KAY-MAAM-NUERA.pdf

1 INTRODUCTION Since the early 1970s, technological advances around the world have occurred at a phe- nomenal rate, transforming the telecommunications industry into a highly sophisticated and extremely dynamic field. Where previously telec...

1 INTRODUCTION Since the early 1970s, technological advances around the world have occurred at a phe- nomenal rate, transforming the telecommunications industry into a highly sophisticated and extremely dynamic field. Where previously telecommunications systems had only voice to accommodate, the advent of very large-scale integration chips and the accom- panying low- cost microprocessors, computers, and peripheral equipment has dramati- cally increased the need for the exchange of digital information. This, of course, neces- sitated the development and implementation of higher-capacity and much faster means of communicating. In the data communications world, data generally are defined as information that is stored in digital form. The word data is plural; a single unit of data is a datum. Data com- munications is the process of transferring digital information (usually in binary form) be- tween two or more points. Information is defined as knowledge or intelligence. Information that has been processed, organized, and stored is called data. The fundamental purpose of a data communications circuit is to transfer digital in- formation from one place to another. Thus, data communications can be summarized as the transmission, reception, and processing of digital information. The original source infor- mation can be in analog form, such as the human voice or music, or in digital form, such as binary-coded numbers or alphanumeric codes. If the source information is in analog form, it must be converted to digital form at the source and then converted back to analog form at the destination. A network is a set of devices (sometimes called nodes or stations) interconnected by media links. Data communications networks are systems of interrelated computers and computer equipment and can be as simple as a personal computer connected to a printer or two personal computers connected together through the public telephone network. On the other hand, a data communications network can be a complex com- munications system comprised of one or more mainframe computers and hundreds, thousands, or even millions of remote terminals, personal computers, and worksta- tions. In essence, there is virtually no limit to the capacity or size of a data communi- cations network. Years ago, a single computer serviced virtually every computing need. Today, the single-computer concept has been replaced by the networking concept, where a large num- ber of separate but interconnected computers share their resources. Data communications networks and systems of networks are used to interconnect virtually all kinds of digital computing equipment, from automatic teller machines (ATMs) to bank computers; per- sonal computers to information highways, such as the Internet; and workstations to main-  frame computers. Data communications networks can also be used for airline and hotel reservation systems, mass media and news networks, and electronic mail delivery systems. The list of applications for data communications networks is virtually endless. 2 HISTORY OF DATA COMMUNICATIONS It is highly likely that data communications began long before recorded time in the form of smoke signals or tom-tom drums, although they surely did not involve electricity or an elec- tronic apparatus, and it is highly unlikely that they were binary coded. One of the earliest means of communicating electrically coded information occurred in 1753, when a proposal submitted to a Scottish magazine suggested running a communications line between vil- lages comprised of 26 parallel wires, each wire for one letter of the alphabet. A Swiss in- ventor constructed a prototype of the 26-wire system, but current wire-making technology proved the idea impractical. In 1833, Carl Friedrich Gauss developed an unusual system based on a five-by-five matrix representing 25 letters (I and J were combined). The idea was to send messages over a single wire by deflecting a needle to the right or left between one and five times. The ini- tial set of deflections indicated a row, and the second set indicated a column. Consequently, it could take as many as 10 deflections to convey a single character through the system. If we limit the scope of data communications to methods that use binary-coded electri- cal signals to transmit information, then the first successful (and practical) data communica- tions system was invented by Samuel F. B. Morse in 1832 and called the telegraph. Morse also developed the first practical data communications code, which he called the Morse code. With telegraph, dots and dashes (analogous to logic 1s and 0s) are transmitted across a wire using electromechanical induction. Various combinations of dots, dashes, and pauses represented bi- nary codes for letters, numbers, and punctuation marks. Because all codes did not contain the same number of dots and dashes, Morse’s system combined human intelligence with electron- ics, as decoding was dependent on the hearing and reasoning ability of the person receiving the message. (Sir Charles Wheatstone and Sir William Cooke allegedly invented the first telegraph in England, but their contraption required six different wires for a single telegraph line.) In 1840, Morse secured an American patent for the telegraph, and in 1844 the first tele- graph line was established between Baltimore and Washington, D.C., with the first message conveyed over this system being “What hath God wrought!” In 1849, the first slow-speed telegraph printer was invented, but it was not until 1860 that high-speed (15-bps) printers were available. In 1850, Western Union Telegraph Company was formed in Rochester, New York, for the purpose of carrying coded messages from one person to another. In 1874, Emile Baudot invented a telegraph multiplexer, which allowed signals from up to six different telegraph machines to be transmitted simultaneously over a single wire. The telephone was invented in 1875 by Alexander Graham Bell and, unfortunately, very lit- tle new evolved in telegraph until 1899, when Guglielmo Marconi succeeded in sending ra- dio (wireless) telegraph messages. Telegraph was the only means of sending information across large spans of water until 1920, when the first commercial radio stations carrying voice information were installed. It is unclear exactly when the first electrical computer was developed. Konrad Zuis, a German engineer, demonstrated a computing machine sometime in the late 1930s; how- ever, at the time, Hitler was preoccupied trying to conquer the rest of the world, so the proj- ect fizzled out. Bell Telephone Laboratories is given credit for developing the first special- purpose computer in 1940 using electromechanical relays for performing logical operations. However, J. Presper Eckert and John Mauchley at the University of Pennsylva- nia are given credit by some for beginning modern-day computing when they developed the ENIAC computer on February 14, 1946.  In 1949, the U.S. National Bureau of Standards developed the first all-electronic diode-based computer capable of executing stored programs. The U.S. Census Bureau in- stalled the machine, which is considered the first commercially produced American com- puter. In the 1950s, computers used punch cards for inputting information, printers for outputting information, and magnetic tape reels for permanently storing information. These early computers could process only one job at a time using a technique called batch processing. The first general-purpose computer was an automatic sequence-controlled calculator developed jointly by Harvard University and International Business Machines (IBM) Cor- poration. The UNIVAC computer, built in 1951 by Remington Rand Corporation, was the first mass-produced electronic computer. In the 1960s, batch-processing systems were replaced by on-line processing systems with terminals connected directly to the computer through serial or parallel communica- tions lines. The 1970s introduced microprocessor-controlled microcomputers, and by the 1980s personal computers became an essential item in the home and workplace. Since then, the number of mainframe computers, small business computers, personal computers, and computer terminals has increased exponentially, creating a situation where more and more people have the need (or at least think they have the need) to exchange digital information with each other. Consequently, the need for data communications circuits, networks, and systems has also increased exponentially. Soon after the invention of the telephone, the American Telephone and Telegraph Company (AT&T) emerged, providing both long-distance and local telephone service and data communications service throughout the United States. The vast AT&T system was referred to by some as the “Bell System” and by others as “Ma Bell.” During this time, Western Union Corporation provided telegraph service. Until 1968, the AT&T op- erating tariff allowed only equipment furnished by AT&T to be connected to AT&T lines. In 1968, a landmark Supreme Court decision, the Carterfone decision, allowed non-Bell companies to interconnect to the vast AT&T communications network. This decision started the interconnect industry, which has led to competitive data communi- cations offerings by a large number of independent companies. In 1983, as a direct re- sult of an antitrust suit filed by the federal government, AT&T agreed in a court settle- ment to divest itself of operating companies that provide basic local telephone service to the various geographic regions of the United States. Since the divestiture, the com- plexity of the public telephone system in the United States has grown even more in- volved and complicated. Recent developments in data communications networking, such as the Internet, in- tranets, and the World Wide Web (WWW), have created a virtual explosion in the data com- munications industry. A seemingly infinite number of people, from homemaker to chief ex- ecutive officer, now feel a need to communicate over a finite number of facilities. Thus, the demand for higher-capacity and higher-speed data communications systems is increasing daily with no end in sight. The Internet is a public data communications network used by millions of people all over the world to exchange business and personal information. The Internet began to evolve in 1969 at the Advanced Research Projects Agency (ARPA). ARPANET was formed in the late 1970s to connect sites around the United States. From the mid-1980s to April 30, 1995, the National Science Foundation (NSF) funded a high-speed backbone called NSFNET. Intranets are private data communications networks used by many companies to ex- change information among employees and resources. Intranets normally are used for secu- rity reasons or to satisfy specific connectivity requirements. Company intranets are gener- ally connected to the public Internet through a firewall, which converts the intranet addressing system to the public Internet addressing system and provides security function- ality by filtering incoming and outgoing traffic based on addressing and protocols.  The World Wide Web (WWW) is a server-based application that allows subscribers to access the services offered by the Web. Browsers, such as Netscape Communicator and Mi- crosoft Internet Explorer, are commonly used for accessing data over the WWW. 3 DATA COMMUNICATIONS NETWORK ARCHITECTURE, PROTOCOLS, AND STANDARDS 3-1 Data Communications Network Architecture A data communications network is any system of computers, computer terminals, or com- puter peripheral equipment used to transmit and/or receive information between two or more locations. Network architectures outline the products and services necessary for the individual components within a data communications network to operate together. In essence, network architecture is a set of equipment, transmission media, and proce- dures that ensures that a specific sequence of events occurs in a network in the proper order to produce the intended results. Network architecture must include sufficient information to allow a program or a piece of hardware to perform its intended function. The primary goal of network architecture is to give the users of the network the tools necessary for setting up the network and performing data flow control. A network architecture outlines the way in which a data communications network is arranged or structured and generally includes the concept of levels or layers of functional responsibility within the architecture. The functional respon- sibilities include electrical specifications, hardware arrangements, and software procedures. Networks and network protocols fall into three general classifications: current, legacy, and legendary. Current networks include the most modern and sophisticated net- works and protocols available. If a network or protocol becomes a legacy, no one really wants to use it, but for some reason it just will not go away. When an antiquated network or protocol finally disappears, it becomes legendary. In general terms, computer networks can be classified in two different ways: broadcast and point to point. With broadcast networks, all stations and devices on the network share a single communications channel. Data are propagated through the network in relatively short messages sometimes called frames, blocks, or packets. Many or all subscribers of the net- work receive transmitted messages, and each message contains an address that identifies specifically which subscriber (or subscribers) is intended to receive the message. When mes- sages are intended for all subscribers on the network, it is called broadcasting, and when messages are intended for a specific group of subscribers, it is called multicasting. Point-to-point networks have only two stations. Therefore, no addresses are needed. All transmissions from one station are intended for and received by the other station. With point-to-point networks, data are often transmitted in long, continuous messages, some- times requiring several hours to send. In more specific terms, point-to-point and broadcast networks can be subdivided into many categories in which one type of network is often included as a subnetwork of another. 3-2 Data Communications Protocols Computer networks communicate using protocols, which define the procedures that the sys- tems involved in the communications process will use. Numerous protocols are used today to provide networking capabilities, such as how much data can be sent, how it will be sent, how it will be addressed, and what procedure will be used to ensure that there are no undetected errors. Protocols are arrangements between people or processes. In essence, a protocol is a set of customs, rules, or regulations dealing with formality or precedence, such as diplomatic or military protocol. Each functional layer of a network is responsible for providing a spe- cific service to the data being transported through the network by providing a set of rules, called protocols, that perform a specific function (or functions) within the network. Data communications protocols are sets of rules governing the orderly exchange of data within the network or a portion of the network, whereas network architecture is a set of layers and protocols that govern the operation of the network. The list of protocols used by a system is called a protocol stack, which generally includes only one protocol per layer. Layered net- work architectures consist of two or more independent levels. Each level has a specific set of responsibilities and functions, including data transfer, flow control, data segmentation and reassembly, sequence control, error detection and correction, and notification. 3-2-1 Connection-oriented and connectionless protocols. Protocols can be gen- erally classified as either connection oriented or connectionless. With a connection-ori- ented protocol, a logical connection is established between the endpoints (e.g., a virtual cir- cuit) prior to the transmission of data. Connection-oriented protocols operate in a manner similar to making a standard telephone call where there is a sequence of actions and ac- knowledgments, such as setting up the call, establishing the connection, and then discon- necting. The actions and acknowledgments include dial tone, Touch-Tone signaling, ring- ing and ring-back signals, and busy signals. Connection-oriented protocols are designed to provide a high degree of reliability for data moving through the network. This is accomplished by using a rigid set of procedures for establishing the connection, transferring the data, acknowledging the data, and then clearing the connection. In a connection-oriented system, each packet of data is assigned a unique sequence number and an associated acknowledgement number to track the data as they travel through a network. If data are lost or damaged, the destination station requests that they be re-sent. A connection-oriented protocol is depicted in Figure 1a. Characteris- tics of connection-oriented protocols include the following: 1. A connection process called a handshake occurs between two stations before any data are actually transmitted. Connections are sometimes referred to as sessions, virtual circuits, or logical connections. 2. Most connection-oriented protocols require some means of acknowledging the data as they are being transmitted. Protocols that use acknowledgment procedures provide a high level of network reliability. 3. Connection-oriented protocols often provide some means of error control (i.e., er- ror detection and error correction). Whenever data are found to be in error, the re- ceiving station requests a retransmission. 4. When a connection is no longer needed, a specific handshake drops the connection. Connectionless protocols are protocols where data are exchanged in an unplanned fashion without prior coordination between endpoints (e.g., a datagram). Connectionless protocols do not provide the same high degree of reliability as connection-oriented proto- cols; however, connectionless protocols offer a significant advantage in transmission speed. Connectionless protocols operate in a manner similar to the U.S. Postal Service, where in- formation is formatted, placed in an envelope with source and destination addresses, and then mailed. You can only hope the letter arrives at its destination. A connectionless proto- col is depicted in Figure 1b. Characteristics of connectionless protocols are as follow: 1. Connectionless protocols send data with a source and destination address without a handshake to ensure that the destination is ready to receive the data. 2. Connectionless protocols usually do not support error control or acknowledgment procedures, making them a relatively unreliable method of data transmission. 3. Connectionless protocols are used because they are often more efficient, as the data being transmitted usually do not justify the extra overhead required by connection-oriented protocols.  NETWORK Setup request Setup response Data transmitted Data acknowledgment Connection clear request Station 1 Station 2 Clear response (a) NETWORK Data Data Data Data Data Station 1 Station 2 (b) FIGURE 1 Network protocols: (a) connection-oriented; (b) connectionless 3-2-2 Syntax and semantics. Protocols include the concepts of syntax and se- mantics. Syntax refers to the structure or format of the data within the message, which in- cludes the sequence in which the data are sent. For example, the first byte of a message might be the address of the source and the second byte the address of the destination. Se- mantics refers to the meaning of each section of data. For example, does a destination ad- dress identify only the location of the final destination, or does it also identify the route the data takes between the sending and receiving locations? 3-3 Data Communications Standards During the past several decades, the data communications industry has grown at an astro- nomical rate. Consequently, the need to provide communications between dissimilar com- puter equipment and systems has also increased. A major issue facing the data communi- cations industry today is worldwide compatibility. Major areas of interest are software and programming language, electrical and cable interface, transmission media, communica- tions signal, and format compatibility. Thus, to ensure an orderly transfer of information, it has been necessary to establish standard means of governing the physical, electrical, and procedural arrangements of a data communications system. A standard is an object or procedure considered by an authority or by general consent as a basis of comparison. Standards are authoritative principles or rules that imply a model or pattern for guidance by comparison. Data communications standards are guidelines that  have been generally accepted by the data communications industry. The guidelines outline procedures and equipment configurations that help ensure an orderly transfer of informa- tion between two or more pieces of data communications equipment or two or more data communications networks. Data communications standards are not laws, however—they are simply suggested ways of implementing procedures and accomplishing results. If everyone complies with the standards, everyone’s equipment, procedures, and processes will be compatible with everyone else’s, and there will be little difficulty communicating information through the system. Today, most companies make their products to comply with standards. There are two basic types of standards: proprietary (closed) system and open sys- tem. Proprietary standards are generally manufactured and controlled by one company. Other companies are not allowed to manufacture equipment or write software using this standard. An example of a proprietary standard is Apple Macintosh computers. Advan- tages of proprietary standards are tighter control, easier consensus, and a monopoly. Disadvantages include lack of choice for the customers, higher financial investment, overpricing, and reduced customer protection against the manufacturer going out of business. With open system standards, any company can produce compatible equipment or software; however, often a royalty must be paid to the original company. An example of an open system standard is IBM’s personal computer. Advantages of open system standards are customer choice, compatibility between venders, and competition by smaller compa- nies. Disadvantages include less product control and increased difficulty acquiring agree- ment between vendors for changes or updates. In addition, standard items are not always as compatible as we would like them to be. 4 STANDARDS ORGANIZATIONS FOR DATA COMMUNICATIONS A consortium of organizations, governments, manufacturers, and users meet on a regular ba- sis to ensure an orderly flow of information within data communications networks and sys- tems by establishing guidelines and standards. The intent is that all data communications equipment manufacturers and users comply with these standards. Standards organizations generate, control, and administer standards. Often, competing companies will form a joint committee to create a compromised standard that is acceptable to everyone. The most promi- nent organizations relied on in North America to publish standards and make recommenda- tions for the data, telecommunications, and networking industries are shown in Figure 2. 4-1 International Standards Organization (ISO) Created in 1946, the International Standards Organization (ISO) is the international or- ganization for standardization on a wide range of subjects. The ISO is a voluntary, nontreaty organization whose membership is comprised mainly of members from the standards com- mittees of various governments throughout the world. The ISO creates the sets of rules and standards for graphics and document exchange and provides models for equipment and sys- tem compatibility, quality enhancement, improved productivity, and reduced costs. The ISO is responsible for endorsing and coordinating the work of the other standards organi- zations. The member body of the ISO from the United States is the American National Stan- dards Institute (ANSI). 4-2 International Telecommunications Union— Telecommunications Sector The International Telecommunications Union—Telecommunications Sector (ITU-T), formerly the Comité Consultatif Internationale de Télégraphie et Téléphonie (CCITT), is one of four per- manent parts of the International Telecommunications Union based in Geneva, Switzerland.  ISO ITU-T IEEE ANSI EIA TIA IAB FIGURE 2 Standards organizations IETF IRTF for data and network communications Membership in the ITU-T consists of government authorities and representatives from many countries. The ITU-T is now the standards organization for the United Nations and develops the recommended sets of rules and standards for telephone and data communications. The ITU-T has developed three sets of specifications: the V series for modem interfacing and data trans- mission over telephone lines; the X series for data transmission over public digital networks, e-mail, and directory services; and the I and Q series for Integrated Services Digital Network (ISDN) and its extension Broadband ISDN (sometimes called the Information Superhighway). The ITU-T is separated into 14 study groups that prepare recommendations on the following topics: Network and service operation Tariff and accounting principles Telecommunications management network and network maintenance Protection against electromagnetic environment effects Outside plant Data networks and open system communications Characteristics of telematic systems Television and sound transmission Language and general software aspects for telecommunications systems Signaling requirements and protocols End-to-end transmission performance of networks and terminals General network aspects Transport networks, systems, and equipment Multimedia services and systems 4-3 Institute of Electrical and Electronics Engineers The Institute of Electrical and Electronics Engineers (IEEE) is an international professional organization founded in the United States and is comprised of electronics, computer, and communications engineers. The IEEE is currently the world’s largest professional society with over 200,000 members. The IEEE works closely with ANSI to develop communica- tions and information processing standards with the underlying goal of advancing theory, creativity, and product quality in any field associated with electrical engineering. 4-4 American National Standards Institute The American National Standards Institute (ANSI) is the official standards agency for the United States and is the U.S. voting representative for the ISO. However, ANSI is a completely private, nonprofit organization comprised of equipment manufacturers and users of data pro- cessing equipment and services. Although ANSI has no affiliations with the federal govern- ment of the United States, it serves as the national coordinating institution for voluntary stan- dardization in the United States. ANSI membership is comprised of people from professional societies, industry associations, governmental and regulatory bodies, and consumer groups. 4-5 Electronics Industry Association The Electronics Industries Associations (EIA) is a nonprofit U.S. trade association that es- tablishes and recommends industrial standards. EIA activities include standards develop- ment, increasing public awareness, and lobbying. The EIA is responsible for developing the RS (recommended standard) series of standards for data and telecommunications. 4-6 Telecommunications Industry Association The Telecommunications Industry Association (TIA) is the leading trade association in the communications and information technology industry. The TIA facilitates business devel- opment opportunities and a competitive marketplace through market development, trade promotion, trade shows, domestic and international advocacy, and standards development. The TIA represents manufacturers of communications and information technology prod- ucts and services providers for the global marketplace through its core competencies. The TIA also facilitates the convergence of new communications networks while working for a competitive and innovative market environment. 4-7 Internet Architecture Board In 1957, the Advanced Research Projects Agency (ARPA), the research arm of the Department of Defense, was created in response to the Soviet Union’s launching of Sputnik. The original purpose ofARPA was to accelerate the advancement of technologies that could possibly be use- ful to the U.S. military. When ARPANET was initiated in the late 1970s, ARPA formed a com- mittee to oversee it. In 1983, the name of the committee was changed to the Internet Activities Board (IAB). The meaning of the acronym was later changed to the Internet Architecture Board. Today the IAB is a technical advisory group of the Internet Society with the follow- ing responsibilities: 1. Oversees the architecture protocols and procedures used by the Internet 2. Manages the processes used to create Internet standards and serves as an appeal board for complaints of improper execution of the standardization processes 3. Is responsible for the administration of the various Internet assigned numbers 4. Acts as representative for Internet Society interests in liaison relationships with other organizations concerned with standards and other technical and organiza- tional issues relevant to the worldwide Internet 5. Acts as a source of advice and guidance to the board of trustees and officers of the Internet Society concerning technical, architectural, procedural, and policy mat- ters pertaining to the Internet and its enabling technologies 4-8 Internet Engineering Task Force The Internet Engineering Task Force (IETF) is a large international community of network designers, operators, venders, and researchers concerned with the evolution of the Internet architecture and the smooth operation of the Internet.  4-9 Internet Research Task Force The Internet Research Task Force (IRTF) promotes research of importance to the evolution of the future Internet by creating focused, long-term and small research groups working on topics related to Internet protocols, applications, architecture, and technology. 5 LAYERED NETWORK ARCHITECTURE The basic concept of layering network responsibilities is that each layer adds value to ser- vices provided by sets of lower layers. In this way, the highest level is offered the full set of services needed to run a distributed data application. There are several advantages to using a layered architecture. A layered architecture facilitates peer-to-peer communications pro- tocols where a given layer in one system can logically communicate with its corresponding layer in another system. This allows different computers to communicate at different lev- els. Figure 3 shows a layered architecture where layer N at the source logically (but not nec- essarily physically) communicates with layer N at the destination and layer N of any inter- mediate nodes. 5-1 Protocol Data Unit When technological advances occur in a layered architecture, it is easier to modify one layer’s protocol without having to modify all the other layers. Each layer is essentially in- dependent of every other layer. Therefore, many of the functions found in lower layers have been removed entirely from software tasks and replaced with hardware. The primary dis- advantage of layered architectures is the tremendous amount of overhead required. With layered architectures, communications between two corresponding layers requires a unit of data called a protocol data unit (PDU). As shown in Figure 4, a PDU can be a header added at the beginning of a message or a trailer appended to the end of a message. In a layered ar- chitecture, communications occurs between similar layers; however, data must flow through the other layers. Data flows downward through the layers in the source system and upward through the layers in the destination system. In intermediate systems, data flows up- ward first and then downward. As data passes from one layer into another, headers and trail- ers are added and removed from the PDU. The process of adding or removing PDU infor- mation is called encapsulation/decapsulation because it appears as though the PDU from the upper layer is encapsulated in the PDU from the lower layer during the downward Source Destination Network Peer-to-peer communications Layer N + 1 Layer N + 1 to Layer N + 1 Layer N + 1 Layer N to Layer N Layer N Layer N Layer N – 1 to Layer N – 1 Layer N – 1 Layer N – 1 FIGURE 3 Peer-to-peer data communications  User information Overhead Overhead User information Layer N PDU Header Trailer Layer N PDU (a) (b) FIGURE 4 Protocol data unit: (a) header; (b) trailer System A System B (PDU – data) (PDU – data) Network Layer N + 1 N + 1 Layer N + 1 N + 1 PDU Header PDU Header Encapsulation Decapsulation Layer N N Layer N N PDU Header PDU Header Encapsulation Decapsulation Layer N – 1 N–1 Layer N – 1 N–1 PDU Header PDU Header Encapsulation Decapsulation FIGURE 5 Encapsulation and decapsulation movement and decapsulated during the upward movement. Encapsulate means to place in a capsule or other protected environment, and decapsulate means to remove from a capsule or other protected environment. Figure 5 illustrates the concepts of encapsulation and de- capsulation. In a layered protocol such as the one shown in Figure 3, layer N receive services from the layer immediately below it (N — 1) and provides services to the layer directly above it (N + 1). Layer N can provide service to more than one entity in layer N + 1 by using a service access point (SAP) address to define which entity the service is intended.  Information and network information passes from one layer of a multilayered archi- tecture to another layer through a layer-to-layer interface. A layer-to-layer interface defines what information and services the lower layer must provide to the upper layer. A well-de- fined layer and layer-to-layer interface provide modularity to a network. 6 OPEN SYSTEMS INTERCONNECTION Open systems interconnection (OSI) is the name for a set of standards for communicating among computers. The primary purpose of OSI standards is to serve as a structural guide- line for exchanging information between computers, workstations, and networks. The OSI is endorsed by both the ISO and ITU-T, which have worked together to establish a set of ISO standards and ITU-T recommendations that are essentially identical. In 1983, the ISO and ITU-T (CCITT) adopted a seven-layer communications architecture reference model. Each layer consists of specific protocols for communicating. The ISO seven-layer open systems interconnection model is shown in Figure 6. This hierarchy was developed to facilitate the intercommunications of data processing equip- ment by separating network responsibilities into seven distinct layers. As with any layered architecture, overhead information is added to a PDU in the form of headers and trailers. In fact, if all seven levels of the OSI model are addressed, as little as 15% of the transmitted message is actually source information, and the rest is overhead. The result of adding head- ers to each layer is illustrated in Figure 7. ISO Layer and name Function Layer 7 User networking applications and interfacing to the network Application Layer 6 Encoding language used in transmission Presentation Layer 5 Job management tracking Session Layer 4 Data tracking as it moves through a network Transport Layer 3 Network addressing and packet transmission on the network Network Layer 2 Frame formatting for transmitting data across a physical Data link communications link Layer 1 Transmission method used to propagate bits through a network Physical FIGURE 6 OSI seven-layer protocol hierarchy  Host A Host B Applications data exchange Application Data Application A A Layer 7 H7 Data Layer 7 Applications Applications Layer 6 H6 H7 Data Layer 6 Presentation Presentation Layer 5 H5 H6 H7 Data Layer 5 Session Session Layer 4 H4 H5 H6 H7 Data Layer 4 Transport Transport Layer 3 H3 H4 H5 H6 H7 Data Layer 3 Network Network Layer 2 H2 H3 H4 H5 H6 H7 Data Layer 2 Data link Data link Layer 1 H1 H2 H3 H4 H5 H6 H7 Data Layer 1 Physical Physical Protocol headers (overhead) System A System B FIGURE 7 OSI seven-layer international protocol hierarchy. H7—applications header, H6— presentation header, H5—session header, H4—transport header, H3—network header, H2— data-link header, H1—physical header In recent years, the OSI seven-layer model has become more academic than standard, as the hierarchy does not coincide with the Internet’s four-layer protocol model. However, the basic functions of the layers are still performed, so the seven-layer model continues to serve as a reference model when describing network functions. Levels 4 to 7 address the applications aspects of the network that allow for two host computers to communicate directly. The three bottom layers are concerned with the actual mechanics of moving data (at the bit level) from one machine to another. A brief summary of the services provided by each layer is given here. 1. Physical layer. The physical layer is the lowest level of the OSI hierarchy and is responsible for the actual propagation of unstructured data bits (1s and 0s) through a transmis-  Wall jack User computer Hub Hub Optical fiber cable Patch panel Twisted-pair cable, coax, or optical fiber A B C B E F Hub User computers User computers (a) (b) FIGURE 8 OSI layer 1—physical: (a) computer-to-hub; (b) connectivity devices sion medium, which includes how bits are represented, the bit rate, and how bit synchroniza- tion is achieved. The physical layer specifies the type of transmission medium and the trans- mission mode (simplex, half duplex, or full duplex) and the physical, electrical, functional, and procedural standards for accessing data communications networks. Definitions such as con- nections, pin assignments, interface parameters, timing, maximum and minimum voltage lev- els, and circuit impedances are made at the physical level. Transmission media defined by the physical layer include metallic cable, optical fiber cable, or wireless radio-wave propagation. The physical layer for a cable connection is depicted in Figure 8a. Connectivity devices connect devices on cabled networks. An example of a connec- tivity device is a hub. A hub is a transparent device that samples the incoming bit stream and simply repeats it to the other devices connected to the hub. The hub does not examine the data to determine what the destination is; therefore, it is classified as a layer 1 compo- nent. Physical layer connectivity for a cabled network is shown in Figure 8b. The physical layer also includes the carrier system used to propagate the data signals between points in the network. Carrier systems are simply communications systems that carry data through a system using either metallic or optical fiber cables or wireless arrange- ments, such as microwave, satellites, and cellular radio systems. The carrier can use analog or digital signals that are somehow converted to a different form (encoded or modulated) by the data and then propagated through the system. 2. Data-link layer. The data-link layer is responsible for providing error-free com- munications across the physical link connecting primary and secondary stations (nodes) within a network (sometimes referred to as hop-to-hop delivery). The data-link layer pack- ages data from the physical layer into groups called blocks, frames, or packets and provides a means to activate, maintain, and deactivate the data communications link between nodes. The data-link layer provides the final framing of the information signal, provides synchro- nization, facilitates the orderly flow of data between nodes, outlines procedures for error detection and correction, and provides the physical addressing information. A block dia- gram of a network showing data transferred between two computers (A and E) at the data- link level is illustrated in Figure 9. Note that the hubs are transparent but that the switch passes the transmission on to only the hub serving the intended destination. 3. Network layer. The network layer provides details that enable data to be routed be- tween devices in an environment using multiple networks, subnetworks, or both. Network- ing components that operate at the network layer include routers and their software. The  Switch Hub Hub G Hub Hub Hub I A B E F FIGURE 9 OSI layer 2—data link Subnet Router Subnet Subnet Router Subnets Router Subnet Subnet Router Hub Hub Subnet FIGURE 10 OSI layer 3—network network layer determines which network configuration is most appropriate for the function provided by the network and addresses and routes data within networks by establishing, maintaining, and terminating connections between them. The network layer provides the upper layers of the hierarchy independence from the data transmission and switching tech- nologies used to interconnect systems. It accomplishes this by defining the mechanism in which messages are broken into smaller data packets and routed from a sending node to a receiving node within a data communications network. The network layer also typically provides the source and destination network addresses (logical addresses), subnet informa- tion, and source and destination node addresses. Figure 10 illustrates the network layer of the OSI protocol hierarchy. Note that the network is subdivided into subnetworks that are separated by routers. 4. Transport layer. The transport layer controls and ensures the end-to-end integrity of the data message propagated through the network between two devices, which provides  Data Acknowledgment Network Data Acknowledgment Computer A Computer B FIGURE 11 OSI layer 4—transport Client Service request Hub Client Service response Server Client FIGURE 12 OSI layer 5—session for the reliable, transparent transfer of data between two endpoints. Transport layer re- sponsibilities includes message routing, segmenting, error recovery, and two types of basic services to an upper-layer protocol: connectionless oriented and connectionless. The trans- port layer is the highest layer in the OSI hierarchy in terms of communications and may provide data tracking, connection flow control, sequencing of data, error checking, and ap- plication addressing and identification. Figure 11 depicts data transmission at the transport layer. 5. Session layer. The session layer is responsible for network availability (i.e., data stor- age and processor capacity). Session layer protocols provide the logical connection entities at the application layer. These applications include file transfer protocols and sending e-mail. Session responsibilities include network log-on and log-off procedures and user authentica- tion. A session is a temporary condition that exists when data are actually in the process of be- ing transferred and does not include procedures such as call establishment, setup, or discon- nect. The session layer determines the type of dialogue available (i.e., simplex, half duplex, or full duplex). Session layer characteristics include virtual connections between applications entities, synchronization of data flow for recovery purposes, creation of dialogue units and ac- tivity units, connection parameter negotiation, and partitioning services into functional groups. Figure 12 illustrates the establishment of a session on a data network. 6. Presentation layer. The presentation layer provides independence to the applica- tion processes by addressing any code or syntax conversion necessary to present the data to the network in a common communications format. The presentation layer specifies how end-user applications should format the data. This layer provides for translation between local representations of data and the representation of data that will be used for transfer be- tween end users. The results of encryption, data compression, and virtual terminals are ex- amples of the translation service.  Type Options Images JPEG, PICT, GIF Video MPEG, MIDI Data ASCII, EBCDIC Computer A Network FIGURE 13 OSI layer 6— Computer B presentation Networking Applications File transfer Email Printing PC Applications Database Word processing Spreadsheets FIGURE 14 OSI layer 7— applications The presentation layer translates between different data formats and protocols. Presentation functions include data file formatting, encoding, encryption and decryption of data messages, dialogue procedures, data compression algorithms, synchronization, interruption, and termination. The presentation layer performs code and character set translation (including ASCII and EBCDIC) and formatting information and determines the display mechanism for messages. Figure 13 shows an illustration of the presentation layer. 7. Application layer. The application layer is the highest layer in the hierarchy and is analogous to the general manager of the network by providing access to the OSI environment. The applications layer provides distributed information services and controls the sequence of activities within an application and also the sequence of events between the computer application and the user of another application. The ap- plication layer (shown in Figure 14) communicates directly with the user’s application program. User application processes require application layer service elements to access the net- working environment. There are two types of service elements: CASEs (common applica- tion service elements), which are generally useful to a variety of application processes and SASEs (specific application service elements), which generally satisfy particular needs of application processes. CASE examples include association control that establishes, main- tains, and terminates connections with a peer application entity and commitment, concur- rence, and recovery that ensure the integrity of distributed transactions. SASE examples in- volve the TCP/IP protocol stack and include FTP (file transfer protocol), SNMP (simple network management protocol), Telnet (virtual terminal protocol), and SMTP (simple mail transfer protocol). 7 DATA COMMUNICATIONS CIRCUITS The underlying purpose of a data communications circuit is to provide a transmission path between locations and to transfer digital information from one station to another us- ing electronic circuits. A station is simply an endpoint where subscribers gain access to the circuit. A station is sometimes called a node, which is the location of computers, computer terminals, workstations, and other digital computing equipment. There are al- most as many types of data communications circuits as there are types of data commu- nications equipment. Data communications circuits utilize electronic communications equipment and fa- cilities to interconnect digital computer equipment. Communications facilities are physical means of interconnecting stations within a data communications system and can include virtually any type of physical transmission media or wireless radio system in existence. Communications facilities are provided to data communications users through public tele- phone networks (PTN), public data networks (PDN), and a multitude of private data com- munications systems. Figure 15 shows a simplified block diagram of a two-station data communications circuit. The fundamental components of the circuit are source of digital information, trans- mitter, transmission medium, receiver, and destination for the digital information. Although the figure shows transmission in only one direction, bidirectional transmission is possible by providing a duplicate set of circuit components in the opposite direction. Source. The information source generates data and could be a mainframe computer, personal computer, workstation, or virtually any other piece of digital equipment. The source equipment provides a means for humans to enter data into the system. Transmitter. Source data is seldom in a form suitable to propagate through the trans- mission medium. For example, digital signals (pulses) cannot be propagated through a wireless radio system without being converted to analog first. The transmitter en- codes the source information and converts it to a different form, allowing it to be more efficiently propagated through the transmission medium. In essence, the transmitter acts as an interface between the source equipment and the transmission medium. Transmission medium. The transmission medium carries the encoded signals from the transmitter to the receiver. There are many different types of transmission media, such as free-space radio transmission (including all forms of wireless transmission, such as terrestrial microwave, satellite radio, and cellular telephone) and physical fa- cilities, such as metallic and optical fiber cables. Very often, the transmission path is comprised of several different types of transmission facilities. Receiver. The receiver converts the encoded signals received from the transmission medium back to their original form (i.e., decodes them) or whatever form is used in the destination equipment. The receiver acts as an interface between the transmission medium and the destination equipment. Destination. Like the source, the destination could be a mainframe computer, per- sonal computer, workstation, or virtually any other piece of digital equipment. Digital Digital Transmission information Transmitter Receiver information medium source destination FIGURE 15 Simplified block diagram of a two-station data communications circuit  Transmitted data Transmitted data MSB A3 A3 0 1 1 0 A2 A2 output input A1 A1 A0 Station A Station B Station A Station B TC TC TC TC TC Clock Clock Clock Clock (b) (a) FIGURE 16 Data transmission: (a) parallel; (b) serial 8 SERIAL AND PARALLEL DATA TRANSMISSION Binary information can be transmitted either in parallel or serially. Figure 16a shows how the binary code 0110 is transmitted from station A to station B in parallel. As the figure shows, each bit position (A0 to A3) has its own transmission line. Consequently, all four bits can be transmitted simultaneously during the time of a single clock pulse (TC). This type of transmission is called parallel by bit or serial by character. Figure 16b shows the same binary code transmitted serially. As the figure shows, there is a single transmission line and, thus, only one bit can be transmitted at a time. Con- sequently, it requires four clock pulses (4TC) to transmit the entire four-bit code. This type of transmission is called serial by bit. Obviously, the principal trade-off between parallel and serial data transmission is speed versus simplicity. Data transmission can be accomplished much more rapidly using parallel transmission; however, parallel transmission requires more data lines. As a general rule, parallel transmission is used for short-distance data communications and within a computer, and serial transmission is used for long-distance data communications. 9 DATA COMMUNICATIONS CIRCUIT ARRANGEMENTS Data communications circuits can be configured in a multitude of arrangements depending on the specifics of the circuit, such as how many stations are on the circuit, type of transmis- sion facility, distance between stations, and how many users are at each station. A data com- munications circuit can be described in terms of circuit configuration and transmission mode. 9-1 Circuit Configurations Data communications networks can be generally categorized as either two point or multi- point. A two-point configuration involves only two locations or stations, whereas a multipoint configuration involves three or more stations. Regardless of the configuration, each station can have one or more computers, computer terminals, or workstations. A two-point circuit involves the transfer of digital information between a mainframe computer and a personal computer, two mainframe computers, two personal computers, or two data communications networks. A multipoint network is generally used to interconnect a single mainframe com- puter (host) to many personal computers or to interconnect many personal computers. 9-2 Transmission Modes Essentially, there are four modes of transmission for data communications circuits: simplex, half duplex, full duplex, and full/full duplex.  9-2-1 Simplex. In the simplex (SX) mode, data transmission is unidirectional; in- formation can be sent in only one direction. Simplex lines are also called receive-only, transmit-only, or one-way-only lines. Commercial radio broadcasting is an example of sim- plex transmission, as information is propagated in only one direction—from the broadcast- ing station to the listener. 9-2-2 Half duplex. In the half-duplex (HDX) mode, data transmission is possible in both directions but not at the same time. Half-duplex communications lines are also called two-way-alternate or either-way lines. Citizens band (CB) radio is an example of half-duplex transmission because to send a message, the push-to-talk (PTT) switch must be depressed, which turns on the transmitter and shuts off the receiver. To receive a message, the PTT switch must be off, which shuts off the transmitter and turns on the receiver. 9-2-3 Full duplex. In the full-duplex (FDX) mode, transmissions are possible in both directions simultaneously, but they must be between the same two stations. Full-du- plex lines are also called two-way simultaneous, duplex, or both-way lines. A local tele- phone call is an example of full-duplex transmission. Although it is unlikely that both par- ties would be talking at the same time, they could if they wanted to. 9-2-4 Full/full duplex. In the full/full duplex (F/FDX) mode, transmission is pos- sible in both directions at the same time but not between the same two stations (i.e., one sta- tion is transmitting to a second station and receiving from a third station at the same time). Full/full duplex is possible only on multipoint circuits. The U.S. postal system is an exam- ple of full/full duplex transmission because a person can send a letter to one address and re- ceive a letter from another address at the same time. 10 DATA COMMUNICATIONS NETWORKS Any group of computers connected together can be called a data communications network, and the process of sharing resources between computers over a data communications net- work is called networking. In its simplest form, networking is two or more computers con- nected together through a common transmission medium for the purpose of sharing data. The concept of networking began when someone determined that there was a need to share soft- ware and data resources and that there was a better way to do it than storing data on a disk and literally running from one computer to another. By the way, this manual technique of mov- ing data on disks is sometimes referred to as sneaker net. The most important considerations of a data communications network are performance, transmission rate, reliability, and security. Applications running on modern computer networks vary greatly from company to company. A network must be designed with the intended application in mind. A general cat- egorization of networking applications is listed in Table 1. The specific application affects how well a network will perform. Each network has a finite capacity. Therefore, network designers and engineers must be aware of the type and frequency of information traffic on the network. Table 1 Networking Applications Application Examples Standard office applications E-mail, file transfers, and printing High-end office applications Video imaging, computer-aided drafting, computer-aided design, and soft- ware development Manufacturing automation Process and numerical control Mainframe connectivity Personal computers, workstations, and terminal support Multimedia applications Live interactive video  End stations Applications Local area networks Wide area networks Networks Metropolitan area networks Global area networks FIGURE 17 Basic network components There are many factors involved when designing a computer network, including the following: 1. Network goals as defined by organizational management 2. Network security 3. Network uptime requirements 4. Network response-time requirements 5. Network and resource costs The primary balancing act in computer networking is speed versus reliability. Too often, network performance is severely degraded by using error checking procedures, data en- cryption, and handshaking (acknowledgments). However, these features are often required and are incorporated into protocols. Some networking protocols are very reliable but require a significant amount of over- head to provide the desired high level of service. These protocols are examples of connection- oriented protocols. Other protocols are designed with speed as the primary parameter and, therefore, forgo some of the reliability features of the connection-oriented protocols. These quick protocols are examples of connectionless protocols. 10-1 Network Components, Functions, and Features Computer networks are like snowflakes—no two are the same. The basic components of computer networks are shown in Figure 17. All computer networks include some combi- nation of the following: end stations, applications, and a network that will support the data traffic between the end stations. A computer network designed three years ago to support the basic networking applications of the time may have a difficult time supporting recently  File request Copy of requested file User computer File server FIGURE 18 File server operation developed high-end applications, such as medical imaging and live video teleconferencing. Network designers, administrators, and managers must understand and monitor the most recent types and frequency of networked applications. Computer networks all share common devices, functions, and features, including servers, clients, transmission media, shared data, shared printers and other peripherals, hardware and software resources, network interface card (NIC), local operating system (LOS), and the network operating system (NOS). 10-1-1 Servers. Servers are computers that hold shared files, programs, and the net- work operating system. Servers provide access to network resources to all the users of the network. There are many different kinds of servers, and one server can provide several func- tions. For example, there are file servers, print servers, mail servers, communications servers, database servers, directory/security servers, fax servers, and Web servers, to name a few. Figure 18 shows the operation of a file server. A user (client) requests a file from the file server. The file server sends a copy of the file to the requesting user. File servers allow users to access and manipulate disk resources stored on other computers. An example of a file server application is when two or more users edit a shared spreadsheet file that is stored on a server. File servers have the following characteristics: 1. File servers are loaded with files, accounts, and a record of the access rights of users or groups of users on the network. 2. The server provides a shareable virtual disk to the users (clients). 3. File mapping schemes are implemented to provide the virtualness of the files (i.e., the files are made to look like they are on the user’s computer). 4. Security systems are installed and configured to provide the server with the re- quired security and protection for the files. 5. Redirector or shell software programs located on the users’ computers transpar- ently activate the client’s software on the file server. 10-1-2 Clients. Clients are computers that access and use the network and shared network resources. Client computers are basically the customers (users) of the network, as they request and receive services from the servers. 10-1-3 Transmission media. Transmission media are the facilities used to inter- connect computers in a network, such as twisted-pair wire, coaxial cable, and optical fiber cable. Transmission media are sometimes called channels, links, or lines. 10-1-4 Shared data. Shared data are data that file servers provide to clients, such as data files, printer access programs, and e-mail. 10-1-5 Shared printers and other peripherals. Shared printers and peripherals are hardware resources provided to the users of the network by servers. Resources provided include data files, printers, software, or any other items used by clients on the network.  NIC Card Computer 04 60 8C 49 F2 3B MAC (media access control) address (six bytes – 12 hex characters – 48 bits) FIGURE 19 Network interface card (NIC) 10-1-6 Network interface card. Each computer in a network has a special expan- sion card called a network interface card (NIC). The NIC prepares (formats) and sends data, receives data, and controls data flow between the computer and the network. On the trans- mit side, the NIC passes frames of data on to the physical layer, which transmits the data to the physical link. On the receive side, the NIC processes bits received from the physical layer and processes the message based on its contents. A network interface card is shown in Figure 19. Characteristics of NICs include the following: 1. The NIC constructs, transmits, receives, and processes data to and from a PC and the connected network. 2. Each device connected to a network must have a NIC installed. 3. A NIC is generally installed in a computer as a daughterboard, although some com- puter manufacturers incorporate the NIC into the motherboard during manufacturing. 4. Each NIC has a unique six-byte media access control (MAC) address, which is typically permanently burned into the NIC when it is manufactured. The MAC ad- dress is sometimes called the physical, hardware, node, Ethernet, or LAN address. 5. The NIC must be compatible with the network (i.e., Ethernet—10baseT or token ring) to operate properly. 6. NICs manufactured by different vendors vary in speed, complexity, manageabil- ity, and cost. 7. The NIC requires drivers to operate on the network. 10-1-7 Local operating system. A local operating system (LOS) allows per- sonal computers to access files, print to a local printer, and have and use one or more disk and CD drives that are located on the computer. Examples of LOSs are MS-DOS, PC-DOS, Unix, Macintosh, OS/2, Windows 3.11, Windows 95, Windows 98, Windows 2000, and Linux. Figure 20 illustrates the relationship between a personal computer and its LOS. 10-1-8 Network operating system. The network operating system (NOS) is a pro- gram that runs on computers and servers that allows the computers to communicate over a net- work. The NOS provides services to clients such as log-in features, password authentication,  UNIX MS-DOS Windows Macintosh FIGURE 20 Local operating system Personal computer (LOS) NOS Server NOS NOS Client FIGURE 21 Network operating Client system (NOS) printer access, network administration functions, and data file sharing. Some of the more pop- ular network operating systems are Unix, Novell NetWare, AppleShare, Macintosh System 7, IBM LAN Server, Compaq Open VMS, and Microsoft Windows NT Server. The NOS is soft- ware that makes communications over a network more manageable. The relationship between clients, servers, and the NOS is shown in Figure 21, and the layout of a local network operat- ing system is depicted in Figure 22. Characteristics of NOSs include the following: 1. A NOS allows users of a network to interface with the network transparently. 2. A NOS commonly offers the following services: file service, print service, mail ser- vice, communications service, database service, and directory and security services. 3. The NOS determines whether data are intended for the user’s computer or whether the data needs to be redirected out onto the network. 4. The NOS implements client software for the user, which allows them to access servers on the network. 10-2 Network Models Computer networks can be represented with two basic network models: peer-to-peer client/server and dedicated client/server. The client/server method specifies the way in which two computers can communicate with software over a network. Although clients and servers are generally shown as separate units, they are often active in a single computer but not at the same time. With the client/server concept, a computer acting as a client initiates a software re- quest from another computer acting as a server. The server computer responds and attempts  User 3 User 2 User 4 User 1 User 5 Hub NOS Database Communications server server File Mail server Print server server To other networks and servers Printer FIGURE 22 Network layout using a network operating system (NOS) Client/server 1 Hub Client/server 2 FIGURE 23 Client/server concept to satisfy the request from the client. The server computer might then act as a client and re- quest services from another computer. The client/server concept is illustrated in Figure 23. 10-2-1 Peer-to-peer client/server network. A peer-to-peer client/server network is one in which all computers share their resources, such as hard drives, printers, and so on, with all the other computers on the network. Therefore, the peer-to-peer operating sys- tem divides its time between servicing the computer on which it is loaded and servicing  Client/server 1 Client/server 2 Hub FIGURE 24 Peer-to-peer Client/server 3 Client/server 4 client/server network requests from other computers. In a peer-to-peer network (sometimes called a workgroup), there are no dedicated servers or hierarchy among the computers. Figure 24 shows a peer-to-peer client/server network with four clients/servers (users) connected together through a hub. All computers are equal, hence the name peer. Each computer in the network can function as a client and/or a server, and no sin- gle computer holds the network operating system or shared files. Also, no one com- puter is assigned network administrative tasks. The users at each computer determine which data on their computer are shared with the other computers on the network. In- dividual users are also responsible for installing and upgrading the software on their computer. Because there is no central controlling computer, a peer-to-peer network is an appro- priate choice when there are fewer than 10 users on the network, when all computers are lo- cated in the same general area, when security is not an issue, or when there is limited growth projected for the network in the immediate future. Peer-to-peer computer networks should be small for the following reasons: 1. When operating in the server role, the operating system is not optimized to effi- ciently handle multiple simultaneous requests. 2. The end user’s performance as a client would be degraded. 3. Administrative issues such as security, data backups, and data ownership may be compromised in a large peer-to-peer network. 10-2-2 Dedicated client/server network. In a dedicated client/server network, one computer is designated the server, and the rest of the computers are clients. As the network grows, additional computers can be designated servers. Generally, the designated servers function only as servers and are not used as a client or workstation. The servers store all the network’s shared files and applications programs, such as word processor documents, com- pilers, database applications, spreadsheets, and the network operating system. Client com- puters can access the servers and have shared files transferred to them over the transmission medium. Figure 25 shows a dedicated client/server-based network with three servers and three clients (users). Each client can access the resources on any of the servers and also the re- sources on other client computers. The dedicated client/server-based network is probably  Client 1 Client 2 Client 3 Hub Dedicated Dedica ted Dedicated file server print ser ver mail server FIGURE 25 Dedicated client/server Printer network the most commonly used computer networking model. There can be a separate dedicated server for each function (i.e., file server, print server, mail server, etc.) or one single general- purpose server responsible for all services. In some client/server networks, client computers submit jobs to one of the servers. The server runs the software and completes the job and then sends the results back to the client computer. In this type of client/server network, less information propagates through the network than with the file server configuration because only data and not applications programs are transferred between computers. In general, the dedicated client/server model is preferable to the peer-to-peer client/server model for general-purpose data networks. The peer-to-peer model client/server model is usu- ally preferable for special purposes, such as a small group of users sharing resources. 10-3 Network Topologies Network topology describes the layout or appearance of a network—that is, how the com- puters, cables, and other components within a data communications network are intercon- nected, both physically and logically. The physical topology describes how the network is actually laid out, and the logical topology describes how data actually flow through the network. In a data communications network, two or more stations connect to a link, and one or more links form a topology. Topology is a major consideration for capacity, cost, and reli- ability when designing a data communications network. The most basic topologies are point to point and multipoint. A point-to-point topology is used in data communications networks that transfer high-speed digital information between only two stations. Very of- ten, point-to-point data circuits involve communications between a mainframe computer and another mainframe computer or some other type of high-capacity digital device. A two- point circuit is shown in Figure 26a. A multipoint topology connects three or more stations through a single transmission medium. Examples of multipoint topologies are star, bus, ring, mesh, and hybrid.  (a) (b) Bus (c) (d) Bus Ring (e) (f) FIGURE 26 Network topologies: (a) point-to-point; (b) star; (c) bus; (d) ring; (e) mesh; (f) hybrid 139  Introduction to Data Communications and Networking 10-3-1 Star topology. A star topology is a multipoint data communications net- work where remote stations are connected by cable segments directly to a centrally located computer called a hub, which acts like a multipoint connector (see Figure 26b). In essence, a star topology is simply a multipoint circuit comprised of many two-point circuits where each remote station communicates directly with a centrally located computer. With a star topology, remote stations cannot communicate directly with one another, so they must re- lay information through the hub. Hubs also have store-and-forward capabilities, enabling them to handle more than one message at a time. 10-3-2 Bus topology. A bus topology is a multipoint data communications circuit that makes it relatively simple to control data flow between and among the computers be- cause this configuration allows all stations to receive every transmission over the network. With a bus topology, all the remote stations are physically or logically connected to a sin- gle transmission line called a bus. The bus topology is the simplest and most common method of interconnecting computers. The two ends of the transmission line never touch to form a complete loop. A bus topology is sometimes called multidrop or linear bus, and all stations share a common transmission medium. Data networks using the bus topology generally involve one centrally located host computer that controls data flow to and from the other stations. The bus topology is sometimes called a horizontal bus and is shown in Figure 26c. 10-3-3 Ring topology. A ring topology is a multipoint data communications net- work where all stations are interconnected in tandem (series) to form a closed loop or cir- cle. A ring topology is sometimes called a loop. Each station in the loop is joined by point- to-point links to two other stations (the transmitter of one and the receiver of the other) (see Figure 26d). Transmissions are unidirectional and must propagate through all the stations in the loop. Each computer acts like a repeater in that it receives signals from down-line computers then retransmits them to up-line computers. The ring topology is similar to the bus and star topologies, as it generally involves one centrally located host computer that controls data flow to and from the other stations. 10-3-4 Mesh topology. In a mesh topology, every station has a direct two-point communications link to every other station on the circuit as shown in Figure 26e. The mesh topology is sometimes called fully connected. A disadvantage of a mesh topol- ogy is a fully connected circuit requires n(n — 1)/2 physical transmission paths to in- terconnect n stations and each station must have n — 1 input/output ports. Advantages of a mesh topology are reduced traffic problems, increased reliability, and enhanced security. 10-3-5 Hybrid topology. A hybrid topology is simply combining two or more of the traditional topologies to form a larger, more complex topology. Hybrid topologies are sometimes called mixed topologies. An example of a hybrid topology is the bus star topol- ogy shown in Figure 26f. Other hybrid configurations include the star ring, bus ring, and virtually every other combination you can think of. 10-4 Network Classifications Networks are generally classified by size, which includes geographic area, distance between stations, number of computers, transmission speed (bps), transmission me- dia, and the network’s physical architecture. The four primary classifications of net- works are local area networks (LANs), metropolitan area networks (MANs), wide 140  Introduction to Data Communications and Networking Table 2 Primary Network Types Network Type Characteristics LAN (local area network) Interconnects computer users within a department, company, or group MAN (metropolitan area network) Interconnects computers in and around a large city WAN (wide area network) Interconnects computers in and around an entire country GAN (global area network) Interconnects computers from around the entire globe Building backbone Interconnects LANs within a building Campus backbone Interconnects building LANs Enterprise network Interconnects many or all of the above PAN (personal area network) Interconnects memory cards carried by people and in computers that are in close proximity to each other PAN (power line area network, Virtually no limit on how many computers it can interconnect sometimes called PLAN) and covers an area limited only by the availability of power distribution lines area networks (WANs), and global area networks (GANs). In addition, there are three primary types of interconnecting networks: building backbone, campus backbone, and enterprise network. Two promising computer networks of the future share the same acronym: the PAN (personal area network) and PAN (power line area network, sometimes called PLAN). The idea behind a personal area network is to allow people to transfer data through the human body simply by touching each other. Power line area networks use existing ac distribution networks to carry data wherever power lines go, which is virtually everywhere. When two or more networks are connected together, they constitute an internetwork or internet. An internet (lowercase i) is sometimes confused with the Internet (uppercase I). The term internet is a generic term that simply means to interconnect two or more net- works, whereas Internet is the name of a specific worldwide data communications net- work. Table 2 summarizes the characteristics of the primary types of networks, and Figure 27 illustrates the geographic relationship among computers and the different types of net- works. 10-4-1 Local area network. Local area networks (LANs) are typically privately owned data communications networks in which 10 to 40 computer users share data re- sources with one or more file servers. LANs use a network operating system to provide two-way communications at bit rates typically in the range of 10 Mbps to 100 Mbps and higher between a large variety of data communications equipment within a relatively small geographical area, such as in the same room, building, or building complex (see Figure 28). A LAN can be as simple as two personal computers and a printer or could contain dozens of computers, workstations, and peripheral devices. Most LANs link equipment that are within a few miles of each other or closer. Because the size of most LANs is limited, the longest (or worst-case) transmission time is bounded and known by everyone using the network. Therefore, LANs can utilize configurations that otherwise would not be possible. LANs were designed for sharing resources between a wide range of digital equip- ment, including personal computers, workstations, and printers. The resources shared can be software as well as hardware. Most LANs are owned by the company or organization 141  Introduction to Data Communications and Networking Local area network Single building Metropolitan area network Multiple buildings or entire city Wide area network Entire country Global area network Entire world Personal area network Between people and computers FIGURE 27 Computer network types that uses it and have a connection to a building backbone for access to other departmental LANs, MANs, WANs, and GANs. 10-4-2 Metropolitan area network. A metropolitan area network (MAN) is a high-speed network similar to a LAN except MANs are designed to encompass larger areas, usually that of an entire city (see Figure 29). Most MANs support the trans- mission of both data and voice and in some cases video. MANs typically operate at 142  Introduction to Data Communications and Networking NOS client NOS client software software LAN Laptop PC Workstation Scanner NOS server Wall jack CD-ROM/WORM Patch panel File/application/ print server Hub/repeater FAX machine Router or switch To building backbone FIGURE 28 Local area network (LAN) layout speeds of 1.5 Mbps to 10 Mbps and range from five miles to a few hundred miles in length. A MAN generally uses only one or two transmission cables and requires no switches. A MAN could be a single network, such as a cable television distribution net- work, or it could be a means of interconnecting two or more LANs into a single, larger network, enabling data resources to be shared LAN to LAN as well as from station to station or computer to computer. Large companies often use MANS to interconnect all their LANs. A MAN can be owned and operated entirely by a single, private company, or it could lease services and facilities on a monthly basis from the local cable or telephone company. Switched Multimegabit Data Services (SMDS) is an example of a service of- fered by local telephone companies for handling high-speed data communications for MANs. Other examples of MANs are FDDI (fiber distributed data interface) and ATM (asynchronous transfer mode). 10-4-3 Wide area