Wireless Community Networks

A Guide for Library Boards, Educators, and Community Leaders

Version 1.0

Chapter 6. Building a Local Area Network

Before we can really begin talking about wireless data communications, let's take a step backward and discuss how computer networks work. At this point neither of us has time for a complete study of the subject—that's what college classes are for. So we'll run through just the basics here, one at a time.

Specifically, we'll review all the parts that make up a basic computer network, one that connects the computers in a single location (building or department). Such a network is called a local area network, commonly referred to by its acronym, LAN. When we finish, you should have the basic knowledge you need to be ready for your next discussion of computer networking.

How Computers Converse

The purpose in creating computer networks is to allow multiple computers to share a resource that would be too expensive to provide for every computer user. For example, instead of buying one printer for every computer, an agency might buy a single, high-speed, high-quality laser printer and install it on a computer network. Every computer connected to the network could potentially have access to the centralized printer (agency or departmental network policy determines who can use a centralized resource and a network administrator configures each computer in accordance with the policy). This generally saves money and allows the agency to purchase a higher quality product.

In order for several computers to share limited resources, there have to be rules in place for communication. It's like having a human conversation, except that most of our "rules" are taken for granted. Let's think about a conversation for a moment and look at the rules in place governing our communication.

First, we have to agree on a language to use. In this book's context, we've chosen English. Secondly, we have to formulate rules of transmission "right of way." We recognize that the two people conversing can't both talk at the same time. (Not that we don't try to at times!) So some rules, usually unspoken but learned through experience, are in place. For example, if I'm impatient to make a point, I may begin to utter a sentence only to stop when I realize that the current speaker is not finished. Or perhaps another person might barge in before I can begin. The original speaker may take this as a sign to yield at an opportune moment and give me a chance to ask a question or make a point. In more formal situations we may be required to raise a hand to indicate we would like to interrupt to make a point.

In computer networking, the rules for the "language" used and for the process of starting, interrupting, and continuing communication are called network protocols, or just protocols. In order for a computer to participate in a network of computers, the same protocol software must be installed on each computer. Computers communicating over the Internet use a specific protocol called TCP/IP. The same computers can participate in a local area network using a different protocol, such as IPX or NetBEUI.

How the Signal is Transmitted

When we carry on a conversation, the words of our speech are carried over a communications mediumspecifically through the air around us as sound waves. Our vocal cords and mouth form and transmit the sound waves, and our eardrums receive the sound. (This is, by the way, a very good demonstration of a wireless communication.) But the range of communication is limited, usually to several hundred yards. How loud we can yell (amplify the signal) determines how far the signal will travel as an identifiable sound before it degenerates into something unintelligible. The power of our lungs has a direct bearing on how much the signal is amplified.

On the other hand, let's say you want to speak to a friend in Milwaukee. You will have to use some form of wired connection—a telephone. The wired infrastructure intensifies or amplifies your voice at various points along the wire so that your friend in Milwaukee hears your voice as if you were in the same room. If the sound is not transmitted very well, you can use an amplifier in your phone set to increase the "gain" of the signal, making the signal louder in your ear.

So whether close or far away when we carry on a conversation, we need several components in place to communicate with each other:

  • lungs, vocal cords, tongue, teeth and lips to transmit a signal
  • the air (or some other medium) to carry the sound vibrations (signal)
  • ears, including the eardrum and the fine bones of the inner ear, to receive the signal.

Like humans, computers need the same type of components: a transmitter, a transmission medium, and a receiver. In the next section we discuss the various physical components required to create a computer network.

General Network Equipment

Network Interface Cards

Which is the most basic component of a computer network: the interface card, the transmission medium, or the protocol drivers? It depends on one's philosophy. We'll tackle the interface card first.

In order to get a computer connected to a network, a network interface card (also called a NIC or network card for short) must be used. This is just another of the green electronic circuit boards you've grown used to seeing around computers. This particular circuit board has one or more port connectors attached to one end. The port extends to the outside of the computer case, on the rear panel.

The most common port looks just like a modular telephone jack, only a little larger. This connector is for an eight-wire cable (a regular telephone jack has only four). Technically it's known as a RJ-45 jack. Another common port looks like a silver barrel, not quite a half-inch in diameter. This connector is called BNC port. The type of port your network card requires is determined by the cable type you use for your network. We'll discuss this in more detail in the next section.

The interface card has programming contained on it that provides several network functions, including transmission and reception of network signals. The interface card prepares the data and transmits it over a transmission medium. It also receives any network signals coming to the computer in which it is installed. It converts the signal to the data originally sent and makes it available for use on its host computer.

The network protocol software mentioned above and the computer's operating system software must work very closely with the network interface. This is accomplished by a third piece of software called a driver. A driver is software written specifically for a particular manufacturer's network interface card, allowing the network operating system to control (or "drive") it. 

Common Cabling Types

Cabling doesn't really fit in the category of equipment, but it is one of the physical components required to create a network. The cable connects to the network interface card and, usually, to a centralized network device called a hub. We'll discuss the hub in the next section.

Coaxial. One of the oldest cable types for computer networks is thick coaxial cable, such as the type used in your cable TV connection. This cable is designated as 10Base-5 cable. While there are some good reasons to use regular coaxial cable, it has fallen out of general use in favor of fiber optic cable. We'll give it short shrift here.

Thin Coaxial. Sometimes called thinnet cable, thin coaxial cable gained great popularity is small, departmental LANs in the late 1980s. It is designated as 10Base-2 cable. Thinnet cable is more pliable and is much easier to install and maintain than regular coaxial cable. There are still many LANs partially or completely configured with thinnet.

The most serious drawback to thinnet is that it limits to 30 the number of computers that can be connected to a network without requiring additional, expensive equipment. Once thirty computers are attached to the network, a repeater, costing several hundred dollars, is required to extend the LAN and attach additional computers. Also, it is difficult to move computers from location to another. No "open connections" are allowed along the network cable.

Because of these limitations, and the relatively higher cost of thin coaxial cable, another common type of cabling became popular in the early 1990s.

Twisted Pair. Twisted pair cable is the type you commonly see used for telephone cable. It comes in two flavors, shielded twisted pair (STP) or unshielded twisted pair (UTP). Shielded cabling adds a layer of protective "foil" inside to prevent many types of electromagnetic interference. The shielding makes it is much more expensive.

Unshielded twisted pair cabling is the most common type of cable in use today for computer networking. We will limit our discussion just to the levels of UTP cabling installed in many buildings. There are currently five levels, each of which differs in the number of twists per foot provided. More twists required more cable; therefore, as one goes up the levels, pricing gets higher, but so does signal quality. The more twists in a UTP cable, the more able it is to resist the effects of crosstalk, signal interference created by the other pairs of wires inside the cable. The five levels (or categories) are described below:

  • Category 1. Early unshielded twisted pair cable used for telephone systems. It is not capable of carrying data signals. If your twisted pair wiring was installed prior to 1983, it is probably traditional Category 1 phone cabling.
  • Category 2. This level of UTP cabling is certified to carry data signals at acceptable networking levels. It uses four twisted pairs of copper wiring and provides bandwidth up to 4Mbps. It was used in some original LAN implementations.
  • Category 3. As demand for higher bandwidth network connections developed, a higher-quality UTP cable was designed. It supports bandwidth of up to 10Mbps, the speed of standard Ethernet networks. Many early LAN implementations for Ethernet networks used Category 3 cable.
  • Category 4. With some network architectures supporting speeds of 16Mbps, cable manufacturers increased the performance of their cable again, up to 16Mbps to match this standard. The minimum level of cabling that should be considered for current LANs is Category 4. However, in all current or planned network installations, the next level is recommended for installation simply because it costs almost the same as Category 4 and provides more versatility in future use.
  • Category 5. Also called Cat5 cable. The currently recommended level of UTP cable for new installations or upgrades. It is designed to provide bandwidth up to 100Mbps. All installations of Fast Ethernet should be cabled with certified Cat5 UTP cable.
  • Future Levels. Enhancement of UTP cables is ongoing. Already some Cat5 cables are certified to deliver data at approximately 300Mbps. With gigabit Ethernet (1000Mbps) already appearing in the business and research sector, one can expect continued enhancement of UTP cabling.

Fiber Optics. Fiber optic cabling is very different from any other cabling. It does not use copper at all, and the signals it carries are not electrical signals. Inside the cable jacket are very small glass tubes through which high energy light is transmitted. Because it uses light, fiber optic cabling is immune to any electrical interference. It's also hard to tap by unauthorized users, so it is a very secure medium.

It is primarily used where high-speed transmission or longer distance is required. Cable length can range from six-tenths of a mile to 6.4 miles, depending on the type of cable and network signaling used. Fiber optic cable is much more expensive than copper cables to purchase and install.

Radio Frequency Wireless. Obviously, wireless technology does not involve cables (wires). However, it does provide network connectivity in specialized local area situations. It is used in internal settings where it is not possible to run cabling effectively (such as older historical buildings which cannot be modified) or where mobile computing is required (such as factory or retail floors). Such internal wireless connections can be made over several hundred feet. Over such distances, the signal is strong enough to penetrate some types of walls and doors. 

Network Devices

So far we have discussed network cards and cabling, the two basic building blocks for local area networks. A network card is installed in every computer to be networked. Cable is connected to each network card. The only remaining question is how to connect all the cables together.

Tranceivers. Older coaxial cabling used an Ethernet transceiver to connect the computer cable to the main network cable, referred to as the network backbone. Thin coaxial cable uses a small T-connector to connect the backbone cable directly to a transceiver on the network card. But newer network installations using UTP cable are configured quite differently. These networks use a device called a hub.

Hubs. When using UTP cabling, an Ethernet network is physically configured in an arrangement called a star. Cabling from each computer's network card runs directly to a centralized hub, named after the hub-and-spoke pattern produced when such a network is charted on paper.

The hub acts as a concentrator. It concentrates the signals from every computer attached and then distributes the signals back to all the computers. This is how basic communication occurs over the network. Each computer broadcasts a signal to all other computers on the network. The appropriate computer accepts the signal and sends an acknowledgment. All others ignore it.

Some hubs, generally the more expensive ones, contain electronic circuitry to clean up and regenerate the signals. This function is normally associated with another device called a repeater. (We don't cover repeaters here, because they are seldom used separately unless long runs of cable are required.) Therefore, some documentation refers to them as multi-port repeaters. Due to the popularity of the term hub, we will use it exclusively in this manual.

Each connector on a hub is called a port. So the plug from the cable of each computer on the network is plugged into a hub port. Hubs normally have 4, 8, 12, 16, or 24 ports. It is possible to connect one or more of these ports to another hub. This daisy-chaining of hubs allows the number of computers which can be connected to the network to grow.

Network Operating Systems

Once the hardware is in place, a physical network exists. However, communication is still not possible. Next in the development process is the computer's operating system. A computer operating system simply provides basic control of the various components that make up a computer: the keyboard, mouse, monitor, hard drive, processor, etc. Software developers use the operating system to work with the various components of the computer.

In order for a computer to communicate on a network, a secondary operating system, called the network operating system, is required. This software allows the computer to operate the network card and use shared resources on the network. Most computers sold today have a network operating system embedded in the main operating system: Windows 95/98, Windows NT, Macintosh OS, OS/2, and, for high-end workstations, Unix. There are third-party software packages, such as LANtastic and NetWare, which also provide alternative network capabilities to DOS and Windows-based computers.

Some older Windows 3.1-based computers can be upgraded to Windows for Workgroups, which includes the network operating system components required to allow Windows 3.1 to communicate on a network.

Shared Services

The last, but most important, part of a local area network is the resources one wants to share among multiple computers. These include files, printers, and programs. Each of these requires a computer to act as a servant to the rest of the computer. These are called servers. The following paragraphs describe the most common network servers.

File Servers. Computers with large, fast hard drive systems can be used to provide access to central files for all computers on the network. These enhanced computers are called file servers. One of the primary uses of a file server is to provide access to centralized databases. Electronic library catalogs are one example of such databases. Another is the community non-profit agency database that some libraries maintain. Schools will use a file server to maintain records on their students.

Once an organization gets dedicated access to the Internet, there are a number of other servers the organization may need. To provide information on a web site, a web server is required. To provide e-mail capabilities to staff, an e-mail server is required. Schools may find servers necessary to provide in-house chat or videoconferencing services in support of curriculum.

A specialized form of a file server is used to provide networked access to software applications. These are, quite naturally, called application servers.

Shared Printers. Printers attached to a network server can be made available to any computer on the network. The computer to which such printers are attached is called a print server. Installing a network card in the expansion slot provided on some printers can make them available on a network without a computer. These are called network printers.

Sharing printers can allow an organization either to save money by reducing the number of printers required or to provide a higher quality printer than it normally could afford.

Shared CD-ROM Drives. In many library environments, both in schools and public libraries, many informational databases on CD-ROMs may be available. Some products specifically of interest to municipal and county offices are also available. It is possible to provide access to as many as 21 or more CD-ROM disks at the same time to any computer connected to the network by dedicating one computer to the task. Such computers are called CD-ROM servers. Obviously an organization must pay appropriate licensing fees for each product based on the number of users with access to it.

Electronic Messaging. Another common and beneficial use of networks is to provide electronic messaging capability, commonly called electronic mail, or e-mail for short. Depending on the network messaging software used, messages can be broadcast to all or any subset of the organization's staff. In the case of schools, it can also be used to provide student e-mail access for collaborating with students around the world or communication with teachers or other students.

Summary

In this section we have provided a quick overview of the components in a local area network (LAN). We covered the following concepts:

  • Protocols—the "language" and rules of communication chosen so that a group of computers can communicate.
  • Network Equipment
    • Network cards—an adapter allowing a computer to send and receive data over the network.
    • Cabling—commonly Category 5 unshielded twisted pair, which provides a medium over which the data signals are carried.
    • Hubs—a network device that broadcasts the data signal from one computer to all other computers.
  • Network Operating System—the software that controls the network card and allows software on one computer to send commands and responses to other computers, printers, or CD-ROM drives on another computer.
  • Shared Resources
    • File servers—provide centralized access to organizational databases and other files.
    • Print servers and Network printers—provide shared access to high-end printers.
    • CD-ROM servers—provide centralized access to CD-ROM informational products.
    • E-mail servers—provide electronic communication capabilites to staff and users.

 

BACK | CONTENTS | NEXT 

 

Line
Wireless Community Networks

Written by Robert L. Williams.


The web version was last updated on April 26, 1999.


Send comments to rlwconsult@aol.com.

Line

Page last modified: March 2, 2011