free transmission between different systems and is the stan-dard protocol of the Internet and intranets. (Further discussion of this protocolmay be found in Technology Guide 5.)

Communication andNetwork Protocols

T4.3 NETWORK SYSTEMS: PROTOCOLS, STANDARDS, INTERFACES, AND TOPOLOGIES

t04.qxd 01/27/2005 04:43PM Page T4.21 EQA


Convergence is used to refer to the ability to transfer all types ofinformation—voice, data, video—utilizing a single Internet protocol (IP) networkinfrastructure. In voice-over IP (VoIP) systems, analog voice signals are digi-tized and transmitted as a stream of packets over a digital IP data network. VoIPutilizes a gateway to compress and convert the caller’s voice into digital IP pack-ets. These packets are then sent along the IP network. A second gateway thenputs the voice packets in the correct order, decompresses them, and convertsthe voice packets back into a sound signal that can be received by existingtelephone equipment.

Internet telephony (IPT) is the transport of telephone calls over theInternet, no matter whether traditional telephony devices, multimedia PCs, ordedicated terminals take part in the calls and no matter whether the calls areentirely or only partially transmitted over the Internet.

In 1999, the Internet Engineering Task Force (IETF) published RFC 2543which defined the session initiation protocol (SIP). SIP is the IETF’s take onthe end-to-end model of IP telephony. It can be used to increase speed, scalabil-ity, and functionality for emergency calling and notification systems. In 2001, IETF,working with ITU-T, published megaco protocol (known as H.248 in ITU-T)which is even less end-to-end than H.323. It is designed to support the creationof IP-based phone switches that could mimic traditional phone switches. In 2002,IETF approved publication of an update to RFC 2543 (also, RFC 3264) which doesnot define a new protocol but rather cleans up the details of the old specification,particularly in security. Whether it can be widely accepted depends on how peoplesee its future: If it is seen only as a different way to provide the current conceptof phone service or only a subset of what can be provided today with standardequipment, then there is little incentive to deploy it.

Networks typically have hardware and software from a number of different ven-dors which must communicate with each other by “speaking the same language”and following the same protocols. Unfortunately, commercially available data com-munication devices speak a variety of languages and follow a number of differentprotocols, causing substantial problems with data communications networks.

Attempts at standardizing data communications have been somewhat suc-cessful, but standardization in the United States has lagged behind other countrieswhere the communications industry is more closely regulated. Various organiza-tions, including the Electronic Industries Association (EIA), the Consultative Com-mittee for International Telegraph and Telephone (CCITT), and the InternationalStandards Organization (ISO), have developed electronic interfacing standards thatare widely used within the industry. The major types of standards are networkingstandards, transmission standards, and software standards.

NETWORKING STANDARDS. Typically, the protocols required to achieve com-munication on behalf of an application are actually multiple protocols existingat different levels or layers. Each layer defines a set of functions that are pro-vided as services to upper layers, and each layer relies on services provided bylower layers. At each layer, one or more protocols define precisely how softwareprograms on different systems interact to accomplish the functions for that layer.

This layering notion has been formalized in several architectures. The mostwidely known is the Open Systems Interconnection (OSI) Reference Model developed bythe ISO. There is peer-to-peer communication between software at each layer,

CommunicationsStandards

t04.qxd 01/27/2005 04:43PM Page T4.22 EQA


and each relies on underlying layers for services to accomplish communication.The OSI model has seven layers, each having its own well-defined function:

● Layer 1: Physical layer. Concerned with transmitting raw bits over a com-munications channel; provides a physical connection for the transmission ofdata among network entities and creates the means by which to activate anddeactivate a physical connection.

● Layer 2: Data link layer. Provides a reliable means of transmitting data acrossa physical link; breaks up the input data into data frames sequentially andprocesses the acknowledgment frames sent back by the receiver.

● Layer 3: Network layer. Routes information from one network computer toanother; computers may be physically located within the same network orwithin another network that is interconnected in some fashion; accepts mes-sages from source host and sees to it they are directed toward the destination.

● Layer 4: Transport layer. Provides a network-independent transport serviceto the session layer; accepts data from session layer, splits it up into smallerunits as required, passes these to the network layer, and ensures all piecesarrive correctly at the other end.

● Layer 5: Session layer. User’s interface into network; where user mustnegotiate to establish connection with process on another machine; onceconnection is established the session layer can manage the dialogue in anorderly manner.

● Layer 6: Presentation layer. Here messages are translated from and to the for-mat used in the network to and from a format used at the application layer.

● Layer 7: Application layer. Includes activities related to users, such as sup-porting file transfer, handling messages, and providing security.

The SNA Standard. The Systems Network Architecture (SNA) is a standarddeveloped by IBM that is widely used in private networks. Similar to OSI, SNAuses a layered approach; however, the layers are somewhat different. Certaintypes of expected errors, like those occurring over phone lines, are handledautomatically. Other errors, like software problems, are isolated, logged, andreported to the central technical staff for analysis.

TRANSMISSION STANDARDS. A number of network bandwidth boostersaddress the need for greater bandwidth on networks for advanced computingapplications. These transmission technologies are discussed below.

FDDI. Like token-ring networks, the fiber distributed data interface(FDDI) passes data around a ring, but with a bandwidth of 100 Mpbs—muchfaster than a standard 10–13 Mbps token-ring or bus network. Although theFDDI standard can use any transmission medium, it is based on the high-speed,high-capacity capabilities of fiber optics. FDDI can significantly boost networkperformance, but this technology is about ten times more expensive to imple-ment than most local area networks (LANs).

ATM. Asynchronous transfer mode (ATM) networks are based onswitched technologies, allowing for almost unlimited bandwidth on demand. Ita service that carries voice, data, video, and multimedia at extremely highspeeds. They are packet-switched networks, dividing data into uniform cells,each with 53 bytes, eliminating the need for protocol conversion. ATM allowsmixing of varying bandwidths and data types (e.g., video, data, and voice) and

t04.qxd 01/27/2005 04:43PM Page T4.23 EQA


much higher speeds because the data are more easily “squeezed” in among othervery small packets. ATM currently requires fiber-optic cable, but it can transmitup to 2.5 gigabits per second. On the downside, ATM is more expensive thanISDN and DSL.

Switched hub technologies. Switched hub technologies are often usedto boost local area networks. A switched hub can turn many small LANs intoone big LAN. A network need not be rewired nor adapter cards replaced whenchanges are made; all that is needed is the addition of a switching hub. Switchedhub technology can also add an ATM-like packet-switching capability to existingLANs, essentially doubling bandwidth.

SONET. Synchronous optical network (SONET) is an interface stan-dard for transporting digital signals over fiber-optic links that allows the inte-gration of transmissions from multiple vendors. SONET defines optical line rates,known as optical carrier (OC) signals. The base rate is 51.84 Mbps (OC-1), andhigher rates are direct multiples of the base rate. For example, OC-3 runs at155.52 Mbps, or three times the rate of OC-1.

T-carrier system. The T-carrier system is a digital transmission systemthat defines circuits that operate at different rates, all of which are multiples ofthe basic 64 Kbps used to transport a single voice call. These circuits include T1(1.544 Mbps, equivalent to 24 channels); T2 (6.312 Mbps, equivalent to 96channels); T3 (44.736 Mbps, equivalent to 672 channels); and T4 (274.176Mbps, equivalent to 4032 channels). The T-carrier system uses a techniquecalled multiplexing which enables multiple signals to share the telephone line.The T1 line is the most popular; the T3 line, which is used for the Internet back-bone, is 28 times faster than T1.

ISDN. Integrated services digital network (ISDN) is a high-speed datatransmission technology that allows users to simultaneously transfer voice,video, image, and data at high speed over standard copper telephone lines, usingmultiplexing. ISDN converges all overlapping information networks (telephone,telegraph, data, cable TV, paging, and personal communications services) to anintegrated system. The user can see an incoming telephone message notice andtalk on the phone, while on her PC.

ISDN provides two levels of service: basic-rate ISDN and primary-rate ISDN.Basic-rate ISDN serves a single device with three channels. Two channels are B(bearer) channels with a capacity to transmit 64 Kbps of digital data. The thirdor D channel is a 16-Kbps channel for signaling and control information.Primary-rate ISDN provides 1.5 Mbps of bandwidth. The bandwidth contains 23B channels and one D channel.

A second generation of ISDN is broadband ISDN (BISDN), which providestransmission channels capable of supporting transmission rates greater than theprimary ISDN rate. BISDN supports transmission from 2 Mbps up to muchhigher, but as yet unspecified, rates.

DSL. A Digital Subscriber Line (DSL) provides high-speed, digital datatransmission from homes and businesses over existing telephone lines. It trans-mits at fast speeds on existing standard copper telephone wiring. Some instal-lations can be used for both voice and data. The existing lines are analog andthe transmission is digital, so modems are necessary with DSL technology. Usedunder similar circumstances, DSL is a popular alternative to ISDN. ADSL (asym-metric digital subscriber line) supports faster transfer rates when receiving datathan when sending data.

t04.qxd 2/16/05 12:15 PM Page T4.24 EQA


Infinite Bandwidth. Infinite Bandwidth (InfiniBand) is a new standarddesigned to dramatically increase the velocity of information by overhauling a keybottleneck–today’s general-purpose, shared bus inside the computer. That sharedbus, named the Peripheral Component Interconnect (PCI) bus can carry one mes-sage at a time past many points. The new standard, called a switched fabric net-work, will able to juggle hundreds or thousands of messages at a time both insideand outside the computer, moving them precisely from origin to destination. Besidesfaster data throughput by 10 times, InfiniBand may allow redesign of the computeritself. Table T4.6 shows the difference between the two buses architecture.

Circuit Switching. Circuit switching is an end-to-end circuit that must be setup before the call can begin. A fixed share of network resources is reserved forthe call, and no other call can use those resources until the original connectionis closed. This enables performance guarantees and is much easier to do detailedaccounting for circuit-switched network usage.

SOFTWARE STANDARDS. In order for computers and computing devices fromdifferent vendors to conveniently “talk” to each other, they need an opensystem. Three types of software standards are necessary for an open system:

● Operating systems. A network operating system (NOS) is the system soft-ware that controls the hardware devices, software, and communications me-dia and channels across a network. The NOS enables various devices to com-municate with each other. NetWare by Novell and Windows XP fromMicrosoft are popular network operating systems for LANs.

● Graphical User Interface standard. X Windows is the standard for GUI. Itruns on all types of computers and is used with Unix and the DEC VAX/VMSoperating systems. It permits one display of several applications on onescreen and allows one application to use several windows.

● Software application standards. Because of the large number of applications,it takes more time to reach agreements on standards. In the meantime, theU.S. government is attempting to establish a standard for all of its softwareapplications. The unified standards will cover DBMSs, user interfaces, pro-gramming languages, electronic data interchange, and so on.

Network management software comes in different shapes, and it has manyfunctions in operating a network. These functions reduce time spent on routinetasks, such as remote, electronic installation of new software on many devices

TABLE T4.6 The Difference between PCI Bux and Infiniband Bus

PCI InfiniBand

Communication mode Has the problem of handling InfiniBand will supplant communications with only the PCI bus and overhaulone device at a time that a the I/O architecture of faulty peripheral card may servershang the whole server

Data transfer speed Uses parallel bus which has Uses serial bus which hasspeed of 1Gbps speed of 2.5Gbps

Network protocol Act as a local input/output Act as a distributed I/Obus for a single server architecture

NetworkManagement

Software

t04.qxd 2/16/05 12:15 PM Page T4.25 EQA


across a network. They also provide faster response to network problems,greater control over the network, and remote diagnosing of problems in devicesconnected to the network. In short, network management software performsfunctions that decrease the human resources needed to manage the network.

ACTIVE DIRECTORY. Active Directory is a software product created by Microsoftthat provides applications with a single user interface for accessing network direc-tory services from multiple operating systems. An Active Directory service appli-cation interacts with the directory services of each operating system through theset of interfaces supported for each namespace. Because many of the administra-tive tasks and directory services are performing across namespaces, a similar setof interfaces is supported by each directory service implementation on the organ-ization’s network. Active Directory is a key component of Microsoft’s WindowsServer 2003 operating system which simplifies management tasks, enhances datasecurity, and extends interoperability with other networked operating systems.For details, see microsoft.com/windowsserver2003/technologies/directory/activedirectory/default.mspx.

An interface is a physical connection between two communications devices.One important concept of interfacing concerns the types of data transfer—parallel or serial. Parallel data transfer, most often used for local communi-cation, employs a communications interface with a series of dedicated wires,each serving one purpose. In parallel communication, both data and controlsignals are transmitted simultaneously.

A serial data transfer, most often used for long-distance communications,is bit by bit rather than many bits in parallel. Most data communicationsdevices transmit in serial fashion. While much slower than parallel data trans-fer, serial transfer is simpler and requires much less on the part of the receivingsystem.

The topology of a network is the physical layout and connectivity of a net-work. Specific protocols, or rules of communications, are often used on specifictopologies, but the two concepts are different. Topology refers to the ways thechannels connect the nodes, whereas protocol refers to the rules by which datacommunications take place over these channels. Neither concept should beconfused with the physical cabling of the network.

There are several basic network topologies: star, bus, ring, mesh, and hier-archical. Figure T4.5 illustrates these different types. Hierarchical topologies typ-ically connect desktops and minicomputers to a mainframe. Networks that com-bine more than one type (such as a ring segment connected to a star segment)are considered hybrid topologies. We discuss the various topologies in moredetail below.

STAR. A star network has a central node that connects to each of the othernodes by a single, point-to-point link. Any communication between one nodeand another in a star topology must pass through the central node. It is easyto add a node in a star network, and losing a peripheral node will not causethe entire network to fail. However, the central computer must be powerfulenough to handle communications in a star network. Too many devices in thestar can overload the central computer and cause degradation in performance

Interfaces

Network Topology

t04.qxd 2/16/05 4:18 PM Page T4.26 EQA


across the network. The star topology is typically used in low-cost, slow-speeddata networks.

BUS. In a bus topology, nodes are arranged along a single length of twisted-pair wire, coaxial cable, or fiber-optic cable that can be extended at the ends.Using a bus topology, it is easy and inexpensive to add a node to the network,and losing a node in the network will not cause the network to fail. The maindisadvantages to the bus topology are that a defective bus causes the entire net-work to fail. Also, providing a bus with inadequate bandwidth will degrade theperformance of the network.

RING. In a ring topology, nodes are arranged along the transmission path sothat a signal passes through each station one at a time before returning toits originating node. The nodes, then, form a closed circle. It is relatively easyand inexpensive to add a node to the network, and losing a node does notnecessarily mean that the network will fail.

MESH. A mesh network design is one in which each device is connected toevery other device located on the network, like a spider web. The advantage tothis design is the redundancy of the connected devices; if one link fails, it willnot affect the rest of the network. The disadvantages of this design are the costof all the required medium and limited scalability. If you add a device to a net-work that currently has four devices, then you must connect the new device tothe four existing devices with individual cable drops.

FIGURE T4.5 The mainnetwork topologies.

(a) Star network (b) Bus network

(e) Hierarchical network

(c) Ring network

Mainframe

(d) Meshnetwork design

Printer

Server

Computer

Computer

Computer

Minicomputers

Desktop computers

t04.qxd 01/27/2005 04:43PM Page T4.27 EQA


HIERARCHICAL. In a hierarchical topology, nodes are arranged like an invertedtree with the root (usually the mainframe computer) as the highest level and theleaves (usually the desktop computers) as the lowest level. It is very cheap, butmay have possible traffic jams at the top level.

An advantage of the hierarchical topology is its ability to scale to very largenetworks. This scalability is because of the exponential reduction in size of thevisible topology and amount of received topology state information at eachswitch in the network. These reductions improve the effectiveness of your net-work by reducing the control traffic, memory, and processing required by eachswitch in the network.

A disadvantage of the hierarchical topology is the loss of informationcaused by topology aggregation. Because a hierarchical view of the networkis restricted, compared to a nonhierarchical (flat topology) view, routingdecisions are not as effective as in a flat topology. In both cases, a path to thedestination is selected; however, in most cases the path selected in a flat topol-ogy is more efficient.

The decision to implement a hierarchy depends on many factors, includingthe size of the network, type of network traffic, call setup activity, and theamount of processing and memory required to handle control traffic.

HYBRID. In a hybrid topology, nodes are arranged in more than one topologywhich may include star, ring, and hierarchical (see Figure T4.6). A hybrid topol-ogy can integrate together various computer configurations that may have spe-cial reasons for their own choice of topology (as mentioned in the precedingsections). A hybrid network will allow companies to pick the advantages fromseveral different topologies.

Each topology has strengths and weaknesses. When systems developerschoosea topology, they should consider such performance issues as delay, speed,reliability, and the network’s ability to continue through, or recover after, the fail-ure of one or more of its nodes. A company should also consider such physicalconstraints as the maximum transmission speed of the circuit, the distancesbetween nodes, the circuit’s susceptibility to errors, and the overall system costs.

Hybrid network

Mainframe

Minicomputers

FIGURE T4.6 A hybridnetwork topology.

t04.qxd 2/16/05 12:15 PM Page T4.28 EQA


Because people need to communicate over long as well as short distances, thegeographic size of data communications networks is important. There are twogeneral network sizes: local area networks and wide area networks. A“metropolitan” area network falls between the two in size. In addition, homenetworks are a type of LAN.

LOCAL AREA NETWORKS. A local area network (LAN) connects two ormore communicating devices within a short distance (e.g., 2000 feet), so thatevery user device on the network has the potential to communicate with anyother device. LANs are usually intraorganizational, privately owned, inter-nally administered, and not subject to regulation by the FCC. LANs do notcross public rights-of-way and do not require communications hardware andsoftware that are necessary to link computer systems to existing communi-cations networks. A LAN allows a large number of its intelligent devices toshare corporate resources (such as storage devices, printers, programs, anddata files), and it integrates a wide range of functions into a single system.Many LANs are physically connected as a star, with every device connectedto a hub or switch.

In an office, a LAN can give users fast and efficient access to a common col-lection of information while also allowing the office to pool resources, such asprinters and facsimile machines. A well-constructed LAN can also eliminate theneed to circulate paper documents by distributing electronic memos and othermaterials to each worker’s terminal.

The LAN file server is a repository of various software and data files for thenetwork. The server determines who gets access to what and in what sequence.Servers may be powerful microcomputers with large, fast-access hard drives, orthey may be workstations, minicomputers, or mainframes. The server typicallyhouses the LAN’s network operating system, which manages the server androutes and manages communications on the network.

The network gateway connects the LAN to public networks or other cor-porate networks so that the LAN can exchange information with networksexternal to it. A gateway is a communications processor that can connect dis-similar networks by translating from one set of protocols to another. A bridgeconnects two networks of the same type. A router routes messages throughseveral connected LANs or to a WAN.

A LAN consists of cabling or wireless technology linking individual devices,network interface cards (special adapters serving as interfaces to the cable),and software to control LAN activities. The LAN network interface card specifiesthe data transmission rate, the size of message units, the addressing informationattached to each message, and the network topology.

LANs employ a baseband or a broadband channel technology. In basebandLANs, the entire capacity of the cable is used to transmit a single digitally codedsignal. In broadband LANs, the capacity of the cable is divided into separatefrequencies to permit it to carry several signals at the same time.

Private Branch Exchange. A private branch exchange (PBX) is a type ofLAN. The PBX is a special-purpose computer that controls telephone switchingat a company site. PBXs can carry both voice and data and perform a wide vari-ety of functions to make communications more convenient and effective, suchas call waiting, call forwarding, and voice mail. PBXs also offer functions directedat decreasing costs, such as reducing the number of outside lines, providing

NetworkArchitecture and

Size

t04.qxd 2/16/05 12:15 PM Page T4.29 EQA


internal extensions, and determining least-cost routings. Automatic assignmentof calls to lines reduces the required number of outside lines. Providing internalextension numbers permits people to make calls within the same site using onlyextension numbers and without making a chargeable outside call.

A new technology developed by MCK Communications works to extend thecorporate telephone network to anywhere in the world. The Mobile Extendergives any wireless phone user access to the private branch exchange (PBX) inways not possible before. Its main task is to extend the voice network to anylocation. It can turn any analog phone, touch-tone phone, or cell phone into adigital extension into the PBX using Internet protocal (IP) as the underlyingtechnology. MCK manufactures a box that sits beside the PBX to packetize voiceand send it over a dedicated IP backbone. A box on the end of the networkclosest to the user then depacketizes the voice and sends over the wireless net-work. Moreover, MCK also includes software that allows the user to programPBX applications into the phone. This allows the wireless user to hit *8 totransfer a call or dial 9 to get an outside line.

Wireless local area networks (WLANs). WLAN technologies provide LANconnectivity over short distances, typically limited to less than 150 meters, andusually within one building.

Bluetooth technology. As discussed earlier in the chapter, Bluetooth is awireless technology that allows digital devices such as computers, printers, key-boards, cell phones, and Palm Pilots to communicate with each other over shortdistance (less than 500 feet, or 150 meters) via low-power/milliwatt radio fre-quencies. Bluetooth can also form a home network by linking devices like lights,televisions, the furnace and air conditioning, and the garage door. Bluetooth isnot line-of-sight, which means that transmissions may occur around corners,through walls, and through briefcases. Limiting factors for adoption of Bluetoothinclude security, transmission speed (Bluetooth maximum transmission speed is720 Kbps for half-duplex link), and relatively high cost.

METROPOLITAN AREA NETWORKS. A metropolitan area network (MAN)is a network that interconnects users with computer resources in a geographicarea or region larger than that covered by a local area network (LAN) (evena large LAN), but smaller than the area covered by a wide area network(WAN). The term is applied to the interconnection of networks in a city intoa single larger network (which may then also offer efficient connection to awide area network). It is also used to mean the interconnection of severallocal area networks by bridging them with backbone lines, which are largertransmission lines that carry data gathered from smaller lines that intercon-nect with them. The latter usage is also sometimes referred to as a campusnetwork.

Three important features differentiate MANs from LANs or WANs:

1. The network size falls part-way between LANs and WANs. A MAN typicallycovers an area of between 5- and 50-km diameter. Many MANs cover an areathe size of a city, although in some cases MANs may be as small as a groupof buildings or as large as the North of Scotland.

2. A MAN (like a WAN) is not generally owned by a single organization. Eithera consortium of users or a single network provider that sells the service to theusers generally owns the MAN, its communications links, and equipment.

t04.qxd 2/16/05 4:18 PM Page T4.30 EQA


Each user must therefore negotiate with the MAN operator the level of serv-ice to be provided by the operator, and some performance guarantees are nor-mally specified.

3. A MAN often acts as a high-speed network to allow sharing of regional re-sources (similar to a large LAN). It is also frequently used to provide a sharedconnection to other networks using a link to a WAN.

A typical use of MANs to provide shared access to a wide area network is shownin Figure T4.7.

HOME NETWORKS. Home networks are the computer-networking infra-structure installed at home. The components of a home network are very similarto those used in an office network, but the scale is much smaller. By connectingtheir home computers into a network, users can:

● Share a single printer or expensive equipment between computers.

● Share a single Internet connection among all the computers in your home.

● Access shared files such as photographs, MP3s, Excel spreadsheets, and Worddocuments on any computer in the house.

● Play games that allow multiple users at different computers.

● Send the output of a device like a DVD player or Webcam to the home’sother computer(s).

WIDE AREA NETWORKS. Wide area networks (WANs) are long-haul, broad-band, generally public-access networks covering wide geographic areas that crossrights-of-way where communications media are provided by common carriers.WANs include regional networks such as telephone companies or international net-works such as global communications service providers. They usually have verylarge-capacity circuits with many communications processors to use these cir-cuits efficiently. WANs may combine switched and dedicated lines, microwave,and satellite communications.

A leased line may handle data only, or it may be capable of handling bothvoice and data just as a standard telephone line does. When leased lines havebeen designed specifically for data transmission, they produce less noise andfewer transmission errors than regular telephone lines, and they are moresecure from wiretapping and other security risks. Most importantly, the

MetropolitanAreaNetwork

To othernetworks

Wide AreaNetwork

Local AreaNetworks

Local AreaNetworks

MetropolitanAreaNetwork

FIGURE T4.7 Use ofMANs to provide regionalnetworks, which share thecost of access to a WAN.(Source: http://www.erg.abdn.ac.uk/users/gorry/eg3561/intro-pages/man.html.)

t04.qxd 01/27/2005 04:43PM Page T4.31 EQA


central processor is always accessible through the leased line, and the lineusually transmits data at speeds (e.g., 1.544 Mbps) faster than a standard tele-phone line.

Some WANs are commercial, regulated networks, while others are privatelyowned, usually by large businesses that can afford the costs. Some WANs, how-ever, are “public” in terms of their management, resources, and access. One suchpublic WAN is the Internet, the foundation of the worldwide information super-highway.

Value-Added Networks. A value-added network (VAN) is a type ofWAN. VANs are private, multipath, data-only, third-party-managed networksthat can provide economies in the cost of service and network managementbecause they are used by multiple organizations. VANs can add teleconferenc-ing services and message storage, tracking, and relay services, and can moreclosely tailor communications capabilities to specific business needs.

VANs offer several valuable services. For example, customers do not haveto invest in network hardware and software or perform their own error check-ing, editing, routing, and protocol conversion. Subscribers realize savings in linecharges and transmission costs because many users share the costs of using thenetwork.

Value-added networks also provide economies through packet switching.Packet switching breaks up a message into small, fixed bundles of data (aboutaround 200 bytes in size) called packets, which are then sent out onto the net-work. Each independent packet contains a header with information necessaryfor routing the packet from origination to destination on a first-come-first-served basis. Packet switching is used extensively in Internet communication. Ithas the advantage of permitting ”statistical multiplexing” on the communica-tions lines, in which packets from many different sources can share a line, thusallowing for very efficient use of the fixed capacity.

The VAN continuously uses various communications channels to send thepackets. Each packet travels independently through the network. Packets of dataoriginating at one source can be routed through different paths in the network,and then may be reassembled into the original message when they reach theirdestination.

Frame Relay. Frame relay is a faster and less expensive version of packetswitching. Frame relay is a shared network service that packages data into“frames” that are similar to packets. Frame relay, however, does not performerror correction, because modern digital lines are less error-prone than olderlines and networks are more effective at error checking. Frame relay cancommunicate at speeds of 50 megabits per second.

Frame relay is essentially used for transmitting data. It is not recommendedfor any transmissions that are sensitive to varying delay, such as voice or digitalvideo traffic, and it cannot easily control network congestion. Some companies douse voice-over frame relay, however, because of its low cost. Frame relay is rap-idly replacing leased lines as the primary form of long-distance communicationin WANs.

Virtual Private Networks. A virtual private network (VPN) is a networkthat exists on the hardware, software, and media of a communications carrier(e.g., Sprint) or an ISP (Internet service provider), but it looks to the customeras if he or she owns it. VPN uses a public network (usually the Internet) to con-nect remote sites or users together, using a tunneling technique to encapsulate

t04.qxd 2/16/05 12:15 PM Page T4.32 EQA

T4.4 NETWORK ARCHITECTURE T4.33

the messages. A VPN provides a link between a corporate LAN and the Internetand is a means for allowing controlled access to a private network’s e-mail,shared files, or intranet via an Internet connection. The VPN provider handlessecurity (e.g., encryption, authentication), thus permitting access from the Inter-net to an intranet. The data, from business partners or corporate commuters,travels over the Internet—but it is encrypted. To provide this level of securitywithout a VPN it might otherwise be necessary to make a long-distance call toconnect to a remote access service (RAS) dialup. With VPN, this cost is elimi-nated. In addition, VPN is less expensive than RAS and can provide more ser-vices. VPNs are a fraction of the cost of the dedicated leased lines used for remoteaccess. It uses ‘tunneling’ to establish a secure connection between the PC andthe corporate network gateway.

There are several types of VPNs on the market, and they are priced dif-ferently. VPNs are especially suited for extranets, since they allow the use ofthe Internet between business partners instead of using a more expensive VAN.VPNs are especially important for international trade where long-distancecalls, or VANs, are very expensive. An SSL-VPN combination is becoming morepopular.

T4.4 NETWORK ARCHITECTURE

When two devices on a network are directly connected, there is a point-to-point connection. When the two devices have the same relative standing, aswith two PCs, there is a peer-to-peer connection. So, for example, if thereis a point-of-sale terminal dialing into a credit card checking mainframe viamodem, there is a point-to-point, but not a peer-to-peer connection. In thisexample, there is a client/server, point-to-point network.

CENTRALIZED ARCHITECTURE. Centralized computer systems are centered arounda large computer, known as the host, that provides computational power andinternal storage. Several devices that lack self-contained computer processors,such as dumb terminals and printers, are connected to the host. Information isentered, distributed, stored, or communicated through these devices. There arefour basic types of these devices for direct, temporary interaction with people(usually via a typewriter-like keyboard and on an electronic monitor or screen):output devices such as printers and plotters for generating permanent outputs; inputdevices such as bar-code scanners for reading data; communications devices forexchanging data with other computer systems; and storage devices for electronicallystoring data. (See Technology Guide 1 for a review of computer hardware.) Ineach case, the host computer, usually a mainframe, is responsible for computerprocessing, and it is centrally connected to each device. All information process-ing is orchestrated by the host, and much of it is carried out by the devices. Thisarrangement (see Figure T4.8) is simple, direct, and easily controlled.

Although mainframes represented the dominant centralized form of com-puting for over 30 years, minicomputers, workstations, and powerful PCs arechallenging that dominance. Centralized computing, as an architecture, caninclude all sizes of computer processors, including a conglomeration of com-puters acting in parallel. Mainframes, by themselves, no longer “rule the roost,”but they are still an important part of an IT architecture, along with smaller

TechnologyArchitecture

t04.qxd 01/27/2005 04:43PM Page T4.33 EQA


computers ranging down to palmtops. Whereas architectural decisions were rel-atively simple with one computer—the mainframe—they are now much morevaried and complex with a wide range of computers available.

There are still many business applications for which mainframes reign. Forhigh-volume, rapid-pace, transaction-intensive applications—such as airlinereservation systems or stock trading systems—the mainframe still plays a vitalrole. Continuous availability and rapid response, as well as the benefits ofreliability and security, are all available from the typical mainframe package,making it worthy of ongoing attention.

NONCENTRALIZED COMPUTING. Noncentralized computing architectures aredecentralized or distributed. Decentralized computing breaks centralizedcomputing into functionally equivalent parts, with each part essentially asmaller, centralized subsystem. Almost all telephone utilities operate this way.Local switching stations contain local, centralized computers for the telephonesin their immediate areas—each switching center is functionally identical. Dist-ributed computing, on the other hand, breaks centralized computing intomany computers that may not be (and usually are not) functionally equivalent.For a bank with many regional centers, for example, one location may processloan applications, another foreign currency transactions, and another businessand individual deposit accounts. All branches can access all data, but certaincomputing functions are distributed across different regional centers.

As smaller, midrange computers (commonly called minicomputers) appeared,and as businesses increasingly required systems capable of sharing informationresources, computing evolved toward peer-to-peer architectures. In this architec-ture, which is basically distributed computing, one computing resource sharesprocessing with other computing resources. As peers, they also share devices anddata with each other, although one “peer” may be more important than anotherfor certain tasks or for controlling devices such as printers. Such peer-to-peerrelationships became even more common as PCs began to proliferate in offices,and as they were linked together for communications (see discussion below).

There is nothing inherently good or bad about a decision to centralize ver-sus a decision to distribute an IT architecture. Instead, there are benefits andlimitations to each approach. Most organizations would normally be categorized

FIGURE T4.8 An exampleof a centralized architec-ture. (Source: From SystemsAnalysis and Design, FourthEdition, by Wetherbe andVitalare © 1988. Reprintedwith permission of South-Western College Publishing, adivision of ThomsonLearning.)

Data archive Printer

Terminal Terminal

TerminalCommunication device

Mainframe

t04.qxd 01/27/2005 04:43PM Page T4.34 EQA


somewhere along a continuum between the extremes of completely centralizedand completely distributed.

The basic structure of client/server architecture is a client device(s) and a serverdevice(s) that are distinguishable, but interact with each other. (That is, inde-pendent processing machines in separate locations perform operations separately,but can operate more efficiently when operating together.) This architecturedivides processing between “clients” and “servers” (see Figure T4.9). Both are onthe network, but each processor is assigned functions it is best suited to perform.An example of a client might be a desktop PC used by a financial analyst torequest data from a corporate mainframe. Or it might be a laptop used by a sales-person to get pricing and availability information from corporate systems, calcu-late and print an invoice, and order goods directly—all from a client’s office.

In a client/server approach, the components of an application are distrib-uted over the enterprise rather than being centrally controlled. There are threeapplication components that can be distributed: the presentation component,the applications (or processing) logic, and the data management component.The presentation component is the application interface or how the applicationappears to the user. The applications logic is created by the business rules of theapplication. The data management component consists of the storage and manage-ment of the data needed by the application. The exact division of processingtasks depends on the requirements of each application including its processingneeds, the number of users, and the available resources.

There are five models of client/server implementation, depending on thepartitioning of the three components between the server and the client.

1. Distributed presentation, in which all three components are on the server,but the presentation logic is distributed between the client and the server.

2. Remote presentation, in which applications logic and database manage-ment are on the server, and the presentation logic is located on the client.

3. Distributed function, in which data management is on the server andpresentation logic is on the client, with application logic distributed betweenthe two.

Client/ServerArchitecture

FIGURE T4.9 An exampleof client/server architec-ture. (Source: From SystemsAnalysis and Design, FourthEdition, by Wetherbe andVitalare ©1988. Reprintedwith permission of South-Western College Publishing,a division of ThomsonLearning.)

Printer (server)

Client Client

Data archive (server)

Server

Client

t04.qxd 01/27/2005 04:43PM Page T4.35 EQA


4. Remote data management, in which database management is on theserver, with the other two components on the client.

5. Distributed data management, in which all three components are onthe client, with database management distributed between the client and theserver.

These models led to the ideas of “fat” clients and “thin” clients. Fat clientshave large storage and processing power, and the three components of an appli-cation can be processed. Thin clients may have no local storage and limitedprocessing power, meaning that thin clients can handle presentation only (suchas in a network computer).

The client is the user point-of-entry for the required function, and the usergenerally interacts directly with only the client portion of the application, typ-ically through a graphical user interface. Clients call on the server for servicesrendered by the application software. When a client needs information basedon files in the server, the client sends a message (or remote procedure call)requesting the information from the server.

The server has the data files and the application software. The server storesand processes shared data and performs “back-end” functions not visible tousers, such as managing peripheral devices and controlling access to shared data-bases. When a client makes a remote procedure call, the server processes filedata and provides the information, not the entire file(s), to the client.

Early client/server systems were primarily used for non-mission-criticalapplications due to experiences with (or fears of) poor system stability and lackof robustness. More recently, studies have shown a significant increase inthe number and size of installed business client/server systems—especially forcritical online transaction processing.

When the client/server architecture first appeared, many analysts believedthat it would lower costs because the typical servers were much less expensive(in relation to processing power) than mainframe computers. However, becauseof the complexities of linking many different types of hardware and software, aswell as the relative immaturity of the technology, client/server applications mayactually be substantially more expensive to operate than mainframe applications.The International Technology Group estimated that computing on a client/serversystem cost an average of $6,982 per user per year, versus $2,127 on a main-frame system. However, other experts provided substantially different results.

In summary, client/server architectures offer the potential to use computingresources more effectively through division of labor among specialized processingunits. A client desktop PC can use its strengths for calculations and visually dis-playing results, but it can also access the specialized functionality of a file serverfor the required data. This configuration also facilitates sharing hardware such asprinters and plotters, and sharing data throughout the organization via an intranet.The client/server model fits well in an environment of disparate computing needsthat are distributed throughout an organization, in an environment that is unsta-ble or changes frequently, and in an environment characterized by risk and uncer-tainty. In today’s business environments, we can fully expect that client/serverarchitectures will continue to increase in popularity and sophistication.

A peer-to-peer network architecture allows two or more computers to pooltheir resources together. Individual resources like disk drives, CD-ROM drives,and even printers are transformed into shared, collective resources that are

Peer-to-PeerArchitecture

t04.qxd 01/27/2005 04:43PM Page T4.36 EQA


accessible from every computer. Unlike client/server networks, where networkinformation is stored on a centralized file server and made available to manyclients, the information stored across peer-to-peer networks is uniquelydecentralized. Because peer-to-peer computers have their own disk drives thatare accessible by all computers, each computer acts as both a client and a server.Each computer has transparent access (as assigned for security or integritypurposes) to all files on all other computers (see Figure T4.10).

Popular peer-to-peer network operating systems include Microsoft’s Win-dows, Artisoft’s LANtastic, and Netware Lite. Most of these operating systemsallow each computer to determine which resources will be available for use byother users. When one user’s disk has been configured so that it is “sharable,”it will usually appear as a new drive to the other users.

There are several advantages of peer-to-peer architecture: The network isthe easiest to build, and it is fast and inexpensive to maintain. Each computercan make backup copies of its files to other computers for security. Finally, thereis no need for a network administrator.

Peer-to-peer architecture, also known as P2P, is extremely popular on theInternet. Probably the best-known P2P application is music sharing managed byNapster, KaZaa, and others.

There are three basic types of peer-to-peer processing. The first accessesunused CPU power among networked computers, as in applications such asNapster, Gnutella, and SETI@home. These applications are from open-source proj-ects and can be downloaded at no cost. The second form of peer-to-peer is real-time, person-to-person collaboration, such as America Online’s Instant Messen-ger. Companies such as Groove Networks (groove.net) have introduced P2Pcollaborative applications (e.g., Virtual Office) that use buddy lists to establish aconnection, then allow real-time collaboration within the application. The thirdpeer-to-peer category is advanced search and file sharing. This category is char-acterized by natural-language searches of millions of peer systems and lets usersdiscover other users, not just data and Web pages.

Open systems are those that allow any computing device to be seamlessly con-nected to and to interact with any other computing device, regardless of size,operating system, or application. This has been a goal for information systems

FIGURE T4.10 An exam-ple of peer-to-peer archi-tecture. (Source: FromSystems Analysis and Design,Fourth Edition, by Wetherbeand Vitalare © 1988. Re-printed with permission ofSouth-Western CollegePublishing, a division ofThomson Learning.)

Open Systems andEnterprise

Networking

Printer Data archive

Mainframe

PC PC

Minicomputer

t04.qxd 01/27/2005 04:43PM Page T4.37 EQA


designers for many years, and it is just now becoming a (limited) reality. Opensystems can provide flexibility in implementing IT solutions, optimization ofcomputing effectiveness, and the ability to provide new levels of integrated func-tionality to meet user demands. As discussed earlier in this Tech Guide, opensystems require connectivity across the various components of the system.

Connectivity is the ability of the various computer resources to commu-nicate with each other through network devices without human intervention.Connectivity allows for application portability, interoperability, and scalability.Portability is the ability to move applications, data, and even people from onesystem to another with minimal adjustments. Interoperability refers to theability of systems to work together by sharing applications, data, and computerresources. Scalability refers to the ability to run applications unchanged onany open system where the hardware can range from a laptop PC to a super-computer.

Open systems and connectivity have enabled networks to completely spanorganizations. Most firms have multiple LANs and may have multiple WANs,which are interconnected to form an enterprisewide network (see Figure T4.11).Note that the enterprisewide network shown in Figure T4.11 has a backbonenetwork composed of fiber-optic cable using the FDDI protocol. The LANs arecalled embedded LANs because they connect to the backbone WAN. TheseLANs are usually composed of twisted-pair wire.

Babcock, C., “InfiniBand Aims to Smash the Data Bind,” InteractiveWeek, April 23, 2001.

Baca, H., et al., Local Area Networks with Novell. Belmont, CA:Wadsworth, 1995.

Bradner, S., “Internet Telephony,” IEEE Internet Computing, May–June 2002.

“Broadband,” pcworld.com, February 2002 (accessed May 16,2003).

Burns, S., “The Web In Your Hand,” Far Eastern Economic Review,June 8, 2000.

Comer, D. E., and R. E. Droms, Computer Networks and Internets.Upper Saddle River, NJ: Prentice Hall, 1996.

Dodd, A. Z., The Essential Guide to Telecommunications, 2nd ed. UpperSaddle River, NJ: Prentice Hall, 2000.

Edwards, J., “Building the Optical Networking Infrastructure,”Computer, March 2000, pp. 20–23.

Fitzgerald, J., Business Data Communications: Basic Concepts, Security,and Design, 5th ed. New York: Wiley, 1996.

Foley, T., “Time to Beam Up,” Business2.com, November 28, 2000.

Gauntt, J., et al., “Technology Forecast 2000,” Pricewaterhouse-Coopers, 2000.

Goldman, R., Business Data Communications, 3rd ed. New York:Wiley, 2001.

Graham, S., “Taking the Metro,” Interactive Week, July 9, 2001.

FIGURE T4.11 Enter-prisewide computing.

REFERENCES

Finance Marketing

EmbeddedLAN

EmbeddedLAN

EmbeddedLAN Embedded

LAN

File serverfor backbone

WAN

Backbone widearea network

CorporateBackbone

Accounting

Manufacturing

AdministrativeServicesand HRM

Mainframe

t04.qxd 01/27/2005 04:43PM Page T4.38 EQA

REFERENCES T4.39

“Just Plug It In: Networking Via Power Circuits,” PC World, April2002.

Keen, P. G. W., and J. M. Cummings, Networks in Action: BusinessChoices and Telecommunications Decisions. Belmont, CA: Wadsworth,1994.

Kim, B. G., and P. Wang, “ATM Network: Goals and Challenges,”Communications of the ACM, Vol. 38, No. 2, February 1995.

Kini, R. B., “Peer-to-Peer Technology: A Technology Reborn,” Inf-ormation Systems Management, Summer 2002.

MacKie-Mason, J. K., “What Is Circuit-Switching?” press.umich.edu/jep/works/node18.html, May 2002.

MacKie-Mason, J. K., “What Changes Are Likely in NetworkTechnology?” press.umich.edu/jep/works/node25.html, May 2002.

Metz, C., “Hot Spots Getting Hotter,” PC Magazine, August 2002.

Minoli, D., and A. Schmidt, Internet Architectures. New York: Wiley,1999.

Mueller, M., “Universal Service and the Telecommunications Act:Myth Made Law,” Communications of the ACM, Vol. 40, No. 3,March 1997.

Ploskina, B., “The New Alchemy,” Interactive Week, November 6,2000.

Poe, R., “Optical Illusion,” Business2.com, November 14, 2000.

Pringle, D., “Orange Introduces Mobile Videophone,” Asian WallStreet Journal, March 13, 2001.

Quelch, J. A., and L. R. Klein, “The Internet and InternationalMarketing,” Sloan Management Review, Spring 1996.

Rose, F., “New Modems Are Fast, Cheap,” Wall Street Journal, Feb-ruary 11, 1997.

Scanlon, B., “Going Global,” Interactive Week, December 11,2000.

Schulzrinne, H., and K. Arabshian, “Providing Emergency Ser-vices in Internet Telephony,” IEEE Internet Computing, May–June2002.

Simpson, D., “Cut Costs with Client/Server Computing? Here’sHow!” Datamation, October 1, 1995.

Steers, K., “Rev Up Your Net Connection,” PC World, March 2002.

Stewart, T. A., “Managing in a Wired Company,” Fortune, July 11,1994.

Stott, D., “Storage Area Networking (SAN),” Technology Watch, inMercurynews.com (accessed May 16, 2003.)

Sullivan, D., The New Computer User, 2nd ed. Orlando, FL: HarcourtBrace, 1997.

Vetter, R. J., “ATM Concepts, Architectures, and Protocols,”Communications of the ACM, Vol. 38, No. 2, February 1995.

Wilson, C., “Optical Networking,” White Paper, Interactive Week,February 21, 1998.

Zieger, A., “IP telephony Gets Real,” Infoworld, January 5, 1998.

t04.qxd 01/27/2005 04:43PM Page T4.39 EQA

Documents

Telecommunications - Rutgers Universitycrab.rutgers.edu/~sundares/MIS334Sec40.Sp08/protected/ref_resour… · of telecommunications and is achieved through the use of telecommunication