The client and operating system software are only
part of the network picture. Whether you go with a client/server or a
peer-to-peer setup, you must have some kind of connection between
machines. In other words, before the network of file sharing and email
can become a reality, an underlying physical network must be in place.
The next few sections introduce you to the various components that
comprise the nuts and bolts of this physical network.
The Network Interface
Card
The network
interface card (NIC) is an adapter that, usually, slips into an
expansion bus slot inside a client or server computer. (External NICs
that plug into USB ports or PC Card slots are also on the market.) The
NIC’s main purpose in life is to serve as the connection point between
the PC and the network. The NIC’s backplate (that is, the portion of the NIC you can see after
the card has been installed) contains one or more ports into which you
plug a network cable.
After the physical
connection has been established, the NIC works with a device driver to
process incoming and outgoing network data. As such, the NIC is the
focal point for the computer’s network connection, so it plays a big
part in the overall performance of that connection. Each NIC is designed
for a specific type of network architecture. The most common types are
ethernet, token ring, and ARCnet, although ethernet is by far the most
popular choice for small networks.
Note
If you have a
broadband Internet connection, you’ll need two NICs in your computer:
one for the Internet connection and a second one for the network
connection.
Ethernet works by
using a carrier sense multiple access/collision detection method. This
means that ethernet cards can sense a carrier signal on the network and
so refrain from transmitting data. If no carrier signal is detected, the
card sends data. However, if two or more cards attempt to send data
simultaneously, a collision occurs. This is detected by the other cards
on the network, and no data is sent until the collision has been
resolved. (Specifically, the nodes involved in the collision resend
their packets after waiting a random amount of time.)
There are three main
types of ethernet NIC:
| Ethernet | This
type of NIC provides 10Mbps throughput. |
| Fast Ethernet | This
is a relatively new iteration of the ethernet architecture and, thanks
to its support for 100Mbps throughput, is rapidly becoming the standard
(if it isn’t already). Note, too, that many NICs
are 10/100 cards that support both Ethernet and Fast Ethernet. |
| Gigabit Ethernet | This type of card features
1Gbps throughput. This is impressive speed, to be sure, but it’s
probably overkill on a small network, unless you plan to transmit video
or other high-bandwidth content. |
Note
To achieve efficient
and reliable data transfers, any information sent over a network is
broken down into smaller pieces called packets. (You can think of a packet as the network
equivalent of a single unit of information.) Each packet contains not
only data, but also a header. The header contains information about which
machine sent the data and which machine is supposed to receive the data.
It also includes a few extra tidbits that let the network put all the
original data together in the correct order and check for errors that
might have cropped up during the transmission. A typical packet size is
512 bytes.
The Cable Connection
To set up a communications
pathway between network computers, you have to install cables that
connect the various network nodes together. The starting point
(figuratively speaking) for any cable is the network card. As I
mentioned in the preceding section, an NIC’s backplate has one or more
external ports into which you insert a network cable. The key point here
is that the cable you use must match the configuration of one of the
NIC ports. To understand why, consider the difference between telephone
cables and the coaxial wiring used by cable TV. If you examine the wall
jacks for each type of cable, you’ll see that the ports into which you
plug the cables are completely different. There’s no way, for example,
to plug a coaxial cable into a telephone jack.
It’s the same way with
NIC ports: Each has a particular shape and pin arrangement that’s
designed for a specific type of cable. When buying network adapters, you
need to match the ports with the type of cabling you intend to use.
This used to be a big deal a
few years ago when different types of cables were used. Nowadays,
however, almost all small networks use twisted-pair cable. It is composed of a pair of copper wires that
together form a circuit that can transmit data. The wires are twisted
together to reduce interference. This is similar to the cable used in
telephone wiring, but network cables are usually shielded by braided
metal insulation to further reduce interference problems. (You can use
unshielded twisted-pair cabling, but the poorer line quality will
restrict the distance between nodes and the total number of nodes.)
Twisted-pair cables use
RJ-45 jacks to plug into corresponding RJ-45 connectors in an NIC or
other type of network node, as shown in Figure 1.
In ethernet circles, twisted-pair cables are also often referred to as 10Base-T cables and RJ-45 ports are often
called 10Base-T ports.
Note
Twisted-pair
cable is categorized according to the maximum transmission rates
supported by various types of cable. For network data, for example,
Category 3 cable supports the standard 10Mbps transmission rates
available in most network installations. These days, however, few people
purchase anything but Category 5 cable, which is rated at 100Mbps and
can support higher-end network technologies such as Fast Ethernet.
More Hardware Goodies
The network card/cable
package is all that each PC requires to broadcast and to receive
network packets, but it is by no means the only hardware you might need.
Depending on the physical layout of your network, the types of services
you need, and the type of card and cable you choose, you might have to
spring for a few more trinkets. Just so you know what to expect, here’s a
list of some common network accessories:
| Hub
(also known as a concentrator) | A
hub is a central connection point for network cables. That is, for each
computer, you run a twisted-pair cable from the computer’s NIC to an
RJ-45 port on the hub. Hubs range in size from small boxes with six or
eight RJ-45 ports to large cabinets with dozens of ports for various
cable types. If you’re using Fast Ethernet NICs, be sure to get a hub
that also supports 100 Mbps. There are also 10/100 hubs available if
you’re using a mix of ethernet and Fast Ethernet. |
| Router | A
router is a device that makes decisions about where to send the network
packets it receives. Unlike a bridge (see the next item in this list),
which merely passes along any data that comes its way, a router examines
the address information in each packet and deter mines the most
efficient route the packet must take to reach its eventual destination.
For example, this is useful when the computers share a high-speed Internet connection because
the router ensures that the Internet data goes to the computer that
requested it. You should plug your high-speed modem directly into the
router. |
| Bridge | A
bridge is a device that connects two LANs, provided that the two LANs
are using the same NOS. The bridge can be either a standalone device or
implemented in a server by the addition of a second network card. One of
the most common uses for a server bridge is to split an existing LAN
into two segments. Doing so distributes the network load between the
server’s two NICs and thus improves overall network performance. |
| Repeater | Some
cables suffer from attenuation—the degradation of the electrical signal
carried over the cable is proportional to the distance the signal must
travel. A repeater is a device that boosts the cable’s signal so that
the length of the network can be extended. Some hubs also act as
repeaters (in which case they’re called active
hubs). |
| Gateway | A
gateway is a computer or other device that acts as a middleman between
two otherwise-incompatible systems. The gateway translates the incoming
and outgoing packets so that each system can work the data. For example,
you can use a gateway to connect a PC network and a Macintosh network.
The gateway takes the data from one network type and translates it into a
form that the other network understands. On a home network, a residential
gateway is a computer or router that
connects to the Internet. |
Understanding Wireless Network
Hardware Requirements
The cabling
requirements of a standard ethernet setup, and the restrictions those
requirements impose on a client, have led an increasing number of
network users to consider the cable-free configuration of a wireless
network. Wireless devices transmit data and communicate with other
devices using radio signals that are beamed from one device to another.
Although these radio signals are similar to those used in commercial
radio broadcasts, they operate on a different frequency. A radio transceiver is
a device that can act as both a transmitter and a receiver of radio
signals. All wireless devices that require two-way communications use a
transceiver.
The most
common wireless technology is wireless
fidelity, which is also called Wi-Fi
(rhymes with hi-fi) or 802.11. There are four
main types—802.11, 802.11a, 802.11b, and 802.11g—each of which has its
own range and speed limits:
| Range | All
wireless devices have a maximum range beyond which they can no longer
communicate with other devices. In practice, Wi-Fi networking ranges span from
75 feet for 802.11a to about 150 feet for 802.11b and 802.11g. |
| Speed | Wireless
transmission speed—which is usually measured in megabits per second, or
Mbps—is an important factor to consider when you set
up a wireless network or a wireless Internet connection. Less expensive
wireless networks most often use 802.11b, which has a theoretical top
speed of 11Mbps. The increasingly popular 802.11g standard has a
theoretical speed limit of 54Mbps. |
Note
Another popular wireless
technology is Bluetooth, a wireless networking standard that uses
radio frequencies to set up a communications link between devices. This
is called an ad hoc
network. The Bluetooth name comes from
Harald Bluetooth, a tenth-century Danish king who united the provinces
of Denmark under a single crown, the same way that, theoretically,
Bluetooth will unite the world of portable wireless devices under a
single standard. Why name a modern technology after an obscure Danish
king? Here’s a clue: Two of the most important companies backing the
Bluetooth standard—Ericsson and Nokia—are Scandinavian.
Wireless networks
require two device types:
| Wireless NIC | A wireless NIC is a
special type of NIC that includes (or has built into its circuitry) a
small antenna that receives and transmits data using radio frequencies.
If your network consists of only computers with wireless NICs, you don’t
need any other equipment. (However, you will have to set up your NICs
to use ad hoc mode for direct NIC-to-NIC communication; consult the
operating manual that came with each wireless NIC.) There are four types
of wireless NIC: |
| | PC Card—For notebook computers
that do not have built-in wireless capabilities, you can insert a PC
Card. Circuit
board—You can insert a wireless NIC circuit board into your desktop
computer. USB—For
easier installation on a desktop computer, you can plug a USB wireless
network adapter into a free USB port. Bluetooth adapter—To set up an ad hoc network with any
Bluetooth device, your computer requires a Bluetooth adapter, most of
which plug into a USB port.
|
| Wireless
access pointAP) ( | A
wireless AP is a device that receives and transmits signals from
wireless computers to form a wireless network. (This relatively more
permanent form or network is called an infrastructure wireless network
to differentiate such a network from an ad hoc network.) Many APs also
accept wired connections, which enables both wired and wireless
computers to form a network. If your network has a broadband modem, you
can connect the modem to a type of AP called a wireless gateway, which
extends Internet access to all the computers on the network. |
Note
If you find that your
wireless access point is not reaching certain areas of your home or
office, you can use a wireless range
extender to boost the signal.
Depending on the device and wireless access point, the extender can more
than double the normal wireless range. Bear in mind, however, that
range extenders are notoriously difficult to incorporate into an
existing network. For best results, use an extender from the same
company that makes your wireless access point, and make sure that the
extender is compatible with the access point.
