Chapter 7 - Review

Network+ Guide to Networks, Chapter 7 Review

 WAN's and Remote Connectivity

Now that you understand the basic transmission media, network models, and networking hardware associated with LANs (local area networks), you need to expand that knowledge to encompass WANs (wide area networks). As you have learned, a WAN is a network that connects two or more geographically distinct LANs. You might assume that WANs are the same as LANs, only bigger. Although a WAN is based on the same principles as a LAN, including reliance on the OSI model, its distance requirements affect its entire infrastructure. As a result, WANs differ from LANs in many respects. To understand the difference between a LAN and WAN; think of the hallways and stairs of your house as LAN pathways. These interior passages allow you to go from room to room. To reach destinations outside your house, however, you need to use sidewalks and streets. These public thoroughfares are analogous to WAN pathways—except that WAN pathways are not necessarily public. This chapter discusses the technical differences between LANs and WANs and describes in detail WAN transmission media and methods. It also notes the potential pitfalls in establishing and maintaining WANs. WANs are significant concerns for organizations attempting to meet the needs of telecommuting workers, global business partners, and Internet-based commerce. To pass the Network+ certification exam, you must be familiar with the variety of WAN options. 

WAN Essentials

A WAN is a network that traverses some distance and usually connects LANs, whether across the city or across the nation. You are probably familiar with at least one WAN—the Internet, which is the largest WAN in existence. However, the Internet is not a typical WAN. Many WANs arise from the simple need to connect one location to another. As an organization grows, the WAN might grow to connect more and more sites, located across a city or around the world. Only an organization’s information technology budget and aspirations limit the dimensions of its WAN. Any business or government institution with sites scattered over a wide geographical area needs a way to exchange data between those sites.

Each of the following scenarios demonstrates a need for a WAN:

·         A bank with offices around the state needs to connect those offices to gather transaction and account information into a central database. Furthermore, it needs to connect with global financial clearinghouses to, for example, conduct transactions with other institutions.

·         Regional sales representatives for a national pharmaceutical company need to submit their sales figures to a file server at the company’s headquarters and receive e-mail from the company’s mail server.  

·         An automobile manufacturer in Detroit contracts out its plastic parts manufacturing to a Delaware-based company. Through WAN links, the auto manufacturer can video conference with the plastics manufacturer, exchange specification data, and even examine the parts for quality from a remote location. 

·         A clothing manufacturer sells its products over the Internet to customers throughout the world.

Although all of these businesses need WANs, they might not need the same kinds of WANs.

Depending on the traffic load, budget, geographical breadth, and commercially available technology, each might implement a different transmission method. For every business need, a few appropriate WAN connection types might exist. 

However, many WAN technologies can coexist on the same network. WANs and LANs have several fundamental properties in common. Both are designed to enable communication between clients and hosts for resource sharing. In general, both use the same protocols from Layers 3 and higher of the OSI model. And both networks typically carry digitized data via packet-switched connections. However, LANs and WANs often differ at Layers 1 and 2 of the OSI model in access methods, topologies, and, sometimes, media. They also differ in the extent to which the organization that uses the network is responsible for the network. LANs that depend on wire-bound transmission use a building’s internal cabling, such as twisted pair, which runs from work areas to the walls, through plenum areas, and to a telecommunications closet. Such wiring is private; it belongs to the building owner. In contrast, WANs typically send data over publicly available communications networks, which are owned by local and long-distance telecommunications carriers. Such carriers, which are privately owned corporations, are also known as NSPs (network service providers). Some large NSPs based in the United States include AT&T, Verizon, and Sprint. Customers lease connections from these carriers, with payments based on the amount of bandwidth they need. For best throughput and quality, organizations lease dedicated lines, or continuously available communications channels, from a telecommunications provider, such as a local telephone company or ISP. Dedicated lines come in a variety of types that are distinguished by their capacity and transmission characteristics. The individual geographic locations connected by a WAN are known as WAN sites. A WAN link is a connection between one WAN site (or point) and another site (or point). Most WAN links are point-to-point, connecting one site to only one other site. That is, the link does not connect one site to several other sites, in the way that LAN routers connect multiple segments or workstations. Nevertheless, one location may be connected to more than one location by multiple WAN links. Figure 7-1 on page 295, illustrates the difference between WAN and LAN connectivity. The following section describes different topologies used on WANs.

WAN Topologies

WAN topologies resemble LAN topologies, but their details differ because of the distance they must cover, the larger number of users they serve, and the heavy traffic they often handle. For example, WAN topologies connect sites via dedicated and, usually, high-speed links. As a consequence, WANs use different connectivity devices. For example, to connect two buildings via high-speed T1 carrier lines, each location must use a special type of terminating device, a multiplexer, plus a router. And because WAN connections require routers or other Layer 3 devices to connect locations, their links are not capable of carrying nonroutable protocols. The following sections describe common WAN topologies and special considerations for using each.

Bus

A WAN in which each site is directly connected to no more than two other sites in a serial fashion is known as a bus topology WAN. A bus topology WAN is similar to a bus topology LAN in that each site depends on every other site in the network to transmit and receive its traffic. However, bus topology LANs use computers with shared access to one cable, whereas the WAN bus topology uses different locations, each one connected to another one through point-to-point links. A bus topology WAN is often the best option for organizations with only a few sites and the capability to use dedicated circuits. Some examples of dedicated circuits include T1, DSL, and ISDN connections, all of which are detailed later in this chapter. Dedicated circuits make it possible to transmit data regularly and reliably. Figure 7-2 on page 296, depicts a bus topology WAN using T1 and DSL connections.

Bus WAN topologies are suitable for only small WANs. Because all sites between the sending and receiving location must participate in carrying traffic, this model does not scale well. The addition of more sites can cause performance to suffer. Also, a single failure on a bus topology WAN can take down communications between all sites.

Ring

In a ring topology WAN, each site is connected to two other sites so that the entire WAN forms a ring pattern, as shown in Figure 7-3 on page 296. This architecture is similar to the simple ring topology used on a LAN, except that a WAN ring topology connects locations rather than local nodes. Also, on most modern WANs, a ring topology relies on redundant rings to carry data. Using redundant rings means that a ring topology WAN cannot be taken down by the loss of one site; instead, if one site fails, data can be rerouted around the WAN in a different direction. On the other hand, expanding ring-configured WANs can be difficult, and it is more expensive than expanding a bus topology WAN. For these reasons, WANs that use the ring topology are only practical for connecting fewer than four or five locations.

Star

The star topology WAN, mimics the arrangement of a star topology LAN. A single site acts as the central connection point for several other points, as shown in Figure 7-4 on page 297. This arrangement provides separate routes for data between any two sites. That means that if a single connection fails, only one location loses WAN access. For example, if the T3 link between the Oak Street and Main Street locations fails, the Watertown and Columbus locations can still communicate with the Main Street location because they use different routes. In a bus or ring topology, however, a single connection failure would halt all traffic between all sites. Another advantage of a star WAN is that when all of its dedicated circuits are functioning, a star WAN provides shorter data paths between any two sites. Extending a star WAN is relatively simple and less costly than extending a bus or ring topology WAN. For example, if the organization that uses the star WAN pictured in Figure 7-4 wanted to add a Maple Street, Madison, location to its topology, it could simply lease a new dedicated circuit from the Main Street office to its Maple Street office. None of the other offices would be affected by the change. If the organization were using a bus or ring WAN topology, however, two separate dedicated connections would be required to incorporate the new location into the network. As with star LAN topologies, the greatest drawback of a star WAN is that a failure at the central connection point can bring down the entire WAN. In Figure 7-4, for example, if the Main Street office suffered a catastrophic fire, the entire WAN would fail. Similarly, if he central connection point is overloaded with traffic, performance on the entire WAN will be adversely affected.

Mesh

A mesh topology WAN incorporates many directly interconnected sites. Because every site is interconnected, data can travel directly from its origin to its destination. If one connection suffers a problem, routers can redirect data easily and quickly. Mesh WANs are the most fault-tolerant type of WAN because they provide multiple routes for data to follow between any two points. For example, if the Madison office in Figure 7-5 on page 298, suffered a catastrophic fire, the Dubuque office could still send and transmit data to and from the Detroit office by going directly to the Detroit office. If both the Madison and Detroit offices failed, the Dubuque and

Indianapolis offices could still communicate. The type of mesh topology in which every WAN site is directly connected to every other site is called a full-mesh WAN. One drawback to a full-mesh WAN is the cost.

If more than a few sites are involved, connecting every site to every other requires leasing a large number of dedicated circuits. As WANs grow larger, the expense multiplies. To reduce costs, a network administrator might choose to implement a partial-mesh WAN, in which only critical WAN sites are directly interconnected and secondary sites are connected through star or ring topologies, as shown in Figure 7-5. Partial-mesh WANs are more common in today’s business world than full-mesh WANs because they are more economical.

Tiered

In a tiered topology WAN, sites connected in star or ring formations are interconnected at different levels, with the interconnection points being organized into layers to form hierarchical groupings. Figure 7-6 on page 299, depicts a tiered WAN. In this example, the Madison, Detroit, and New York offices form the upper tier, and the Dubuque, Indianapolis, Toronto, Toledo, Washington, and Boston offices form the lower tier. If the Detroit office suffers a failure, the Toronto and Toledo offices cannot communicate with any other nodes on the WAN, nor can the Washington, Boston, and New York locations exchange data with the other six locations. Yet the Washington, Boston, and New York locations can still exchange data with each other, as can the Indianapolis, Dubuque, and Madison locations. Variations on this topology abound. Indeed, flexibility makes the tiered approach quite practical. A network architect can determine the best placement of top-level routers based on traffic patterns or critical data paths. In addition, tiered systems allow for easy expansion and inclusion of redundant links to support growth. On the other hand, their enormous flexibility means that creation of tiered WANs requires careful consideration of geography, usage patterns, and growth potential. Now that you understand the fundamental shapes that WANs may take, you are ready to learn about specific technologies and types. WAN technologies discussed in the following sections differ in terms of speed, reliability, cost, distance covered, and security. Also, some are defined by specifications at the Data Link layer, whereas others are defined by specifications at the Physical layer of the OSI model. As you learn about each technology, pay attention to its characteristics and think about its possible applications. To qualify for Network+ certification, you must be familiar with the variety of WAN connection types and be able to identify the networking environments that each suits best.

PSTN

PSTN, which stands for Public Switched Telephone Network, refers to the network of lines and carrier equipment that provides telephone service to most homes and businesses. PSTN may also be called POTS (plain old telephone service). The PSTN encompasses the entire telephone system, from the wires that enter homes and businesses to the network centers that connect different regions of a country. Originally, the PSTN carried only analog traffic. All of its lines were copper wires, and switching was handled by operators who manually connected calls upon request. Today, switching is computer controlled, and nearly all of the PSTN uses digital transmission. Signals may deliver voice, video, or data traffic and travel over fiber-optic or twisted pair copper cable, microwave, and satellite connections. This chapter includes examples of PSTN-based network technologies that enable users to connect to WANs. In Chapter 12, you’ll learn about technologies that have replaced the PSTN’s original function as a voice carrier. Boundaries between the PSTN and computer networks have blurred. To appreciate how these spheres overlap, it’s helpful to understand how the PSTN provided WAN connectivity when the Internet first became popular in the 1990s. At that time, most home users logged on to the Internet via a dial-up connection. A dial-up connection is one in which a user connects her computer, via a modem, to a distant network and stays connected for a finite period of time. Unlike other types of WAN connections, dial-up connections provide a fixed period of access to the network, just as the phone call you make to a friend has a fixed length, determined by when you initiate and terminate the call. The term dial-up usually refers to a connection that uses a PSTN line. When computers connect to a public or private data network via the PSTN, modems are almost always necessary because even today, certain elements of the PSTN can’t handle digital transmission. Recall that a modem converts a computer’s digital pulses into analog signals before it issues them to the telephone line, then converts the analog signals back into digital pulses at the receiving computer’s end. Between the modems at either end, a signal travels through a carrier’s network of switches and, possibly, long-distance connections. Tracing the path of a dial-up connection is one good way to learn about the traditional PSTN. Imagine you are vacationing at a remote cabin in Alaska and the only way to connect to the Internet is via the phone line. You decide to dial into your ISP to pick up e-mail using an older laptop’s 56-Kbps modem. To do so, you first initiate a call through your computer’s dial-up software, which instructs your modem to dial the number for your ISP’s remote access server. Next, your modem attempts to establish a connection. It converts the digital signal from your laptop into an analog signal that travels over the phone line to the local telephone company’s network until it reaches the CO (central office). A CO is the place where a telephone company terminates lines and switches calls between different locations. Between your vacation cabin and the nearest CO, signals might go through one or more of the telephone company’s remote switching facilities. Modern remote switching facilities (sometimes called pedestals, because of their shape) usually contain digital equipment to convert the analog signal back to a digital signal before forwarding it to the CO. Whether at a remote switching facility, or at the CO, your signal is converted back to digital pulses. If the cabin and the ISP share the same CO, the signal is switched from your incoming connection directly to the ISP. In most cases, the ISP would have a dedicated connection to a CO. If so your signal is issued over this dedicated connection multiplexed together with many other signals. Yet it’s likely that your vacation cabin doesn’t share the same CO as your ISP. In that case, the first part of the process is the same as if you were at home—you initiate a call and connect to the local telephone company’s CO, and along the way, your signal is converted to digital pulses. However, this time your signal cannot go straight to your ISP because your ISP doesn’t have a connection in that carrier’s CO. Instead, the local telephone company that serves the cabin’s geographical area forwards the signal from its local CO to a regional CO through a dedicated connection between the two. This regional office most likely connects to a much larger regional or national WAN. The signal travels over the WAN to the regional CO closest to your ISP. That regional office directs the signal to your ISP’s local CO, or straight to the ISP’s network. Figure 7-7 on page 301, illustrates the path a signal might take in a long-distance dial-up connection. Notice that in this figure, the WAN to which regional offices connect is represented as a cloud. On networking diagrams, packet-switched networks (including the Internet) are depicted as clouds, because of the indeterminate nature of their traffic patterns. The portion of the PSTN that connects any residence or business to the nearest CO is known as the local loop, or the last mile (though it is not necessarily a mile long), and is illustrated in Figure 7-8 on page 302. It’s the part of the PSTN most likely to still use copper wire and carry analog signals. That’s because extending fiber-optic cable or high-speed wireless connections to every residence and business is costly. However, fully digital connections do exist. No matter what kind of media is used, the end of the local loop, and also the end of the carrier’s responsibility for the network, is the customer’s demarcation point, where wires terminate at an NIU (network interface unit). An example of a digital local loop is a service called FTTH (fiber to the home), which simply means that a telephone company connects residential users to its network with fiber-optic cable. A more generic term is FTTP (fiber to the premises), which refers to the use of a fiber-optic cable to connect either a residence or a business. Fiber-optic local loops may be part of a PON (passive optical network), a network in which a carrier uses fiber-optic cabling to connect with multiple endpoints—for example, several homes in a neighborhood or many businesses on a city block. The word passive applies because in a PON no repeaters or other connectivity devices intervene between a carrier and its customer. In the point-to-multipoint structure of a PON, the single endpoint at the carrier’s central office is known as the network’s OLT (optical line terminal). The OLT is a device with multiple optical ports, or PON interfaces, similar to interfaces on a router. In fact, a router can be equipped with a special OLT interface card. The OLT contains a splitter that subdivides the capacity of each port into up to 32 logical channels, one per subscriber. Physically, the PON consists of fiber-optic distribution cable leading to the vicinity of its many customers—for example, to the edge of a city block where multiple businesses will use its services or to a building of multiple residential subscribers. There, as shown in Figure 7-9 on page 302, the fiber-optic connection terminates at an ONU (optical network unit). The ONU distributes signals to multiple endpoints via fiber-optic cable, in the case of FTTP, or via copper or coax cable. PONs can handle one of several transmission technologies, including Ethernet, ATM, and SONET (discussed later in this chapter). The carrier controls which standard and how much throughput it delivers to each customer. As you can imagine, fiber-optic connections to a business or residence dramatically increase potential throughput and, therefore, the range of services available to customers. Video- and Voice-over-IP services benefit significantly from FTTP. Several WAN technologies use the public telephone network. One of the first to do so was X.25.

X.25 and Frame Relay

X.25 is an analog, packet-switched technology designed for long-distance data transmission and standardized by the ITU in the mid-1970s. The original standard for X.25 specified a maximum of 64-Kbps throughput, but by 1992 the standard was updated to include maximum throughput of 2.048 Mbps. It was originally developed as a more reliable alternative to the voice telephone system for connecting mainframe computers and remote terminals. Later, it was adopted as a method of connecting clients and servers over WANs. The X.25 standard specifies protocols at the Physical, Data Link, and Network layers of the OSI model. It provides excellent flow control and ensures data reliability over long distances by verifying the transmission at every node. Unfortunately, this verification also renders X.25 comparatively slow and unsuitable for time-sensitive applications, such as audio or video. On the other hand, X.25 benefits from being a long-established, well-known, and low-cost technology. X.25 was never widely adopted in the United States, but was accepted by other countries and was for a long time the dominant packet-switching technology used on WANs around the world.

Recall that, in packet switching, packets belonging to the same data stream may follow different, optimal paths to their destination. As a result, packet switching uses bandwidth more efficiently and allows for faster transmission than if each packet in the data stream had to follow the same path, as in circuit switching. Packet switching is also more flexible than circuit switching because packet sizes may vary.

Frame relay is a digital version of X.25 that also relies on packet switching. ITU and ANSI standardized frame relay in 1984. However, because of a lack of compatibility with other WAN technologies at the time, frame relay did not become popular in the United States and Canada until the late 1980s. Frame relay protocols operate at the Data Link layer of the OSI model and can support multiple different Network and Transport layer protocols.

The name is derived from the fact that data is separated into frames, which are then relayed from one node to another without any verification or processing. An important difference between frame relay and X.25 is that frame relay does not guarantee reliable delivery of data. X.25 checks for errors and, in the case of an error, either corrects the damaged data, or retransmits the original data. Frame relay, on the other hand, simply checks for errors. It leaves the error correction up to higher-layer protocols. Partly because it doesn’t perform the same level of error correction that X.25 performs (and, thus, has less overhead), frame relay supports higher throughput than X.25. It offers throughputs between 64 Kbps and 45 Mbps. A frame relay customer chooses the amount of bandwidth he requires and pays for only that amount. Both X.25 and frame relay rely on virtual circuits. Virtual circuits are connections between network nodes that, although based on potentially disparate physical links, logically appear to be direct, dedicated links between those nodes. One advantage to virtual circuits is their configurable use of limited bandwidth, which can make them more efficient. Several virtual circuits can be assigned to one length of cable or even to one channel on that cable. A virtual circuit uses the channel only when it needs to transmit data. Meanwhile, the channel is available for use by other virtual circuits. X.25 and frame relay may be configured as SVCs (switched virtual circuits) or PVCs (permanent virtual circuits). SVCs are connections that are established when parties need to transmit, then terminated after the transmission is complete. PVCs are connections that are established before data need to be transmitted and maintained after the transmission is complete. Note that in a PVC, the connection is established only between the two points (the sender and receiver); the connection does not specify the exact route the data will travel. Thus, in a PVC data may follow any number of paths from point A to point B. For example, a transmission traveling over a PVC from Baltimore to Phoenix might go from Baltimore to Washington, D.C., to Chicago, then to Phoenix; the next transmission over that PVC, however, might go from Baltimore to Boston to St. Louis to Denver to Phoenix. PVCs are not dedicated, individual links. When you lease an X.25 or frame relay circuit from your local carrier, your contract reflects the endpoints you specify and the amount of bandwidth you require between those endpoints. The service provider guarantees a minimum amount of bandwidth, called the CIR (committed information rate). Provisions usually account for bursts of traffic that occasionally exceed the CIR. When you lease a PVC, you share bandwidth with the other X.25 and frame relay users on the backbone. PVC links are best suited to frequent and consistent data transmission. The advantage to leasing a frame relay circuit over leasing a dedicated service is that you pay for only the amount of bandwidth required. Another advantage is that frame relay is less expensive than some other WAN technologies, depending on your location and its network availability. Also, frame relay is a long-established worldwide standard. Figure 7-10 on page 305, illustrates a WAN using frame relay. On the other hand, because frame relay and X.25 use shared lines, their throughput remains at the mercy of variable traffic patterns. In the middle of the night, data over your frame relay network may zip along at 1.544 Mbps; during midday, when everyone is surfing the Web, it may slow down to less than your CIR. In addition, frame relay circuits are not as private (and potentially not as secure) as dedicated circuits. Nevertheless, because they use the same connectivity equipment as T-carriers, they can easily be upgraded to T-carrier dedicated lines. In all but the most remote locations, frame relay connections have been replaced with newer WAN technologies such as those described in the next section.

 

ISDN

ISDN (Integrated Services Digital Network) is an international standard, originally established by the ITU in 1984, for transmitting digital data over the PSTN. In North America, a standard ISDN implementation wasn’t finalized until 1992 because telephone switch manufacturers couldn’t agree on compatible technology for supporting ISDN. The technology’s uncertain start initially made telephone companies reluctant to invest in it, and ISDN didn’t catch on as quickly as predicted. However, in the 1990s ISDN finally became a popular method of connecting WAN locations to exchange both data and voice signals. ISDN specifies protocols at the Physical, Data Link, and Transport layers of the OSI model. These protocols handle signaling, framing, connection setup and termination, routing, flow control, and error detection and correction. ISDN relies on the PSTN for its transmission medium. Connections can be either dial-up or dedicated. Dial-up ISDN is distinguished from the workstation dial-up connections discussed previously because it relies exclusively on digital transmission. In other words, it does not convert a computer’s digital signals to analog before transmitting them over the PSTN. Also, ISDN is distinguished because it can simultaneously carry as many as two voice calls and one data connection on a single line. Therefore, ISDN can eliminate the need to pay for separate phone lines to support faxes, modems, and voice calls at one location. All ISDN connections are based on two types of channels: B channels and D channels. The B channel is the “bearer” channel, employing circuit-switching techniques to carry voice, video, audio, and other types of data over the ISDN connection. A single B channel has a maximum throughput of 64 Kbps (although it is sometimes limited to 56 Kbps by the ISDN provider). The number of B channels in a single ISDN connection may vary. The D channel is the “data” channel, employing packet-switching techniques to carry information about the call, such as session initiation and termination signals, caller identity, call forwarding, and conference calling signals.A single D channel has a maximum throughput of 16 or 64 Kbps, depending on the type of ISDN connection. Each ISDN connection uses only one D channel. In North America, two types of ISDN connections are commonly used: BRI (Basic Rate Interface) and PRI (Primary Rate Interface). BRI uses two B channels and one D channel, as indicated by the notation 2B+D. The two B channels are treated as separate connections by the network and can carry voice and data or two data streams simultaneously and separate from each other. In a process called bonding, these two 64-Kbps B channels can be combined to achieve an effective throughput of 128 Kbps—the maximum amount of data traffic that a BRI connection can accommodate. Most consumers who subscribe to ISDN from home use BRI, which is the most economical type of ISDN connection. Figure 7-11 on page 307, illustrates how a typical BRI link supplies a home consumer with an ISDN link. From the telephone company’s lines, the ISDN channels connect to a Network Termination 1 device at the customer’s site. The NT1 (Network Termination 1) device connects the twisted pair wiring at the customer’s building with the ISDN terminal equipment via RJ-11 (standard telephone) or RJ-45 data jacks. The ISDN TE (terminal equipment) may include cards or stand-alone devices used to connect computers to the ISDN line (similar to a network adapter used on Ethernet networks). So that the ISDN line can connect to analog equipment, the signal must first pass through a terminal adapter. A TA (terminal adapter) converts digital signals into analog signals for use with ISDN phones and other analog devices. (Terminal adapters are sometimes called ISDN modems, though they are not, technically, modems.) Typically, telecommuters who want more throughput than their analog phone line can offer choose BRI as their ISDN connection. For a home user, the terminal adapter would most likely be an ISDN router, whereas the terminal equipment could be an Ethernet card in the user’s workstation plus, perhaps, a phone.

The BRI configuration depicted in Figure 7-11 applies to installations in North America only. Because transmission standards differ in Europe and Asia, different numbers of B channels are used in ISDN connections in those regions.

PRI (Primary Rate Interface) uses 23 B channels and one 64-Kbps D channel, as represented by the notation 23B+D. PRI is less commonly used by individual subscribers than BRI is, but it may be selected by businesses and other organizations that need more throughput. As with BRI, the separate B channels in a PRI link can carry voice and data, independently of each other or bonded together. The maximum potential throughput for a PRI connection is 1.544 Mbps. PRI and BRI connections may be interconnected on a single network. PRI links use the same kind of equipment as BRI links, but require the services of an extra network termination device, called an NT2 (Network Termination 2), to handle the multiple ISDN lines. Figure 7-12 depicts a typical PRI link as it would be installed in North America. Individual customers who need to transmit more data than a typical modem can handle or who want to use a single line for both data and voice may use ISDN lines. ISDN, although not available in every location of the United States, can be purchased from most local telephone companies. Costs vary depending on the customer’s location. PRI is more expensive than BRI. Dial-up ISDN service is less expensive than dedicated ISDN service. In some areas, ISDN providers charge customers additional usage fees based on the total length of time they remain connected. One disadvantage of ISDN is that it can span a distance of only 18,000 linear feet before repeater equipment is needed to boost the signal. For this reason, it is only feasible to use for the local loop portion of the WAN link.

T-Carriers

Another WAN transmission method that grew from a need to transmit digital data at high speeds over the PSTN is T-carrier technology, which includes T1s, fractional T1s, and T3s. T-carrier standards specify a method of signaling, which means they belong to the Physical layer of the OSI model. A T-carrier uses TDM (time division multiplexing) over two wire pairs (one for transmitting and one for receiving) to divide a single channel into multiple channels. For example, multiplexing enables a single T1 circuit to carry 24 channels, each capable of 64-Kbps throughput; thus, a T1 has a maximum capacity of 24 × 64 Kbps, or 1.544 Mbps. Each channel may carry data, voice, or video signals. The medium used for T-carrier signaling can be ordinary copper wire, fiber-optic cable, or wireless links. AT&T developed T-carrier technology in 1957 in an effort to digitize voice signals and thereby enable such signals to travel longer distances over the PSTN. Before that time, voice signals, which were purely analog, were expensive to transmit over long distances because of the number of connectivity devices needed to keep the signal intelligible. In the 1970s, many businesses installed T1s to obtain more voice throughput per line. In the 1990s, with increased data communication demands, such as Internet access and geographically dispersed offices, T1s became a popular way to connect WAN sites. The next section describes the various types of T-carriers, and then the chapter moves on to describe T-carrier connectivity.

Types of T-Carriers

A number of T-carrier varieties are available to businesses today, as shown in Table 7-1. The most common T-carrier implementations are T1 and, for higher bandwidth needs, T3. A T1 circuit can carry the equivalent of 24 voice or data channels, giving a maximum data throughput of 1.544 Mbps. A T3 circuit can carry the equivalent of 672 voice or data channels, giving a maximum data throughput of 44.736 Mbps (its throughput is typically rounded up to 45 Mbps for the purposes of discussion).

The speed of a T-carrier depends on its signal level. The signal level refers to the T-carrier’s Physical layer electrical signaling characteristics as defined by ANSI standards in the early 1980s. DS0 (digital signal, level 0) is the equivalent of one data or voice channel. All other signal levels are multiples of DS0.

You may hear signal level and carrier terms used interchangeably—for example, DS1 and T1. In fact, T1 is the implementation of the DS1 standard used in North America and most of Asia. In Europe, the standard high-speed carrier connections are E1 and E3. Like T1s and T3s, E1s and E3s use time division multiplexing. However, an E1 allows for 30 channels and offers 2.048-Mbps throughput. An E3 allows for 480 channels and offers 34.368-Mbps throughput. Using special hardware, T1s can interconnect with E1s and T3s with E3s for international communications.

Table 7-1 Carrier specifications

Signal level	Carrier	Number of T1s	Number of channels	Throughput (Mbps)
DS0	—	1/24	1	.064
DS1	T1	1	24	1.544
DS1C	T1C	2	48	3.152
DS2	T2	4	96	6.312
DS3	T3	28	672	44.736
DS4	T4	168	4032	274.176
DS5	T5	240	5760	400.352

As a networking professional, you are likely to work with T1 or T3 lines. In addition to knowing their capacity, you should be familiar with their costs and uses. T1s are commonly used by businesses to connect branch offices or to connect to a carrier, such as an ISP. Telephone companies also use T1s to connect their smaller COs. ISPs may use multiple T1s or T3s, depending on their size, to connect to their Internet carriers. Because a T3 provides 28 times more throughput than a T1, some organizations find that multiple T1s—rather than a single T3—can accommodate their throughput needs. For example, suppose a university research laboratory needs to transmit molecular images over the Internet to another university, and its peak throughput need (at any given time) is 10 Mbps. The laboratory would require seven T1s (10 Mbps divided by 1.544 Mbps equals 6.48 T1s). Leasing seven T1s would prove much less expensive for the university than leasing a single T3. The cost of T1s varies from region to region. On average, leasing a full T1 might cost approximately $500 to install, plus an additional $300 to $800 per month in access fees. The longer the distance between the provider (such as an ISP or a telephone company) and the subscriber, the higher a T1’s monthly charge. For example, a T1 between Houston and New York will cost more than a T1 between Washington, D.C., and New York. Similarly, a T1 from a suburb of New York to the city center will cost more than a T1 from the city center to a business three blocks away. For organizations that do not need as much as 1.544-Mbps throughput, a fractional T1 might be a better option. A fractional T1 lease allows organizations to use only some of the channels on a T1 line and be charged according to the number of channels they use. Thus, fractional T1 bandwidth can be leased in multiples of 64 Kbps. A fractional T1 is best suited to businesses that expect their traffic to grow and that may require a full T1 eventually, but can’t currently justify leasing a full T1. T3s are more expensive than T1s and are used by more data-intensive businesses—for example, computer consulting firms that provide online data backups and warehousing for a number of other businesses or large long-distance carriers.

A T3 might cost as much as $1000 to install, plus monthly service fees based on usage. If a customer uses the full T3 bandwidth of 45 Mbps, for example, the monthly charges might be as high as $10,000. Of course, T-carrier costs will vary depending on the service provider, your location, and the distance covered by the T3.

T-Carrier Connectivity

The approximate costs mentioned previously include monthly access and installation, but not connectivity hardware. Every T-carrier line requires connectivity hardware at both the customer site and the local telecommunications provider’s switching facility. Connectivity hardware may be purchased or leased. If your organization uses an ISP to establish and service your T-carrier line, you might lease the connectivity equipment. If you lease the line directly from the local carrier and you anticipate little change in your connectivity requirements over time, however, you might want to purchase the hardware. T-carrier lines require specialized connectivity hardware that cannot be used with other WAN transmission methods. In addition, T-carrier lines require different media, depending on their throughput. In the following sections, you will learn about the physical components of a T-carrier connection between a customer site and a local carrier.

Wiring

As mentioned earlier, the T-carrier system is based on AT&T’s original attempt to digitize existing long-distance PSTN lines. T1 technology can use UTP or STP (unshielded or shielded twisted pair) copper wiring—in other words, plain telephone wire—coaxial cable, microwave, or fiber-optic cable as its transmission media. However, because the digital signals require a clean connection (that is, one less susceptible to noise and attenuation), STP is preferable to UTP. For T1s using STP, repeaters must regenerate the signal approximately every 6000 feet. Twisted pair wiring cannot adequately carry the high throughput of multiple T1s or T3 transmissions. Thus, for multiple T1s, fiber-optic cabling is the medium of choice.

Termination

In Chapter 3, you learned how to terminate UTP cable in an RJ-45 connector, the type used for most patch cables on LANs. However, when copper cabling is used to carry T1 traffic, it terminates in an RJ-48 connector. RJ-48 and RJ-45 connectors are the same size and contain the same number of pins, but wire pairs used to carry T1 traffic are terminated differently in an RJ-48, as shown in Figure 7-13 on page 311. T1 traffic uses pins 1 and 2 for the receive pair and pins 4 and 5 for the transmit pair. As with LAN patch cables, T1 cables can be straight-through, in which case pinouts at both ends match those pictured in Figure 7-13. Straight-through cables might be used at a carrier’s facility to connect a patch panel with a T1 router interface. T1 crossover cables also exist. In a crossover cable, the transmit and receive pairs are reversed, as shown in Figure 7-14 on page 311. A T1 crossover cable could be used to connect two connectivity devices, such as CSUs/DSUs (discussed later) or WAN interface cards that act as CSU/DSUs. At the customer’s demarc (demarcation point), either inside or outside the building, RJ-48 connectors terminate in a smart jack, a type of NIU. In addition to terminating the line, a smart jack functions as a monitoring point for the connection. If the line between the carrier and customer experiences significant data errors, the smart jack will report this fact to the carrier. Technicians can also check the status of the line at the smart jack. Most include LEDs associated with transmitted and received signals. For example, a steady green light on the display indicates no connectivity problems, whereas a flickering light indicates data errors. A power light indicates whether or not the smart jack is receiving any signal.

Figure 7-15 on page 312, shows a smart jack (or network interface) designed to be used with a T1. Pin assignments on Plug B. The smart jack is not capable of interpreting data, however. For that, the T-carrier signals depend on a SU/DSU.

CSU/DSU (Channel Service Unit/Data Service Unit)

Although CSUs (channel service units) and DSUs (data service units) are actually two separate devices, they are typically combined into a single stand-alone device or an interface card called a CSU/DSU.  The CSU/DSU is the connection point for a T1 line. The CSU provides termination for the digital signal and ensures connection integrity through error correction and line monitoring. The DSU converts the T-carrier frames into frames the LAN can interpret and vice versa. It also connects T-carrier lines with terminating equipment. Finally, a DSU usually incorporates a multiplexer. (In some T-carrier installations, the multiplexer can be a separate device connected to the DSU.) For an incoming T-carrier line, the multiplexer separates its combined channels into individual signals that can be interpreted on the LAN. For an outgoing T-carrier line, the multiplexer combines multiple signals from a LAN for transport over the T-carrier. After being demultiplexed, an incoming T-carrier signal passes on to devices collectively known as terminal equipment. Examples of terminal equipment include switches, routers, or telephone exchange devices that accept only voice transmissions (such as a telephone switch).

Figure 7-17 on page 313, depicts a typical use of smart jacks and CSU/DSUs with a point-to-point T1-connected WAN. In the following sections, you will learn how routers and switches integrate with CSU/DSUs and multiplexers to connect T-carriers to a LAN.

Terminal Equipment

On a typical T1-connected data network, the terminal equipment consists of switches or routers. Usually, a router or Layer 3 or higher switch is the best option because these devices can translate between different Layer 3 protocols that might be used on the WAN and LAN. The router or switch accepts incoming signals from a CSU/DSU and, if necessary, translates Network layer protocols, then directs data to its destination exactly as it does on any LAN.

On some implementations, the CSU/DSU is not a separate device, but is integrated with the router or switch as an expansion card. Compared with a stand-alone CSU/DSU, which must connect to the terminal equipment via a cable, an integrated CSU/DSU offers faster signal processing and better network performance. In most cases, it is also a less-expensive and lower maintenance solution than using a separate CSU/DSU device. Figure 7-18 on page 313, illustrates one way a router with an integrated CSU/DSU can be used to connect a LAN with a T1 WAN link.

DSL (Digital Subscriber Line)

DSL (digital subscriber line) is a WAN connection method introduced by researchers at Bell Laboratories in the mid-1990s. It operates over the PSTN and competes directly with ISDN and T1 services, as well as broadband cable services, which are discussed later in this chapter. Like ISDN, DSL can span only limited distances without the help of repeaters and is, therefore, best suited to the local loop portion of a WAN link. Also, like ISDN and T-carriers, DSL can support multiple data and voice channels over a single line. DSL uses data modulation techniques at the Physical layer of the OSI model to achieve extraordinary throughput over regular telephone lines. To understand how DSL and voice signals can share the same line, it’s helpful to note that telephone lines carry voice signals over a very small range of frequencies, between 300 and 3300 Hz. This leaves higher, inaudible frequencies unused and available for carrying data. Also recall that in data modulation, a data signal alters the properties of a carrier signal.

Depending on its version, DSL connection may use a modulation technique based on amplitude or phase modulation. However, in DSL, modulation follows more complex patterns than the modulation you learned about earlier in this book. The details of DSL modulation techniques are beyond the scope of this book. However, you should understand that the types of modulation used by a DSL version affect its throughput and the distance its signals can travel before requiring a repeater. The following section describes the different versions of DSL.

Types of DSL

The term xDSL refers to all DSL varieties, of which at least eight currently exist. The better known

DSL varieties include ADSL (Asymmetric DSL), G.Lite (a version of ADSL), HDSL (High Bit-Rate DSL), SDSL (Symmetric or Single-Line DSL), VDSL (Very High Bit-Rate DSL), and SHDSL (Single-Line High Bit-Rate DSL)—the x in xDSL is replaced by the variety name. DSL types can be divided into two categories: asymmetrical and symmetrical. To understand the difference between these two categories, you must understand the concepts of downstream and upstream data transmission. The term downstream refers to data traveling from the carrier’s switching facility to the customer. Upstream refers to data traveling from the customer to the carrier’s switching facility. In some types of DSL, the throughput rates for downstream and upstream traffic differ.

In other words, if you were connected to the Internet via a DSL link, you could download images from the Internet more rapidly than you could upload them because the downstream throughput would be greater. A technology that offers more throughputs in one direction than in the other is considered asymmetrical. In asymmetrical communications, downstream throughput is higher than upstream throughput. Asymmetrical communication is well suited to users who receive more information from the network than they send to it—for example, people watching videoconferences or people surfing the Web. ADSL and VDSL are examples of asymmetrical DSL. Conversely, symmetrical technology provides equal capacity for data traveling both upstream and downstream. Symmetrical transmission is suited to users who both upload and download significant amounts of data—for example, a bank’s branch office that sends large volumes of account information to the central server at the bank’s headquarters and, in turn, receives large amounts of account information from the central server at the bank’s headquarters. HDSL, SDSL, and SHDSL are examples of symmetrical DSL. DSL versions also differ in the type of modulation they use. Some, such as the popular full-rate ADSL and VDSL, create multiple narrow channels in the higher frequency range to carry more data. For these versions, a splitter must be installed at the carrier and at the customer’s premises to separate the data signal from the voice signal before it reaches the terminal equipment (for example, the phone or the computer). G.Lite, a slower and less-expensive version of ADSL, eliminates the splitter but requires the use of a filter to prevent high-frequency DSL signals from reaching the telephone. Other types of DSL, such as HDSL and SDSL, cannot use the same wire pair that is used for voice signals. Instead, these types of DSL use the extra pair of wires contained in a telephone cable (that are otherwise typically unused). The types of DSL also vary in terms of their capacity and maximum line length. A VDSL line that carries as much as 52 Mbps in one direction and as much as 6.4 Mbps in the opposite direction can extend only a maximum of 1000 feet between the customer’s premises and the carrier’s switching facility. This limitation might suit businesses located close to a telephone company’s CO (for example, in the middle of a metropolitan area), but it won’t work for most individuals. The most popular form of DSL, ADSL, provides a maximum of 6.144 Mbps downstream and a maximum of 640 Kbps upstream. However, the distance between the customer and the central office affects the actual throughput a customer experiences. Close to the central office, DSL achieves its highest maximum throughput.

The farther away the customer’s premises, the lower the throughput. In the case of ADSL, a customer 9000 feet from the central office can potentially experience ADSL’s maximum potential throughput of 6.144 Mbps downstream. At 18,000 feet away, the farthest allowable distance, the customer will experience as little as 1.544-Mbps throughput. Still, this throughput and this distance (approximately 3.4 miles) renders ADSL suitable for most telecommuters. Table 7-2, compares current specifications for six DSL types. In addition to their data modulation techniques, capacity, and distance limitations, DSL types vary according to how they use the PSTN. Next, you will learn about how DSL connects to a business or residence over the PSTN.

Table 7-2 Comparison of DSL types

DSL type	Maximum upstream throughput (Mbps)	Maximum downstream throughput (Mbps)	Distance limitation (feet)
ADSL (“full rate”)	0.640	6.144	18,000
G.Lite (a type of ADSL)	0.512	1.544	25,000
HDSL or HDSL-2	1.544 or 2.048	1.544 or 2.048	18,000 or 12,000
SDSL	1.544	1.544	12,000
SHDSL	2.36 or 4.7	2.36 or 4.7	26,000 or 18,000
VDSL	1.6, 3.2, or 6.4 12.9,	25.9, or 51.8	1000–4500

Published distance limitations and throughput can vary from one service provider to another, depending on how far the provider is willing to guarantee a particular level of service. In addition, service providers may limit each user’s maximum throughput based on terms of the service agreement. For example, in 2011 AT&T capped the total amount of data transfer allowed for each of its DSL subscribers to 150 GB per month. The company instituted the new policy in response to a dramatic spike in downstream bandwidth usage due to Netflix streaming—in particular, online gaming. In fact, in 2010, Netflix accounted for nearly 30 percent of all downstream Internet traffic requested by fixed users in the United States.

DSL Connectivity

This section follows the path of an ADSL connection from a home computer, through the local loop, and to the telecommunications carrier’s switching facility. Although variations exist, this describes the most common implementation of DSL. Suppose you have an ADSL connection at home. One evening you open your Web browser and request the home page of your favorite sports team to find the last game’s score. As you know, the first step in this process is establishing a TCP connection with the team’s Web server. Your TCP request message leaves your computer’s NIC and travels over your home network to a DSL modem. A DSL modem is a device that modulates outgoing signals and demodulates incoming DSL signals. Thus, it contains receptacles to connect both to your incoming telephone line and to your computer or network connectivity device. Because you are using ADSL, the DSL modem also contains a splitter to separate incoming voice and data signals. The DSL modem may be external or internal (as an expansion card, for example) to the computer. If external, it may connect to a computer’s NIC via an RJ-45, USB, or wireless interface. If your home network contains more than one computer and you want all computers to share the DSL bandwidth, the DSL modem must connect to a device such as a switch or router, instead of just one computer.

In fact, rather than using two separate devices, you could buy a router that combines DSL modem functionalities with the ability to connect multiple computers and share DSL bandwidth. A DSL modem is shown in Figure 7-19 on page 316. When your request arrives at the DSL modem, it is modulated according to the ADSL specifications. Then, the DSL modem forwards the modulated signal to your local loop—the lines that connect your home with the rest of the PSTN. For the first stretch of the local loop, the signal continues over four-pair UTP wire. At some distance less than 18,000 feet, it is combined with other modulated signals in a telephone switch, usually at a remote switching facility. (To accept DSL signals, your telecommunications carrier must have newer digital switching equipment. In the few remaining locales where carriers have not updated their switching equipment, DSL service is not available.) Inside the carrier’s remote switching facility, a splitter separates your line’s data signal from any voice signals that are also carried on the line. Next, your request is sent to a device called a DSLAM (DSL access multiplexer), which aggregates multiple DSL subscriber lines and connects them to the carrier’s CO. Finally, your request is issued from your carrier’s network to the Internet backbone, as pictured in Figure 7-20 on page 317. The request travels over the Internet until it reaches your sports team’s Web server. Barring line problems and Internet congestion, the entire journey happens in a fraction of a second. After your team’s Web server accepts the connection request, the data follow the same path, but in reverse. Telecommunications carriers and manufacturers have positioned DSL as a competitor for T1, ISDN, and broadband cable services. The installation, hardware, and monthly access costs for DSL are slightly less than those for ISDN lines and significantly less than the cost for T1s. (At the time of this writing, ADSL costs approximately $30 per month in the United States, though prices vary by speed and location.) Generally speaking, DSL throughput rates, especially upstream, are lower than broadband cable, its main competition among residential customers.

Broadband Cable

While local and long-distance phone companies strive to make DSL the preferred method of

Internet access for consumers; cable companies are pushing their own connectivity option.

This option, called broadband cable or cable modem access, is based on the coaxial cable wiring used for TV signals. Such wiring can theoretically transmit as much as 150 Mbps downstream and as much as 10 Mbps upstream. Thus, broadband cable is an asymmetrical technology. However, actual broadband cable throughput is typically limited (or throttled) by the cable companies and further diminished by the fact that physical connections are shared. Customers might be allowed, at most, 10 Mbps downstream and 2 Mbps upstream throughput. During peak times of use, they might see data rates of 3 Mbps downstream and 1 Mbps upstream, for example. The asymmetry of broadband cable makes it a logical choice for users who want to surf the Web or download data from a network. Broadband cable connections require that the customer use a special cable modem, a device that modulates and demodulates signals for transmission and reception via cable wiring. Cable modems operate at the Physical and Data Link layer of the OSI model, and, therefore, do not manipulate higher-layer protocols, such as IP. The cable modem then connects to a customer’s PC via an RJ-45, USB, or wireless interface to a NIC. Alternately, the cable modem could connect to a connectivity device, such as a switch or router, thereby supplying bandwidth to a LAN rather than to just one computer. It’s also possible to use a device that combines cable modem functionality with a router; this single device can then provide both the broadband cable connection and the capability of sharing the bandwidth between multiple nodes. Figure 7-21 on page 318, provides an example of a cable modem. Before customers can subscribe to broadband cable, however, their local cable company must have the necessary infrastructure. Traditional cable TV networks supply the infrastructure for downstream communication (the TV programming), but not for upstream communication. To provide Internet access through its network, the cable company must have upgraded its equipment to support bidirectional, digital communications. For starters, the cable company’s network wiring must be replaced with HFC (hybrid fiber-coax), an expensive fiber-optic link that can support high frequencies. The HFC connects the cable company’s offices to a node location near the customer. Most large cable companies, such as Comcast and Charter, long ago upgraded their infrastructure to use HFC. Either fiber-optic or coaxial cable may connect the node to the customer’s business or residence via a connection known as a cable drop. All cable drops for the cable subscribers in the same neighborhood connect to the local node. These nodes then connect to the cable company’s central office, which is known as its head-end. At the head-end, the cable company can connect to the Internet through a variety of means (often via fiber-optic cable) or it can pick up digital satellite or microwave transmissions. The head-end can transmit data to as many as 1000 subscribers, in a one-to-many communication system. Figure 7-22 on page 319, illustrates the infrastructure of a cable system. Like DSL, broadband cable provides a dedicated, or continuous, connection that does not require dialing up a service provider. Unlike DSL, broadband cable requires many subscribers to share the same local line, thus raising concerns about security and actual (versus theoretical) throughput. For example, if your cable company supplied you and five of your neighbors with broadband cable services, your neighbors could, with some technical prowess, capture the data that you transmit to the Internet. (Modern cable networks provide encryption for data traveling to and from customer premises; however, these encryption schemes can be thwarted.) Moreover, the throughput of a cable line is fixed. As with any fixed resource, the more one claims, the less that is left for others. In other words the greater the number of users sharing a single line, the less throughput available to each individual user.

Cable companies counter this perceived disadvantage by rightly claiming that at some point (for example, at a remote switching facility or at the DSLAM interface), a telecommunications carrier’s DSL bandwidth is also fixed and shared among a group of customers.

In the United States, broadband cable access costs approximately $45 per month for customers who already subscribe to cable TV service. Broadband cable is less often used in businesses than DSL, primarily because most office buildings do not contain a coaxial cable infrastructure.

BPL (Broadband over Powerline)

In addition to coaxial, twisted pair, and fiber-optic cable, power lines can be used to deliver broadband Internet service. Starting around the year 2000, electric utilities began offering BPL (broadband over powerline), or high-speed Internet access, over the electrical grid. The service promised potential for connecting remote users who might not be within reach of DSL or cable services, but who were connected to the power lines, to finally receive high-speed Internet access. BPL is shared among multiple customers, which limits practical throughputs to no more than 1 Mbps. Each customer accesses the network using a modem plugged into an electrical outlet. BPL requires users to be within 2 km of a repeater. BPL didn’t take off as planned, however, and promise for the service’s widespread deployment peaked in the mid-2000s. Standards were subjected to opposition from many telecommunications groups and took a long time to develop. Necessary infrastructure upgrades, including numerous repeaters, cost more than anticipated. Also, signals transmitted via power lines are subject to much more noise than those carried by DSL or cable services. And finally, amateur radio operators who claimed its frequencies interfered with their signals protested the service. Most U.S. utility companies that invested in BPL have abandoned it. Installations do exist, however, in some European countries.

ATM (Asynchronous Transfer Mode)

So far, you have learned about several WAN transmission methods, such as ISDN, T-carriers, and DSL, which achieve high throughput by manipulating signals at the Physical layer. You also learned about some, such as X.25 and frame relay, that operate at the Data Link layer. ATM (Asynchronous Transfer Mode) is a third WAN technology that functions at the Data Link layer. Its ITU standard prescribes both network access and signal multiplexing techniques. Asynchronous refers to a communications method in which nodes do not have to conform to any predetermined schemes that specify the timing of data transmissions. In asynchronous communications, a node can transmit at any instant, and the destination node must accept the transmission as it comes. To ensure that the receiving node knows when it has received a complete frame, asynchronous communications provide start and stop bits for each character transmitted. When the receiving node recognizes a start bit, it begins to accept a new character. When it receives the stop bit for that character, it ceases to look for the end of that character’s transmission. Asynchronous data transmission, therefore, occurs in random stops and starts. ATM was first conceived by researchers at Bell Labs in the early 1980s, but it took a dozen years before standards organizations could reach an agreement on its specifications. ATM may run over fiber-optic cable or Cat 5 or better UTP or STP cable. Though it is less popular now than in the late 1990s, ATM may still be found on WANs particularly those owned by large, public telecommunications carriers. Like Ethernet, ATM specifies Data Link layer framing techniques. But what sets ATM apart from Ethernet is its fixed packet size. In ATM, a packet is called a cell and always consists of 48 bytes of data plus a 5-byte header. This fixed-sized, 53-byte packet allows ATM to provide predictable network performance. However, recall that a smaller packet size requires more overhead. In fact, ATM’s smaller packet size does decrease its potential throughput, but the efficiency of using cells compensates for that loss. Like X.25 and frame relay, ATM relies on virtual circuits. On an ATM network, switches determine the optimal path between the sender and receiver and then establish this path before the network transmits data. Because ATM packages data into cells before transmission, each of which travels separately to its destination, ATM is typically considered a packet switching technology. At the same time, the use of virtual circuit’s means that ATM provides the main advantage of circuit switching—that is, a point-to-point connection that remains reliably available to the transmission until it completes, making ATM a connection-oriented technology. Establishing a reliable connection allows ATM to guarantee a specific QoS (quality of service) for certain transmissions. ATM networks can supply four QoS levels, from a “best effort” attempt for noncritical data to a guaranteed, real-time transmission for time-sensitive data. This is important for organizations using networks for time-sensitive applications, such as video and audio transmissions. For example, a company that wants to use its physical connection between two offices located at opposite sides of a state to carry voice phone calls might choose the ATM network technology with the highest possible QoS. On the other hand, the company might assign a low QoS to routine e-mail messages exchanged between the two offices. Without QoS guarantees, cells belonging to the same message may arrive in the wrong order or too slowly to be properly interpreted by the receiving node. ATM’s developers have made certain it is compatible with other leading network technologies. Its cells can support multiple types of higher-layer protocols. In addition, the ATM networks can be integrated with Ethernet or token ring networks through the use of LANE (LAN Emulation). LANE encapsulates incoming Ethernet or token ring frames, and then converts them into ATM cells for transmission over an ATM network. ATM’s throughput potential rivals any other described in this chapter, ranging from 25 Mbps to 622 Mbps. When leasing ATM connections, you can choose to pay for only as much throughput as you think you’ll need. This also allows you to tailor your desired QoS.

Currently, ATM is relatively expensive, is rarely used on small LANs, and is almost never used to connect typical workstations to a network. Gigabit Ethernet, a cheaper technology, has replaced ATM on many networks. In addition to its lower cost, Gigabit Ethernet is a more natural upgrade for the multitude of Fast Ethernet users. It overcomes the QoS issue by simply providing a larger pipe for the greater volume of traffic using the network. Although ATM caught on among the very largest carriers in the late 1990s, most networking professionals have followed the Gigabit Ethernet standard rather than spending extra dollars on ATM infrastructure. Where ATM is still used, it’s often deployed over the popular SONET WAN technology, discussed next.

SONET (Synchronous Optical Network)

SONET (Synchronous Optical Network) is a high-bandwidth WAN signaling technique developed by Bell Communications Research in the 1980s, and later standardized by ANSI and ITU. SONET specifies framing and multiplexing techniques at the Physical layer of the OSI model. Its four key strengths are that it can integrate many other WAN technologies, it offers fast data transfer rates, it allows for simple link additions and removals, and it provides a high degree of fault tolerance. (The word synchronous as used in the name of this technology means that data being transmitted and received by nodes must conform to a timing scheme. A clock maintains time for all nodes on a network. A receiving node in synchronous communications recognizes that it should be receiving data by looking at the time on the clock.) Perhaps the most important SONET advantage is that it provides interoperability. Before SONET, telecommunications carriers that used different signaling techniques (or even the same technique but different equipment) could not be assured that their networks could communicate.Now, SONET is often used to aggregate multiple T1s, T3s, or ISDN lines. SONET is also used as the underlying technology for ATM transmission. Furthermore, because it can work directly with the different standards used in different countries, SONET has emerged as the best choice for linking WANs between North America, Europe, and Asia. Internationally, SONET is known as SDH (Synchronous Digital Hierarchy). SONET’s extraordinary fault tolerance results from its use of a double-ring topology over fiber-optic cable. In this type of layout, one ring acts as the primary route for data, transmitting in a clockwise direction. The second ring acts as a backup, transmitting data counterclockwise around the ring. If, for example, a backhoe operator severs the primary ring, SONET would automatically reroute traffic to the backup ring without any loss of service. This characteristic, known as self-healing, makes SONET very reliable. (To lower the potential for a single accident to sever both rings, the cables that make up each ring should not lay adjacent to each other.) Figure 7-23 on page 322, illustrates a SONET ring and its dual-fiber connections. A SONET ring begins and ends at the telecommunications carrier’s facility. In between, it connects an organization’s multiple WAN sites in a ring fashion. It may also connect with multiple carrier facilities for additional fault tolerance. Companies can lease an entire SONET ring from a telecommunications carrier, or they can lease part of a SONET ring—for example, a circuit that offers T1 throughput—to take advantage of SONET’s reliability. At both the carrier and the customer premises, a SONET ring terminates at a multiplexer. A multiplexer combines individual SONET signals on the transmitting end, and another multiplexer separates combined signals on the receiving end. On the transmitting end, multiplexers accept input from different network types (for example, a T1 or ISDN line) and format the data in a standard SONET frame. That means that many different devices might connect to a SONET multiplexer, including, for example, a private telephone switch, a T1 multiplexer, and an ATM data switch. On the receiving end, multiplexers translate the incoming signals back into their original format. Most SONET multiplexers allow for easy additions or removals of connections to the SONET ring, which makes this technology easily adaptable to growing and changing networks.Figure 7-24 on page 323, shows the devices necessary to connect a WAN site with a SONET ring. This is the simplest type of SONET connection; however, variations abound. The data rate of a particular SONET ring is indicated by its OC (Optical Carrier) level, a rating that is internationally recognized by networking professionals and standards organizations. OC levels in SONET are analogous to the digital signal levels of T-carriers. Table 7-3 lists the OC levels and their maximum throughput.

SONET technology is typically not implemented by small or medium-sized businesses because of its high cost. It is commonly used, for example, by large companies; long-distance companies linking metropolitan areas and countries; ISPs that want to guarantee fast, reliable access to the Internet; or telephone companies connecting their COs. SONET is particularly suited to audio, video, and imaging data transmission. As you can imagine, given its reliance on fiber-optic cable and its redundancy requirements, SONET technology is expensive to implement.

Table 7-3 SONET OC levels

OC level	Throughput (Mbps
OC1	51.84
OC3	155.52
OC12	622
OC24	1244
OC48	2488
OC96	4976
OC192	9953
OC768	39,813

WAN Technologies Compared

You have learned that WAN links offer a wide range of throughputs, from 56 Kbps for a

PSTN dial-up connection to potentially 39.8 Gbps for a full-speed SONET connection.

Table 7-4 summarizes the media and throughputs offered by each technology discussed in this chapter. Bear in mind that each technology’s transmission techniques (for example, switching for frame relay versus point-to-point for T1) will affect real throughput, so the maximum transmission speed is a theoretical limit. Actual transmission speeds will vary. In addition, this table omits wireless and satellite WAN technologies, which are discussed in the next chapter.

Table 7-4 A comparison of WAN technology throughputs

WAN technology	Typical media	Maximum throughput
Dial-up over PSTN	UTP or STP	56 Kbps theoretical; actual limit is 53 Kbps
X.25	UTP/STP (DS1 or DS3)	64 Kbps or 2.048 Mbps
Frame relay	UTP/STP (DS1 or DS3)	45 Mbps
BRI (ISDN)	UTP/STP(PSTN)	128 Kbps
PRI (ISDN)	UTP/STP (PSTN)	1.544 Mbps
T1	UTP/STP (PSTN), microwave, or fiber-optic cable	1.544 Mbps
Fractional T1	UTP/STP (PSTN), microwave, or fiber-optic cable	n times 64 Kbps (where n = number of channels leased)
T3	Microwave link or fiber-optic cable	45 Mbps
xDSL	UTP/STP (PSTN)	Theoretically, 1.544 Mbps–52 Mbps (depending on the type), but typical residential DSL throughputs are limited to 1.5 Mbps
Broadband cable	Hybrid fiber-coaxial cable	Theoretically, 56 Mbps downstream, 10 Mbps upstream, but actual throughputs are approximately 1.5–3 Mbps upstream and 256–768 Kbps downstream
BPL	Power line	Up to 1 Mbps actual throughput
ATM	Fiber-optic cable, UTP/STP (PSTN)	25 Mbps to 622 Mbps (depending on the customer’s preferred bit rate)
SONET	Fiber-optic cable	51, 155, 622, 1244, 2488, 4976, 9952, or 39813 Mbps (depending on the OC level)

Chapter Summary

■ WANs are distinguished from LANs by the fact that WANs traverse a wider geographical area. They usually employ point-to-point, dedicated communications rather than point-to-multipoint communications. They also use different connectivity devices, depending on the WAN technology in use.

■ A WAN in which each site is connected in a serial fashion to no more than two other sites is known as a bus topology WAN. This topology often provides the best solution for organizations with only a few sites and access to dedicated circuits.

■ In a ring topology WAN, each site is connected to two other sites so that the entire WAN forms a ring pattern. This architecture is similar to the LAN ring topology, except that most ring topology WANs have the capability to reverse the direction data travel to avoid a failed site.

■ In the star topology WAN, a single site acts as the central connection point for several other points. This arrangement allows one connection to fail without affecting other connections. Therefore, star topology WANs are more fault tolerant than bus or ring WANs.

■ A mesh topology WAN consists of many directly interconnected sites. In partial-mesh

WANs, only some of the WAN sites are directly interconnected. In full-mesh WANs, every site is directly connected to every other site. The full-mesh topology is the most fault tolerant and also the most expensive WAN topology to implement.

■ A tiered topology WAN is one in which sites that are connected in star or ring formations are interconnected at different levels, with the interconnection points being organized into layers to form hierarchical groupings.

■ The PSTN (Public Switched Telephone Network) is the network of lines and carrier equipment that provides telephone service to most homes and businesses. It was originally composed of analog lines alone, but now also uses digital transmission over fiber-optic or copper twisted pair cable, microwave, and satellite connections. The local loop portion of the PSTN is still primarily UTP; it is this portion that limits throughput on the PSTN. The

PSTN provides the foundation for several types of WAN connections, including dial-up networking, X.25, frame relay, T-carriers, and DSL.

■ FTTP (fiber to the premises) refers to the use of a fiber-optic cable to complete a carrier’s connection to a subscriber, whether residential or business. FTTP can be found on PONs (passive optical networks), which are point-to-multipoint networks. In PONs, a single port on a carrier’s OLT (optical line terminal) is capable of carrying 32 channels, each assigned to a customer. Fiber connects the OLT port with an ONU (optical network unit) near a group of subscribers. From the ONU, fiber-optic or copper cabling brings services to multiple customers.

■ X.25 is an analog, packet-switched technology optimized for reliable, long-distance data transmission. It can support 2-Mbps throughput. X.25 was originally developed and used for communications between mainframe computers and remote terminals.

■ Frame relay, like X.25, relies on packet switching, but carries digital signals. It does not analyze frames to check for errors, but simply relays them from node to node, so frame relay supports higher bandwidth than X.25, offering a maximum of 45-Mbps throughput.

■ Both X.25 and frame relay are configured as PVCs (permanent virtual circuits), or point-to-point connections over which data may follow different paths. When leasing an X.25 or frame relay circuit from a telecommunications carrier, a customer specifies endpoints and the amount of bandwidth required between them.

■ ISDN (Integrated Services Digital Network) is an international standard for protocols at the Physical, Data Link, and Transport layers that allows the PSTN to carry digital signals. ISDN lines may carry voice and data signals simultaneously but require an ISDN phone to carry voice traffic and an ISDN router and ISDN terminal adapter to carry data.

■ Two types of ISDN connections are commonly used by consumers in North America: BRI (Basic Rate Interface) and PRI (Primary Rate Interface). Both use a combination of bearer channels (B channels) and data channels (D channels). The B channel transmits and receives data or voice from point to point. The D channel carries information about the call, such as session initiation and termination signals, caller identity, call forwarding, and conference calling signals.

■ BRI uses two 64-Kbps circuit-switched B channels and a 16-Kbps D channel. The maximum throughput for a BRI connection is 128 Kbps. PRI uses 23 B channels and one 64-Kbps D channel. The maximum potential throughput for a PRI connection is 1.544 Mbps. Individual subscribers rarely use PRI, preferring BRI instead, but PRI may be used by businesses and other organizations that need more throughput.

■ T-carrier technology uses TDM (time division multiplexing) to divide a single channel into multiple channels for carrying voice, data, video, or other signals. Devices at the sending end arrange the data streams (multiplex), then devices at the receiving end filter them back into separate signals (demultiplex).

■ The most common T-carrier implementations are T1 and T3. A T1 circuit can carry the equivalent of 24 voice channels, giving a maximum data throughput of 1.544 Mbps. A T3 circuit can carry the equivalent of 672 voice channels, giving a maximum data throughput of 44.736 Mbps.

■ The signal level of a T-carrier refers to its Physical layer electrical signaling characteristics, as defined by ANSI standards. DS0 is the equivalent of one data or voice channel. All other signal levels are multiples of DS0.

■ T1 technology can use UTP or STP, preferably the latter. However, twisted pair wiring cannot adequately carry the high throughput of multiple T1s or T3 transmissions. For T3 transmissions, fiber-optic cable or microwave connections are necessary.

■ Incoming T-carrier lines terminate in RJ-48 connectors at a smart jack at the customer’s demarc. Next, signals are processed by a CSU/DSU. The CSU/DSU ensures connection integrity through error correction and line monitoring and converts the T-carrier frames into frames the LAN can interpret, and vice versa. It also connects T-carrier lines with terminating equipment. A CSU/DSU often includes a multiplexer.

■ DSL (digital subscriber line) is a WAN connection method that uses advanced phase or amplitude modulation in the higher (inaudible) frequencies on a phone line to achieve throughputs of up to 51.8 Mbps. DSL comes in eight different varieties, each of which either asymmetrical or symmetrical. In asymmetrical transmission, more data can be sent in one direction than in the other direction. In symmetrical transmission, throughput is equal in both directions. The most popular form of DSL is ADSL.

■ DSL technology creates a dedicated circuit. At the consumer end, a DSL modem connects computers and telephones to the DSL line. At the carrier end, a DSLAM (DSL access multiplexer) aggregates multiple incoming DSL lines before connecting them to the Internet or to larger carriers.

■ Broadband cable is a dedicated service that relies on the cable wiring used for TV signals. The service can theoretically provide as much as 36-Mbps downstream and 10-Mbps upstream throughput, though actual throughput is much lower.

■ Broadband cable connections require that the customer use a special cable modem to transmit and receive signals over coaxial cable wiring. In addition, cable companies must have replaced their coaxial cable plant with hybrid fiber-coax cable to support bidirectional, digital communications.

■ In some locations, users can access the Internet using BPL (broadband over powerline). The service is shared among multiple customers, which limits practical throughputs to no more than 1 Mbps. Each customer accesses the network using a modem plugged into an electrical outlet. BPL requires users to be within 2 km of a repeater.

■ ATM (Asynchronous Transfer Mode) is a Data Link layer standard that relies on fixed packets, called cells, consisting of 48 bytes of data plus a 5-byte header. It’s a connection-oriented technology based on virtual circuits. Having a reliable connection enables ATM to guarantee QoS (quality of service) levels for designated transmissions.

■ SONET (Synchronous Optical Network) is a high-bandwidth WAN signaling technique that specifies framing and multiplexing techniques at the Physical layer of the OSI model. Its four key strengths are that it can integrate many other WAN technologies (for example, T-carriers, ISDN, and ATM technology), it offers fast data transfer rates, it allows for simple link additions and removals, and it provides a high degree of fault tolerance. Internationally, SONET is known as SDH.

■ SONET depends on fiber-optic transmission media and uses multiplexers to connect to network devices (such as routers or telephone switches) at the customer’s end. A typical SONET network takes the form of a dual-ring topology. If one ring breaks, SONET technology automatically reroutes traffic along a backup ring. This characteristic, known as self-healing, makes SONET very reliable.

Key Terms

ADSL (Asymmetric DSL) - A variation of DSL that offers more throughput when data travels downstream, downloading from a local carrier’s switching facility to the customer, than when data travels upstream, uploading from the customer to the local carrier’s switching facility.

asymmetrical - The characteristic of a transmission technology that affords greater bandwidth in one direction (either from the customer to the carrier, or vice versa) than in the other direction.

asymmetrical DSL - See ADSL.

asynchronous - A transmission method in which data being transmitted and received by nodes do not have to conform to any timing scheme. In asynchronous communications, a node can transmit at any time and the destination node must accept the transmission as it comes.

Asynchronous Transfer Mode - See ATM.

ATM (Asynchronous Transfer Mode) - A Data Link layer technology originally conceived in the early 1980s at Bell Labs and standardized by the ITU in the mid-1990s. ATM relies on fixed packets, called cells, that each consist of 48 bytes of data plus a 5-byte header. ATM relies on virtual circuits and establishes a connection before sending data. The reliable connection ensured by ATM allows network managers to specify QoS levels for certain types of traffic.

B channel - In ISDN, the “bearer” channel, so named because it bears traffic from point to point.

Basic Rate Interface - See BRI.

bonding  - The process of combining more than one bearer channel of an ISDN line to increase throughput. For example, BRI’s two 64-Kbps B channels are bonded to create an effective throughput of 128 Kbps.

BPL (broadband over powerline) - High-speed Internet access delivered over the electrical grid.

BRI (Basic Rate Interface) - A variety of ISDN that uses two 64-Kbps bearer channels and one 16-Kbps data channel, as summarized by the notation 2B+D. BRI is the most common form of ISDN employed by home users.

broadband cable - A method of connecting to the Internet over a cable network. In broadband cable, computers are connected to a cable modem that modulates and demodulates signals to and from the cable company’s head-end.

broadband over powerline - See BPL.

bus topology WAN - A WAN in which each location is connected to no more than two other locations in a serial fashion.

cable drop - The fiber-optic or coaxial cable that connects a neighborhood cable node to a customer’s house.

cable modem - A device that modulates and demodulates signals for transmission and reception via cable wiring.

cable modem access - See broadband cable.

cell - A packet of a fixed size. In ATM technology, a cell consists of 48 bytes of data plus a 5-byte header.

central office - See CO.

channel service unit - See CSU.

CIR (committed information rate) - The guaranteed minimum amount of bandwidth selected when leasing a frame relay circuit. Frame relay costs are partially based on CIR.

CO (central office) - The location where a local or long-distance telephone service provider terminates and interconnects customer lines.

committed information rate - See CIR.

CSU (channel service unit) - A device used with T-carrier technology that provides termination for the digital signal and ensures connection integrity through error correction and line monitoring. Typically, a CSU is combined with a DSU in a single device, a CSU/DSU.

CSU/DSU - A combination of a CSU (channel service unit) and a DSU (data service unit) that serves as the connection point for a T1 line at the customer’s site. Most modern CSU/DSUs also contain a multiplexer. A CSU/DSU may be a separate device or an expansion card in another device, such as a router.

D channel - In ISDN, the “data” channel is used to carry information about the call, such as session initiation and termination signals, caller identity, call forwarding, and conference calling signals.

data service unit - See DSU.

dedicated - A continuously available link or service that is leased through another carrier.

Examples of dedicated lines include ADSL, T1, and T3.

dial-up - A type of connection in which a user connects to a distant network from a computer and stays connected for a finite period of time. Most of the time, the term dial-up refers to a connection that uses a PSTN line.

digital subscriber line - See DSL.

downstream - A term used to describe data traffic that flows from a carrier’s facility to the customer. In asymmetrical communications, downstream throughput is usually much higher than upstream throughput. In symmetrical communications, downstream and upstream throughputs are equal.

DS0 (digital signal, level 0) - The equivalent of one data or voice channel in T-carrier technology, as defined by ANSI Physical layer standards. All other signal levels are multiples of DS0.

DSL (digital subscriber line) - A dedicated WAN technology that uses advanced data modulation techniques at the Physical layer to achieve extraordinary throughput over regular phone lines. DSL comes in several different varieties, the most common of which is Asymmetric DSL (ADSL).

DSL access multiplexer - See DSLAM.

DSL modem - A device that demodulates an incoming DSL signal, extracting the information and passing it to the data equipment (such as telephones and computers) and modulates an outgoing DSL signal.

DSLAM (DSL access multiplexer) - A connectivity device located at a telecommunications carrier’s office that aggregates multiple DSL subscriber lines and connects them to a larger carrier or to the Internet backbone.

DSU (data service unit) - A device used in T-carrier technology that converts the digital signal used by bridges, routers, and multiplexers into the digital signal used on cabling. Typically, a DSU is combined with a CSU in a single device, a CSU/DSU.

E1 - A digital carrier standard used in Europe that offers 30 channels and a maximum of 2.048-Mbps throughput.

E3 - A digital carrier standard used in Europe that offers 480 channels and a maximum of 34.368-Mbps throughput.

fiber to the home - See FTTH.

fiber to the premises - See FTTP.

fractional T1 - An arrangement that allows a customer to lease only some of the channels on a T1 line.

frame relay - A digital, packet-switched WAN technology whose protocols operate at the Data Link layer. The name is derived from the fact that data is separated into frames, which are then relayed from one node to another without any verification or processing. Frame relay offers throughputs between 64 Kbps and 45 Mbps. A frame relay customer chooses the amount of bandwidth he requires and pays for only that amount.

FTTH (fiber to the home) - A service in which a residential customer is connected to his carrier’s network with fiber-optic cable.

FTTP (fiber to the premises) - A service in which a residential or business customer is connected to his carrier’s network using fiber-optic cable.

full-mesh WAN - A version of the mesh topology WAN in which every site is directly connected to every other site. Full-mesh WANs are the most fault-tolerant type of WAN.

head-end - A cable company’s central office, which connects cable wiring to many nodes before it reaches customers’ sites.

HFC (hybrid fiber-coax) - A link that consists of fiber cable connecting the cable company’s offices to a node location near the customer and coaxial cable connecting the node to the customer’s house. HFC upgrades to existing cable wiring are required before current TV cable systems can provide Internet access.

hybrid fiber-coax - See HFC.

Integrated Services Digital Network - See ISDN.

ISDN (Integrated Services Digital Network) - An international standard that uses PSTN lines to carry digital signals. It specifies protocols at the Physical, Data Link, and Transport layers of the OSI model. ISDN lines may carry voice and data signals simultaneously. Two types of ISDN connections are used in North America: BRI (Basic Rate Interface) and PRI (Primary Rate Interface). Both use a combination of bearer channels (B channels) and data channels (D channels).

LAN Emulation - See LANE.

LANE (LAN Emulation) - A method for transporting token ring or Ethernet frames over ATM networks. LANE encapsulates incoming Ethernet or token ring frames, then converts them into ATM cells for transmission over an ATM network.

last mile - See local loop.

local loop - The part of a phone system that connects a customer site with a telecommunications carrier’s switching facility.

mesh topology WAN - A type of WAN in which several sites are directly interconnected. Mesh WANs are highly fault tolerant because they provide multiple routes for data to follow between any two points.

network interface unit - See NIU.

network service provider - See NSP.

Network Termination 1 - See NT1.

Network Termination 2 - See NT2.

NIU (network interface unit) - The point at which PSTN-owned lines terminate at a customer’s premises. The NIU is usually located at the demarc.

NSP (network service provider) - A carrier that provides long-distance (and often global) connectivity between major data-switching centers across the Internet. AT&T, Verizon, and Sprint are all examples of network service providers in the United States. Customers, including ISPs, can lease dedicated private or public Internet connections from an NSP.

NT1 (Network Termination 1) - A device used on ISDN networks that connects the incoming twisted pair wiring with the customer’s ISDN terminal equipment.

NT2 (Network Termination 2) - An additional connection device required on PRI to handle the multiple ISDN lines between the customer’s network termination connection and the local phone company’s wires.

OC (Optical Carrier) - An internationally recognized rating that indicates throughput rates for SONET connections.

OLT (optical line terminal) - A device located at the carrier’s endpoint of a passive optical network. An OLT contains multiple optical ports, or PON interfaces and a splitter that subdivides the capacity of each port into up to 32 logical channels, one per subscriber.

ONU (optical network unit) - In a passive optical network, the device near the customer premises that terminates a carrier’s fiber-optic cable connection and distributes signals to multiple endpoints via fiber-optic cable, in the case of FTTP, or via copper or coax cable.

Optical Carrier - See OC.

optical line terminal - See OLT. optical network unit See ONU.

partial-mesh WAN - A version of a mesh topology WAN in which only critical sites are directly interconnected and secondary sites are connected through star or ring topologies. Partial-mesh WANs are less expensive to implement than full-mesh WANs.

passive optical network - See PON.

permanent virtual circuit - See PVC.

plain old telephone service (POTS) - See PSTN.

PON (passive optical network) - A network in which a carrier uses fiber-optic cabling to connect with multiple endpoints—for example, many businesses on a city block. The word passive applies because in a PON no repeaters or other connectivity devices intervene between a carrier and its customer.

POTS - See PSTN.

PRI (Primary Rate Interface) - A type of ISDN that uses 23 bearer channels and one 64-Kbps data channel, represented by the notation 23B+D. PRI is less commonly used by individual subscribers than BRI, but it may be used by businesses and other organizations needing more throughput.

Primary Rate Interface - See PRI.

PSTN (Public Switched Telephone Network) - The network of lines and carrier equipment that provides telephone service to most homes and businesses. Now, except for the local loop, nearly all of the PSTN uses digital transmission. Its traffic is carried by fiber-optic or copper twisted pair cable, microwave, and satellite connections.

Public Switched Telephone Network -  See PSTN.

PVC (permanent virtual circuit) - A point-to-point connection over which data may follow any number of different paths, as opposed to a dedicated line that follows a predefined path. X.25, frame relay, and some forms of ATM use PVCs.

registered jack 48 - See RJ-48.

ring topology WAN - A type of WAN in which each site is connected to two other sites so that the entire WAN forms a ring pattern.

RJ-48 (registered jack 48) - A standard for terminating wires in an eight-pin connector. RJ-48 is the preferred connector type for T1 connections that rely on twisted pair wiring.

SDH (Synchronous Digital Hierarchy) - The international equivalent of SONET.

self-healing - A characteristic of dual-ring topologies that allows them to automatically reroute traffic along the backup ring if the primary ring becomes severed.

signal level - An ANSI standard for T-carrier technology that refers to its Physical layer electrical signaling characteristics. DS0 is the equivalent of one data or voice channel. All other signal levels are multiples of DS0.

smart jack - A termination for T-carrier wire pairs that is located at the customer demark and which functions as a connection protection and monitoring point.

SONET (Synchronous Optical Network) - A high-bandwidth WAN signaling technique that specifies framing and multiplexing techniques at the Physical layer of the OSI model. It can integrate many other WAN technologies (for example, T-carriers, ISDN, and ATM technology) and allows for simple link additions and removals. SONET’s topology includes a double ring of fiber-optic cable, which results in very high fault tolerance.

star topology WAN - A type of WAN in which a single site acts as the central connection point for several other points. This arrangement provides separate routes for data between any two sites; however, if the central connection point fails, the entire WAN fails.

SVC (switched virtual circuit) - A logical, point-to-point connection that relies on switches to determine the optimal path between sender and receiver. ATM technology uses SVCs.

switched virtual circuit - See SVC.

symmetrical - A characteristic of transmission technology that provides equal throughput for data traveling both upstream and downstream and is suited to users who both upload and download significant amounts of data.

symmetrical DSL - A variation of DSL that provides equal throughput both upstream and downstream between the customer and the carrier.

synchronous - A transmission method in which data being transmitted and received by nodes must conform to a timing scheme.

Synchronous Digital Hierarchy - See SDH.

Synchronous Optical Network - See SONET.

T1- A digital carrier standard used in North America and most of Asia that provides 1.544-Mbps throughput and 24 channels for voice, data, video, or audio signals. T1s rely on time division multiplexing and may use shielded or unshielded twisted pair, coaxial cable, fiber optics, or microwave links.

T3 - A digital carrier standard used in North America and most of Asia that can carry the equivalent of 672 channels for voice, data, video, or audio, with a maximum data throughput of 44.736 Mbps (typically rounded up to 45 Mbps for purposes of discussion). T3s rely on time division multiplexing and require either fiber-optic or microwave transmission media.

T-carrier - The term for any kind of leased line that follows the standards for T1s, fractional T1s, T1Cs, T2s, T3s, or T4s.

TA (terminal adapter) - A device used to convert digital signals into analog signals for use with ISDN phones and other analog devices. TAs are sometimes called ISDN modems.

TE (terminal equipment) - The end nodes (such as computers and printers) served by the same connection (such as an ISDN, DSL, or T1 link).

terminal adapter - See TA.

terminal equipment - See TE.

tiered topology WAN - A type of WAN in which sites that are connected in star or ring formations are interconnected at different levels, with the interconnection points being organized into layers to form hierarchical groupings.

upstream - A term used to describe data traffic that flows from a customer’s site to a carrier’s facility. In asymmetrical communications, upstream throughput is usually much lower than downstream throughput. In symmetrical communications, upstream and downstream throughputs are equal.

virtual circuit - A connection between network nodes that, although based on potentially disparate physical links, logically appears to be a direct, dedicated link between those nodes.

WAN link - A point-to-point connection between two nodes on a WAN.

X.25 - An analog, packet-switched WAN technology optimized for reliable, long-distance data transmission and standardized by the ITU in the mid-1970s. The X.25 standard specifies protocols at the Physical, Data Link, and Network layers of the OSI model. It provides excellent flow control and ensures data reliability over long distances by verifying the transmission at every node. X.25 can support a maximum of only 2-Mbps throughput.

xDSL - The term used to refer to all varieties of DSL.

Review Questions

1. Which of the following elements of the PSTN is most likely capable of transmitting only analog signals?

a. Central office

b. Local loop

c. CSU/DSU

d. Remote switching facility

2. Which of the following WAN topologies comes with the highest availability and the greatest cost?

a. Bus

b. Tiered

c. Partial mesh

d. Full mesh

3. A customer calls your ISP's technical support line, complaining that his connection to the Internet usually goes as fast as 128 Kbps, but today it is only reaching 64 Kbps. He adds that he has tried dialing up three different times with the same result. What type of connection does this customer have?

a. PSTN dial-up

b. DSL

c. ISDN

d. T1

            4. What is the purpose of ISDN's D channel?

            a. To carry call session information

            b. To carry error checking information

c. To enable time division multiplexing

d. To carry data

5. Suppose you work for a bank and are leasing a frame relay connection to link an automatic teller machine located in a rural grocery store with your bank's headquarters. Which of the following circuits would be the best option, given the type of use this automatic teller machine will experience?

a. DLC

b. PVC

c. SVC

d. HLC

6. On an ISDN connection, what device separates the voice signal from the data signals at the customer premises?

a. Network termination

b. Terminal equipment

c. Multiplexer

d. Terminal adapter

7. Which of the following WAN technologies operates at Layer 3 of the OSI model?

a. ATM

b. DSL

c. SONET

d. None of the above

8. You work for a regional common carrier and have been asked to design a passive optical network that will bring high speed Internet access to a new neighborhood in a nearby suburb. The area served by the network will bring fiber-optic cable to as many as 45 households. At least how many ports on the carrier’s OLT should you expect to configure?

a. 1

b. 2

c. 3

d. 4

9. You offer your networking expertise to a small nonprofit organization. The three employees each have a desktop computer and want to share the single broadband cable Internet connection that comes to their office. You donate a router you have sitting around to make this sharing possible. Where on your network should you install the router?

a. Attached to one of the end workstations

b. Between the cable modem and the cable drop

c. Attached to a server that's connected to the cable drop

d. Between the cable modem and the workstations

10. How does ATM differ from every other WAN technology described in this chapter?

a. It does not use packet switching.

b. It requires fiber-optic media.

c. It does not provide error detection or correction.

d. It uses fixed-sized cells to carry data.

11. You work for an Internet service provider that wants to lease a T3 over a SONET ring. What is the minimum Optical Carrier level that the SONET ring must have to support the bandwidth of a T3?

a. OC1

b. OC3

c. OC12

d. OC24

12. Which two of the following are asymmetrical versions of DSL?

a. SDSL

b. ADSL

c. HDSL

d. VDSL

e. FDSL

13. What technique does T1 technology use to transmit multiple signals over a single telephone line?

a. Amplitude modulation

b. Wave division multiplexing

c. Time division multiplexing

d. Frequency modulation

14. Where on the PSTN would you most likely find a DSLAM?

a. In a remote switching facility

b. At the demarc

c. In a border router

d. In a CSU/DSU

15. The science museum where you work determines that it needs an Internet connection capable of transmitting and receiving data at 12 Mbps at any time. Which of the following T-carrier solutions would you advise?

a. A T1

b. A T3

c. Ten T1s

d. Ten T3s

16. You're troubleshooting a problem with a T1 connection between your business and the service provider’s facility. The T1 connection intermittently goes down. When you call the service provider for assistance, they say that they will only engage one of their service technicians after you have verified that all of your customer premises equipment is in working order. Given that requirement, which of the following do you not examine for faults?

a. The CSU/DSU interface card where the T1 terminates.

b. The router that contains the CSU/DSU interface card.

c. The RJ-48 connectors in your smart jack.

d. The cable that enters the smart jack from outside your building.

17. Your company has decided to order ADSL from its local telecommunications carrier. You call the carrier and find out that your office is located 17,000 feet from the nearest CO. Given ADSL's potential throughput and your distance from the CO, what is the maximum downstream throughput you can realistically expect to achieve through this connection?

a. 16 Mbps

b. 8 Mbps

c. 2 Mbps (Note: This applies to “Full Rate” ADSL, whose maximum potential downstream throughput is 8 Mbps.)

d. 200 Kbps

18. What part of a SONET network allows it to be self-healing?

a. Its double-ring topology

b. Its use of error correction protocols

c. Its use of fiber-optic cable

d. Its independence from local carriers' switching facilities

19. Which of the following may limit a DSL connection's capacity?

a. The number of nodes connected to the incoming DSL line

b. The distance from the carrier's switching facility to the ISP

c. The existence of more than one copper wire phone line at the customer's location

d. The distance from the customer to the carrier's switching facility

20. You have just started working for a regional network service provider. The company provides several T1s to businesses around your city. In addition, your company supplies Internet service to three ISPs, using multiple T1s. Two T3s connect your employer’s data center with two even larger service providers. You notice a router in the data center that handles the multiple T1s and T3s. Which of the following routing protocols is this router almost certainly running to exchange route information with routers on the other service providers’ networks?

a. OSPF

b. RIP

c. BGP

d. EIGRP

Sample Quiz

1. In asynchronous communications, a node can transmit ________ .

a. at any instant         

b. at prescribed intervals determined by the sender            

c. only when the receiver signals it is ready

d. at prescribed intervals determined by the receiver         

2. Which mesh topology directly connects every WAN site to every other site?

a. Full-tiered  

b. Full-mesh   

c. Integrated  

d. Distributed

3. Typically, SONET technology is implemented by small- and medium-sized businesses because of its low cost.

a. True

b. False           

4. Boundaries between the PSTN and computer networks have blurred over time.

a. True

b. False           

5. ISDN is a North American standard for transmitting digital data over the PSTN, but has not yet caught on globally.

a. True

b. False

6. A T-carrier uses ________ over two wire pairs to divide a single channel into multiple channels.

a. self-healing technology     

b. frequency division multiplexing    

c. bonding      

d. time division multiplexing

7. Broadband cable is a(n) ________ technology.

a. passive       

b. symmetrical           

c. asymmetrical         

d. active         

8. A T1 circuit has a maximum data throughput of ________ Mbps.

a. 0.1544        

b. 154.4          

c. 15.44          

d. 1.544

9. Which WAN topology provides the most fault tolerance?

a. Star topology WAN            

b. Ring topology WAN            

c. Mesh topology WAN          

d. Bus topology WAN

           

10. ATM technology functions at the OSI model ________ layer.

a. Physical      

b. Transport   

c. Data Link    

d. Network     

           

11. A WAN link is a connection between one WAN site and another site.

a. True

b. False           

12. The ________ is the connection point for a T1 line.

a. smart jack  

b. distribution hub      

c. DSLAM        

d. CSU/DSU

13. Each technology's maximum transmission speed is typically stated in terms of the actual limit.

a. True

b. False

14. DSL uses data modulation techniques at the ________ layer of the OSI model to achieve extraordinary throughput over regular telephone lines.

a. Data Link    

b. Network     

c. Physical      

d. Transport

15. ________ is a digital version of X.25 that relies on packet switching.

a. ISDN           

b. SONET        

c. Frame relay            

d. X.25

Practice Test

1.    ____ are connections between network nodes that, although based on potentially disparate physical links, logically appear to be direct, dedicated links between those nodes.

a.      Virtual circuits

b.      Credentials

c.       D channels

d.      Frame relays

2. BPL is a cost-effective technology that will maximize the widely deployed electrical grid as infrastructure.

a.      True

b.      False

3.  ____ is the standard for connecting home computers to an ISP via DSL or broadband cable.

a.      SLIP

b.      RAS

c.       PPPoE

d.      RDP

4. WANs that use the ____ topology are only practical for connecting fewer than four or five locations.

a.      tiered

b.      ring

c.       star

d.      mesh

5. The ____ is largest WAN in existence today.

internet

6. ISDN BRI (Basic Rate Interface) uses ____ B channels and one D channel.

a.      zero

b.      one

c.       two

d.      three

7. The typical media for ____ is UTP/STP (DS1 or DS3).

a.      PRI (ISDN)

b.      T1

c.       frame relay

d.      Fractional T1

8. ATM is a widely deployed technology that could one day overtake the Ethernet.

a.      True

b.      False

9. The maximum throughput for a T1 line is 128 Kbps.

a.      True

b.      False

10. LANs and WANs often differ at Layers 1 and 2 of the OSI model in access methods, topologies, and sometimes, media.

a.      True

b.      False

11. A ____ provides T-carrier digital signal termination.

a.      DSU

b.      CSU

c.       smart jack

d.      terminal adapter

12. An advantage of SONET is its fault tolerance.

a.      True

b.      False

13.  ____ sets ATM apart from Ethernet.

a.      Fixed packet size

b.      Security

c.       Wiring

d.      Throughput

14. In packet switching, packets belonging to the same data stream may follow different, optimal paths to their destination.

a.      True

b.      False

15. The maximum throughput for a T3 line is ____.

45 Mbps

16. Most WAN ____ are point-to-point, connecting one site to only one other site.

a.      hubs

b.      links

c.       servers

d.      switches

17. X.25 is quite suitable for time-sensitive applications, such as audio or video.

a.      True

b.      False

18. A WAN link is a connection between one WAN site and another site.

a.      True

b.      False

19. On most modern WANs, a ____ topology relies on redundant rings to carry data.

a.      star

b.      bus

c.       mesh

d.      ring

20. A ____ lease allows organizations to use only some of the channels on a T1 line and be charged according to the number of channels they use.

fractional T1

21. A star topology WAN is often the best option for organizations with only a few sites and the capability to use dedicated circuits.

a.      True

b.      False

22. The type of mesh topology in which every WAN site is directly connected to every other site is called a(n) ____.

full-mesh WAN

23.  ____ communication occurs when the downstream throughput is higher than the upstream throughput.

a.      DSU

b.      CSU

c.       Symmetrical

d.      Asymmetrical

24. Broadband cable relies on the PSTN for transmission medium.

a.      True

b.      False

25. PVCs are dedicated, individual links.

a.      True

b.      False

Chapter Test

1. A____ is the place where a telephone company terminates lines and switches calls between different locations.

            a. CO

            b. EO

            c. DO

            d. TO

2. After SONET, BPL has the best maximum throughput available.

a.      True

b.      False

3. WANs typically send data over ____ available communications networks.

            a. serially

            b. individually

            c. privately

            d. publicly

4. Broadband cable requires many subscribers to share the same local line, thus raising concerns about ____ and actual (versus theoretical) throughput.

            a. costs

            b. security

            c. noise

            d. access

5. The data rate of a particular SONET ring is indicated by its ____, a rating that is internationally recognized by networking professionals and standards organizations.

            a. CIR (committed information rate)

            b. DS0 (digital signal, level 0)

            c. BRI (Basic Rate Interface)

            d. OC (Optical Carrier) level

6. The individual geographic locations connected by a WAN are known as ____.

            a. WAN sites

            b. WAN links

            c. central offices

            d. network service providers

7.   ____ means that a telephone company connects residential users to its network with fiber-optic cable.

            a. OLT

            b. PON

            c. FTTH

            d. FITL

 

8. SONET specifies framing and multiplexing techniques at the Physical layer.

a.      True

b.      False

9. Because WAN connections require routers or other Layer 3 devices to connect locations, their links are not capable of carrying ____ protocols.

            a. nonroutable

            b. open

            c. routable

            d. standard

10. A ____ aggregates multiple DSL subscriber lines and connects them to the carrier’s CO.

            a. terminal adapter

            b. terminator

            c.  smart jack

            d. DSLAM

11.  ____________________ encapsulates incoming Ethernet or token ring frames, then converts them into ATM cells for transmission over an ATM network.

LANE

12. The portion of the PSTN that connects any residence or business to the nearest CO is known as the ____.

            a. demarcation point

            b. central office

            c. local loop

            d. NIU (Network Interface Unit)

13. A ____________________ node in synchronous communications recognizes that it should be receiving data by looking at the time on the clock.

receiving

14. The speed of a T-carrier depends on its ____ level.

            a. signal

            b. hierarchy

            c. channel

            d. traffic

15. A ____ converts digital signals into analog signals for use with ISDN phones and other analog devices.

            a. DSLAM

            b. smart jack

            c. terminal adapter

            d. terminator

16. A ____ connection is one in which a user connects her computer, via a modem, to a distant network and stays connected for a finite period of time.

            a. direct

            b. virtual

            c.  dial-up

            d. remote

17.  In a process called ____, two 64-Kbps ISDN B channels can be combined to achieve an effective throughput of 128 Kbps.

            a. merging

            b. slicing

            c. bonding

            d. linking

18. ATM is a WAN technology that functions in the Data Link layer.

a.      True

b.      False

19. A____________________ uses TDM (time division multiplexing) over two wire to divide a single channel into multiple channels.

T-carrier

20. On most modern WANs, a ring topology relies on ____ rings to carry data.

            a. serial

            b. open

            c. redundant

            d. flexible

21. ISDN PRI uses ____ B channels and one 64-Kbps D channel.

            a. 10

            b. 12

            c. 23

            d. 32

22. The ____ encompasses the entire telephone system, from the wires that enter homes and businesses to the network centers that connect different regions of a country.

            a. POTS

            b. PPP

            c. PSTN

            d. SLIP

23. SONET’s extraordinary ____ results from its use of a double-ring topology over fiber-optic cable.

            a. fault tolerance

            b. throughput

            c. latency

            d. low cost

24.  ____ is an updated, digital version of X.25 that also relies on packet switching.

            a. DSL

            b. ATM

            c. ISDN

            d. Frame relay

25. Dial-up ISDN does not convert a computer’s digital signals to analog before transmitting them over the PSTN.

a.      True

b.      False