Network technicians are often charged with troubleshooting a problem on a network with which they are not familiar. A technician may work for an organization with a large network and be summoned to a remote location, or he or she may work for a consulting company and regularly travel to various client sites. As a technician, when you are faced with an unfamiliar network installation, your first order of business should be to determine the network's basic hardware configuration. In a perfect world, the people who designed and built the network would have meticulously documented their work, and you would have access to schematic diagrams of the cabling and an inventory of the network hardware. This is rarely the case, however, and often you must approach a problem with no information about the network infrastructure and the equipment used to build it. When this happens, you must be able to examine the equipment yourself to find out what's what. This lesson is designed to help you recognize the various types of network components when you see them.
Computers have many different ports on them, and although most of the ports on computers manufactured today are labeled, you might come across units that have no markings. When this occurs, you must be able to discern the functions of the various ports so that, when you are working with the computer, you connect it properly to external devices.
The ports on a typical computer are located either on the motherboard or on expansion cards that plug into the system bus. In a few cases where the same type of port can have more than one function, you can sometimes tell what a port does by where it is located. The location of the motherboard ports can vary depending on the design of the computer. Figure 17.1 shows the back of an older computer, which has the motherboard ports grouped in a separate area away from the motherboard itself. In this type of computer, the motherboard ports are separate modules that are bolted to the back of the computer case and connected to the motherboard with cables. The computer's expansion slots are grouped in a different area, where you find the ports provided by the expansion cards installed in the system.
Figure 17.1 The back panel of an older computer
Figure 17.2 shows a newer computer with all of the motherboard ports lined up in a row. On this computer, the ports are actually part of the motherboard, which, in this figure, runs horizontally along the bottom of the computer. All ports on this computer are located on the motherboard, including the video port, which, in most cases, is provided by an expansion card.
Figure 17.2 The back panel of a newer computer
Serial ports are nearly always used by modems, but there are a few proprietary networking systems that use standard unshielded twisted pair (UTP) cable that plugs into an adapter connected to a serial port. Computers manufactured today typically have two serial ports on the motherboard, both of which use male DB-9 connectors, as shown in Figure 17.3.
Figure 17.3 Today's computers have two DB-9 serial ports
Older computers may also use DB-25 connectors for serial ports. Because nine pins more than satisfy the requirements for serial communications, a computer can use either type of connector. The typical configuration used to be one DB-9 and one DB-25, as shown in Figure 17.4. Serial connections can run at speeds that range from 110 to 115,200 bits per second (bps).
Figure 17.4 Older computers use both DB-9 and DB-25 connectors for serial ports
D-type connectors are not used exclusively for serial ports. Parallel ports also use DB-25 connectors (although they are female instead of male), and, in rare cases, SCSI adapters use DB-25 connectors as well. In addition, DB-9 connectors were at one time used as video ports for connections to monitors, although they were female instead of male. Be sure not to confuse them, because their functions are not interchangeable.
Parallel ports were designed for use by printers, and in recent years they have been enhanced to provide more efficient bidirectional communications between the computer and the attached device. Parallel ports always use a female DB-25 connector, and the average computer today is only furnished with one (see Figure 17.5). In addition to printers, there is a select group of SCSI devices that plug into a computer's parallel port. These devices use the same basic principles of communication as standard SCSI devices, but they use a different interface to the computer. The parallel port interface is slower than that of a dedicated SCSI host adapter but, for many users, the ability to move a SCSI device from computer to computer without installing an adapter card is worth the sacrifice in speed.
Figure 17.5 The female DB-25 connector used by parallel ports
Although nearly all computers today have two serial ports and one parallel port, it is still possible to add additional ports using an input/output (I/O) expansion card. However, because of the rise in popularity of the universal serial bus (USB) and the gradual elimination of the Industry Standard Architecture (ISA) slots that most of the I/O cards plug into, it is rarely necessary to add more serial and parallel ports.
The video connector on computers today cannot be mistaken for any other port because no other devices use the same female 15-pin D-shell connector with three rows of five pins each (shown in Figure 17.6). Before the introduction of the Video Graphics Array (VGA) standard in 1987, which eventually gave its name to the 15-pin connector, computers used digital monitors that connected to a female DB-9 port in the video adapter.
Figure 17.6 The 15-pin VGA connector used for connecting a monitor to the computer
The computer's keyboard and mouse (or other type of pointing device) always require ports, which are typically on the computer's motherboard. On today's computers, both devices use the same type of port, which has a round, female, six-pin connector called a mini-DIN (DIN stands for Deutsche Industrie Norm, the name of the German organization that developed the standard for the connector), shown in Figure 17.7. This connector has also come to be known as a PS/2 connector, which comes from the name of the IBM computer model that first used it. Although it took far too many years to become standard practice, the mini-DIN connectors on computers are now usually labeled with pictograms that indicate which one is for the keyboard and which is for the mouse, as shown here in Figure 17.7. Accidentally cross-connecting the keyboard and the mouse does not cause any damage to the hardware, but the two ports are not interchangeable.
Figure 17.7 The six-pin mini-DIN connectors used for the keyboard and mouse today
Before the mini-DIN connector became standard equipment, keyboards used a larger, female five-pin DIN connector, illustrated in Figure 17.8. Mice and other pointing devices used either DB-9 serial ports or a port on a dedicated "bus mouse" card.
Figure 17.8 The five-pin DIN connector formerly used for keyboard connections
In a few cases, computers and other peripherals use the older five-pin DIN connector to carry electrical power from a transformer to the device. Plugging a power connector into a keyboard port can severely damage a computer. However, because few computers today use the five-pin DIN keyboard connector, the danger of this happening is remote.
The universal serial bus (USB) is a relatively recent innovation that is rapidly replacing many of the ports commonly included on personal computers, such as the serial, parallel, keyboard, and mouse ports. It is a multipurpose bus that runs at up to 12 megabits per second (Mbps) and supports a wide range of devices, from keyboards and mice to cameras and disk drives, all using the same interface. Computers today typically have two USB ports, which use rectangular, female, four-conductor connectors (shown in Figure 17.9) that are totally unlike any other port in the computer. This port is called an A-connector; USB devices have B-connectors on them, which are more square-shaped.
Figure 17.9 USB A-connectors
The small computer systems interface (SCSI) is a mass storage interface that supports many different internal and external devices at speeds up to 160 Mbps. Network servers often use SCSI for internal hard drives and other storage media, but they can also have external SCSI connections using any one of several different cable connectors. Since its inception in the early 1980s, this interface has undergone several revisions to increase its speed and capabilities. These revisions have necessitated the use of different cables and, consequently, different connectors.
Implementations of SCSI typically involve a host adapter card that plugs into the computer's expansion bus. Occasionally, you might encounter a computer with a SCSI adapter integrated into the motherboard. The host adapter usually has both internal and external connectors. Internally, SCSI uses ribbon connectors that attach to hard drives and other devices. External SCSI cables are thick and relatively inflexible because of their heavy shielding and the tight bundling of wires contained inside. The earliest SCSI implementations used a 50-pin Centronics connector for external connections, which was identical in design to the Centronics connector used on most printers, except that it was larger. This SCSI bus is 8 bits wide and runs at 5 Mbps. Sometimes this same bus also used a standard DB-25 connector, such as on early IBM, Apple Macintosh, and Sun Microsystems computers.
Newer SCSI implementations use unique 50-pin high-density or 68-pin high-density connectors on their cables. These connectors have two rows containing equal numbers of pins that are smaller than those in a DB-25 connector. The 50-pin version has clips on the connector that connect it to the port on the computer, and the 68-pin version has thumbscrews instead. The 50-pin connectors are used primarily on Fast SCSI implementations, running at 10 Mbps, whereas the 68-pin connector is used for Fast/Wide SCSI, running at 20 Mbps.
Another type of SCSI used by some drives today is called Single Connector Attachment (SCA) SCSI, which uses a special 80-pin connector that supplies both the power and the data connections to the drive. The connector is similar in appearance to a Centronics connector, although a bit wider, and in many cases people use an adapter to attach SCA drives to a standard power connector and a 50-pin or 68-pin high-density SCSI connector. Figure 17.10 illustrates the various types of external SCSI connectors you might find on a computer.
Figure 17.10 SCSI host adapter connectors: DB-25, 50-pin Centronics, 50-pin high-density, 68-pin high-density, and 80-pin SCA
Network interface cards (NICs) can often have several different types of connectors on them, but you can only use one of the connectors at a time. This type of adapter is called a combination NIC or combo NIC. Additional connectors often add substantially to the cost of the NIC. The type of connector you use to attach your computer to a network depends on the data-link layer protocol and the type of cable your network uses. Ethernet NICs often have three network connectors on them: an Attachment Unit Interface (AUI) connector, a Bayonet-Neill-Concelman (BNC) connector, and an RJ-45 connector. These network connectors are shown in Figure 17.11.
For more information about NICs and network cables, see Chapter 2, "Network Hardware."
Figure 17.11 A combination Ethernet NIC with RJ-45, AUI, and BNC connectors
The AUI connector, the oldest of the Ethernet cable connectors, is a 15-pin female D-shell connector with two rows of pins. You use the AUI connector to attach an AUI cable to the NIC in your computer. The other end of the AUI cable connects to a thick Ethernet network. Although it's still common to see Ethernet NICs with an AUI connector (in addition to others), it's rare to see the connector actually being used. The BNC connector is for attaching a computer to a thin Ethernet network. You must attach a special T fitting to the connector on the NIC, and you then attach the Ethernet cables to the arms of the T, as shown in Figure 17.12. This enables you to run the cable from computer to computer, forming a bus topology. In some cases, the T connector is included with the NIC; in other cases, you must purchase it separately.
Figure 17.12 Thin Ethernet connections require a T connector, which you must attach directly to the BNC connector on the NIC
The connector used on most Ethernet networks installed today is called the RJ-45. Because the vast majority of Ethernet networks today use UTP cable, purchasing NICs with just a single RJ-45 connector, like that shown in Figure 17.13, is usually the most economical course of action. RJ-45 connectors are similar in appearance to telephone jacks, except that the RJ-11 telephone connector has four (or sometimes six) contacts and the RJ-45 has eight. An RJ-11 plug is slightly narrower than an RJ-45, but it's easy to mistake the two. You can insert an RJ-11 plug into an RJ-45 jack, but you can't plug an RJ-45 plug into an RJ-11 jack.
Many computers have internal modems, so it's quite common to see two expansion cards in one computer and RJ-11 and RJ-45 jacks that look similar. Obviously, confusing the two connectors can lead to problems. Plugging a telephone cable into a NIC's RJ-45 connector won't cause any damage, but neither will any network communication take place. (This is unlike other situations in which confusing this type of connector can be more serious; plugging a standard analog modem into a digital telephone jack connected to a switchboard, for example, can ruin the modem.) In most cases, the easiest way to tell modem connectors from NIC connectors on an installed expansion card is by the number of connectors. Most modems have two RJ-11 connectors, one for the connection to the telephone line and one for a telephone, whereas NICs only have one RJ-45 connector.
Figure 17.13 An Ethernet NIC that uses an RJ-45 jack
Token Ring networks can use UTP cables, just like Ethernet, so the connector you see on the NIC is the same female RJ-45. Token Ring networks that use IBM Type 1 cabling have female DB-9 connectors on their NICs, however. This type of Token Ring cable uses a male DB-9 connector on one end to connect to the NIC, and an IBM data connector (IDC) on the other end to connect the computer to its multistation access unit (MAU). In some cases, however, administrators trying to avoid the expense of the Type 1 cabling use devices called Token Ring media filters to connect their Type 1 NICs to a UTP network. The media filter is essentially an adapter with a male DB-9 connector and a female RJ-45 connector on it, as shown in Figure 17.14. You plug the DB-9 connector into the NIC and connect a standard UTP network cable to the RJ-45 connector.
Figure 17.14 A Token Ring media filter
In addition to being knowledgeable about the ports used to connect computers to the network and to other devices, network technicians must also be familiar with the other types of equipment found at a network location.
Network interface adapters, or NICs, are similar in appearance to other types of expansion cards. NICs are manufactured for each of the bus types commonly found in computers, although the most commonly seen today are Peripheral Component Interconnect (PCI) and ISA NICs. The primary way to distinguish a NIC from another type of card (apart from examining the chips, which often have the name of the manufacturer printed on them) is by looking at the connectors, which can be any one of the types mentioned in the earlier section, "Network Cable Connectors."
Motherboard-resident network interface adapters don't require a bus slot, and the electronics can easily be lost in the maze of motherboard circuitry. However, the presence of an RJ-45 jack on the back of the computer indicates support for UTP Ethernet.
See Lesson 2: Network Interface Adapters, in Chapter 2, "Network Hardware," for more information about NICs.
A hub can be either a stand-alone box or a unit that mounts into a standard 19-inch-wide rack used for large network installations. Whatever the form, the basic identifying feature of a hub is one or more rows of female connectors, as shown in Figure 17.15. Most of the hubs you'll encounter have rows of RJ-45 connectors for Ethernet or Token Ring cables, but there are also hubs that have straight-tip fiber optic connectors, which are bayonet-style connectors; Type 1 Token Ring connectors, which are IBM data connectors; and others. A hub can have as few as four ports in it, or as many as 24.
Figure 17.15 The back of a 10Base-T/100Base-TX Ethernet hub
On an externally wired network segment, a single cable connects the network interface adapter in each computer to a hub port. On an internally wired network segment, each hub port is connected to a patch panel port using a short patch cable. The patch panel is connected to all of the cable runs inside the building's walls and ceilings, terminating at wall plates throughout the installation. Another patch cable connects each wall plate to a computer.
In most cases, a hub has rows of light-emitting diode (LED) lamps that correspond to the network cable ports, as shown in Figure 17.16. These LEDs are activated by the link pulse signal generated by a network interface adapter when it is active and functioning properly. An LED that is not lit indicates a malfunction in the hub or the network interface adapter, a broken or disconnected cable, or a computer that is turned off or missing.
Figure 17.16 The LED display on the same 10Base-T/100Base-TX Ethernet hub
Hubs can support any one of several data-link layer protocols. The most common is Ethernet, but you can also find hubs that support the Fiber Distributed Data Interface (FDDI), Token Ring, or other protocols. These hubs might be called by different names (a Token Ring hub is called a MAU, for example) and use different types of connectors, but their general appearance is the same.
See Lesson 3: Network Hubs, in Chapter 2, "Network Hardware," for more information about hubs.
Switches are very similar in appearance to hubs, and units made by the same manufacturer can indeed look identical, except for their markings. The difference between a hub and a switch is in the internal manipulation of incoming data. A hub forwards incoming traffic out through all the other ports, but a switch forwards traffic only to the device for which it is destined. Switches are available in most of the same configurations as hubs, and they range from small units intended for home or small business networks to large rack-mounted devices.
Some switches have an additional nine-pin serial port on them, which you use to connect the device to a computer using a null modem cable. This interface enables you to perform the initial configuration of the unit before it's connected to the network. After the initial configuration is complete, you can access the switch's management functions by using a Telnet session from a computer on the network.
Patch panels have rows of female RJ-45 jacks or other connectors, just like hubs, but they don't perform any other function besides providing connections between internal cable installations and patch cables. Patch panels, also called punchdown blocks, are typically mounted either on a wall or in a rack; they are not stand-alone units. In many cases, patch panels are modular, consisting of a framework and interchangeable connectors, as shown in Figure 17.17.
Figure 17.17 A modular patch panel
As with hubs and switches, bridges and routers can be stand-alone or rack-mounted units. Because bridges and routers only connect two network segments together, they don't have as many ports as hubs do, and are often more difficult to recognize as a result. A stand-alone bridge or router can take the form of any type of box with two or more ports in it. It sometimes has a few LED lamps for status displays, as shown in Figure 17.18.
Figure 17.18 A typical stand-alone router
Bridges and routers can connect local area networks (LANs) of the same type or different types, or they can connect a LAN to a remote network using a leased telephone line or other wide area network (WAN) connection. For a router connected to a WAN link, the identifying element of the unit is a serial port that is connected to a channel service unit/data service unit (CSU/DSU). The CSU/DSU provides the interface to the leased line, much like a modem provides the interface to a standard telephone line, except that the CSU/DSU is a digital device. In addition to the serial port, the bridge or router also has one or more ports for connecting to the LAN and uses any of the standard cable connectors, such as female RJ-45s.
A bridge or router used to connect two internal network segments can have two ports of the same type, such as two RJ-45 jacks, or two different ports, which connect to different types of networks. In most cases, bridges used internally are transparent bridges, which connect the same type of network, so they have two identical ports. A device with two different ports, such as an RJ-45 jack and a BNC connector, is more likely to be a router. However, routers can also have two identical ports, either because they connect two LANs of the same type or because they connect two different types of LANs that use the same cable, such as 10Base-T and UTP Token Ring. Larger routers can be modular, consisting of a frame with multiple slots into which you plug separate modules supporting a specific type of network, as shown in Figure 17.19. This enables you to custom-build a router configuration to suit virtually any combination of network technologies.
Figure 17.19 A router frame with modules installed
A print server is a device that receives print jobs from clients on a network and feeds them to the printer at the appropriate rate. The print server sometimes spools the print jobs, meaning that it stores them on a hard drive while waiting for the printer to be ready to receive them. This enables network users to share a printer without experiencing delays in applications while they wait for the printer to become available.
In some cases, a print server is a computer. You can connect a printer directly to a computer and share it with other users on the network. However, there are also stand-alone print servers, which take the form of either an expansion card that you install in the printer itself or a small box with one or more ports for connecting to the printer (or printers) and another for connecting to a network hub, as shown in Figure 17.20. The device has an Internet Protocol (IP) address of its own, and computers on the network send their print jobs to the print server, which relays them to the appropriate printer. This enables users to locate a printer anywhere on the network.
Figure 17.20 External and printer-internal print server devices
An uninterruptible power supply (UPS) provides equipment with a temporary supply of electrical power in the event of an outage in the building's main power supply. A UPS is essentially a battery that is continuously charged when the main power is on. You plug computers or other equipment into the UPS, and if the power fails, the battery supplies power for a short period of time, long enough to safely turn off a computer without damaging its data. UPS devices are available in models ranging from small units designed to protect a single computer to larger devices that can protect several computers at one time, as shown in Figure 17.21. It's even possible to install a huge UPS that can protect an entire data center or building.
Figure 17.21 Uninterruptible power supplies
A UPS (see Figure 17.22) is a heavy box with a series of power outlets for your equipment (typically on the back) and a standard electrical plug that connects to the building's power source. Larger devices might require special power connections. In many cases, a group of LEDs displays the amount of power left in the battery and the load generated by the connected equipment. Every network server containing important data should be connected to a UPS.
Figure 17.22 The back panel of a UPS
The best UPS units are those that run connected devices off the battery at all times, while the battery is continuously charged. This way, when the building power fails, the equipment experiences no interruption at all. Lower priced units run the equipment from the building's power supply until it fails, at which time it switches over to battery power. The brief interruption during the switchover can sometimes be enough to interrupt a disk writing process, causing data corruption. High-quality units also have one or more serial ports that you can use to connect the UPS to a computer. A program or service running on the computer receives a signal from the UPS when building power is cut off, and after a specified interval the computer shuts itself down in a controlled fashion. In this way, the UPS continues to provide protection against data loss, even when no one is present to shut down the computer.
Match the ports listed in the left column with the correct connector descriptions in the right column.
1. VGA video port
2. Serial port
3. SCA SCSI port
4. 10Base-T Ethernet port
5. USB port
6. AUI port
7. Keyboard/mouse port
8. Parallel port
9. 5 Mbps SCSI port
10. Thin Ethernet port
a. Rectangular four-conductor connector
b. Combined power and data connector
c. Mini-DIN connector
d. Three-row, 15-pin D-shell connector
e. Female DB-25 connector
f. 50-pin Centronics connector
g. Thick Ethernet connector
h. Male DB-9 connector
i. BNC connector
j. RJ45 connector