Read an Excerpt
Before you grab a screwdriver, drill a hole for cable, or open your computer to add networking components, you need a plan. When the ultimate plan is a small network in your home or in your place of business, you need to take a step back and make some decisions before you pick up your tools. (Maybe I should call that a preplan to a plan?) You have a lot of choices about the way you design and install your network, and you can�t do anything definitive until you understand them. In Part I, you�ll learn what a network is, how it works, and the pros and cons of the various hardware options. You�ll also learn how to install hardware networking components and how to add peripherals (printers, modems, DSL devices, and so on) to your network.
A network is nothing more than two or more computers that are connected so they can exchange data. The largest network in the world is the Internet, which comprises all the servers that provide services to Internet users. The smallest networks are found in millions of homes, where home users have connected two or more computers. In between those extremes are millions of corporate and small business networks comprising anywhere from two to hundreds of thousands of connected computers. What you may not realize is that all networks, regardless of their size, share common characteristics:
- Every computer contains a hardware device that controls communication with the other computers on the network.
- Every computer has a connection method (wireless or wired) that sends data to the other computers, using the hardware communication device.
When you create a home network or a network for your business, you�ll be using the same approach as the network designers who create enormous corporate enterprises. As far as hardware and connection requirements, your network is no different than any other network in the world.While this may seem overwhelmingly complicated, you�ll be amazed at how simple and logical it is if you take it one step at a time.
You do have some decisions to make, regarding the type of connections you want to use between your computers and the way in which users will join the network. This chapter provides an overview of your choices, and all the other chapters in this book will help you implement those choices.
Types of NetworksNetworks are designed to be either client/server networks or peer-to-peer networks. Each type has its advantages and disadvantages, but small networks are almost always created as peer-to-peer networks. Regardless of the network type you opt for, the hardware and connection devices you have to install in each computer are the same.
Client/Server NetworksClient/server networks are built for security and controls. A server, equipped with special software tools, checks the computers and users who are logging in to the network to make sure that only authorized components can access the computers and other resources on the network. Information, including names and passwords, about each computer and user is stored in a database on the server. As each computer boots, and each user logs on, their logon data is checked against the information in the database.
The server can also limit the rights and privileges of computers and users who log on to the network. Some users may be restricted from working with files in certain folders on network computers, or they may not be able to change settings on their own computers. These controls are imposed on a computer-by-computer or user-by-user basis, but to make it easier for administrators, computers and users are placed into groups. The restrictions applied to a group apply to each member of the group. Assigning rights and privileges on a group basis is easier than dealing with each computer and user one at a time.
Most client/server networks are also designed for server-based work. Servers are set up all over the network (and they must also log on and be authenticated). Most of these servers are assigned a specific task; consider the following, for instance:
- A mail server that holds the company�s e-mail software and the users� mailboxes
- The accounting software server, which can be accessed only by members of the accounting department
- Print servers that control the printers attached to them, so users can share printers
- Data servers that hold the documents that users create with software applications such as word processors or spreadsheets
NOTE: In client/server environments, the software that each user runs is frequently configured to save documents to a data server instead of to the user�s own computer. This makes it easier to share data among users and also makes it easier to back up the data. (All administrators know that users rarely obey company mandates to perform backups of their own computers.)
In some large companies, a server may be reserved to host the company�s Web site (although many companies let Web hosting services run the company�s Web servers).
In addition to being able to access network servers, client computers in a Windows network can access other client computers, which is a peer-to-peer method (discussed in the next section). This permits users to exchange or share data files directly. However, each computer can be set up to maintain permissions that allow access only by certain users, or to deny access to all client computers.
In fact, users can opt to log on to their own computers, rather than the network, if they wish. They won�t be able to access any network resources, but this may still be the way to log on if a server is down or if the user wants to work on software applications that are installed on the local computer without saving the data to a network server.
NOTE: The client/server paradigm described here represents a Windows Network Operating System (NOS), but you may come across a different NOS in corporate environments, such as UNIX, Linux, or NetWare. All of those systems authenticate users to maintain controls and security, but the methods vary. In addition, the client/server relationship is usually absolute, so individual computers on a client/server network cannot access each other in a peer-to-peer fashion, nor can users log on to their local computers- they must join the network. Those last two distinctions enhance the network security.
The biggest advantage of a client/server network is the level of security you can achieve, but for small networks, that advantage is usually outweighed by the cost of buying a server to maintain those controls and the technical knowledge required to set up and maintain the authentication processes. Because small networks have fewer computers and users, the need to develop security controls on a user-by-user or computer-by-computer basis is less daunting.
NOTE: In a Windows client/server network, the servers that authenticate computers and users when they log on are called domain controllers. When you log on to a client/server network, you�re logging on to a domain. Peer-to-peer networks log on to a workgroup.
Peer-To-Peer NetworksA peer-to-peer network is exactly what the term implies-everybody is equal. No servers equipped with authentication software utilities interfere with the process of logging on to a computer or to the network. All the computers on the network can communicate directly with all the other computers on the network.
You can apply security measures to resources, such as files and folders, on each computer on the network. In fact, if one user decides that his computer is totally private, he can refuse to share any files or folders on his computer. That doesn�t stop him from being able to access resources on another computer, though, if that other computer is sharing resources.
In addition, with peer-to-peer configuration, you don�t entirely lose the concept, or efficiency, of a server. For example, you can use any computer on the network as a print server without interfering with a user�s ability to work on that computer.
A print server is a computer that has a printer connected to it, and that printer is configured as a shared printer, meaning everyone on the network can use that printer. The user who works at the computer that is connected to the printer doesn�t have to share any other resources, such as files or folders, because shared resources are set up on a resource-by-resource basis.
You can also maintain data files for certain software applications on one computer, which essentially treats the computer as a server. For example, many accounting software applications work in this mode (an example is QuickBooks).
In some versions of Windows (Windows 2000 and Windows XP), you can impose security on the logon process, so that nobody can log on except an existing, recognized user who knows the password. In addition, each logged-on user can be restricted in their ability to make changes to the system�s configuration.
When you log on to a peer-to-peer network, you join a workgroup, which is a group of computers that are connected to each other.
Network AdministrationA certain amount of administrative work is involved in creating and maintaining a network. Luckily, if you�re creating a peer-to-peer network, most of the administrative tasks don�t fall into the category of �real work�; this is all more like �set it and forget it.� The fact that administrative tasks must be performed is the result of some important networking concepts and rules, which I�ll discuss in this section. This book is really about network hardware, and it won�t cover most of the software and operating system issues involved in networking. (See Appendix C for a list of books that cover these issues in detail.) However, it�s important to present an overview of what you�re going to face as you create, configure, and run your network.
Computer and Workgroup NamesEvery computer that joins a network must have a name, and the name must be unique on the network. The computers communicate with each other every time a user accesses any network resource, and the computers don�t have a way of determining that you meant this Bob, not the other Bob. One computer named Bob to a network, please. In addition, when you create a network, you must also name the group of networked computers, which is called a workgroup.
You can name a computer when you install the operating system, and if you bought a computer from a manufacturer who installed the operating system for you, it may already have a name (something really creative, descriptive, and easy to remember like DEO77495FG077MR). Some people name their first computer Brown, because that�s the family name; then when they buy another computer, they name it Brown, and perhaps even when a third computer arrives, it too is named Brown. That�s neither unusual, nor harmful, until you decide to create a network with all three computers.
When you install your network using one of the Windows networking wizards (available in Windows Me and Windows XP) or by setting configuration options manually (Windows 98SE and Windows 2000), you can name (or rename) the computer and the workgroup during the network setup process.
It�s a good idea to name a computer for the room in which it resides, so there�s no chance of duplication and so that everyone understands where the computer is located. For example, Den, Kitchen, Basement, Attic, Garage, Marysroom, and so on, are good choices.
The workgroup name can be anything you want, and each computer on your network must have the same workgroup name (the Windows networking wizards usually suggest something like MSHome or evenWorkgroup, but you can substitute another name for your workgroup).
Computer and workgroup names cannot be longer than 15 characters, and the following characters are forbidden in the name:
/ \ * , . @ space
You can change the existing name of a computer (usually a task that�s performed because of duplicate computer names) in the following way:
- In Windows XP, right-click My Computer and choose Properties from the shortcut menu. In the Properties dialog, go to the Computer Name tab. Click Change to enter a new name.
- In Windows 2000, right-click My Computer and choose Properties from the shortcut menu. In the Properties dialog, go to the Network Identification tab and click Properties to change the name.
- In Windows 98SE/Me, open Control Panel, and then open the Network applet. Go to the Identification tab and type in a new name.
User SettingsYou can configure any computer to ask a user for a logon name and a password when the operating system starts, and this approach provides several advantages to users and to your efforts to make your computers, and your network, secure.
If more than one person uses the computer, each user�s configuration preferences are loaded when that user logs on. For example, desktop icons, the listings on the Programs menu, the personal My Documents folder, and other components are all exactly the way the user wants them. This is called a user profile, and each user�s profile is stored on the computer.
For computers running Windows XP (and Windows 2000), individual logons also provide security measures that protect the computer, because you can configure the rights and privileges for each user. Users who have accounts configured for limited rights cannot perform certain functions; for instance, they may not be able to install software, and they cannot delete or modify system files. If your network is part of your business, limited accounts make a great deal of sense. Even in a home network, a limited account may prevent damage if one of your household members tends to plunge into computer tasks without enough knowledge to stay out of trouble (usually, that means one of the parents-most kids are far more sophisticated about computers than their parents are). All network administrators can identify those users who �know enough to be dangerous�; these are the users they restrict.
For computers running Windows 98/Me, the only advantage to logons and passwords is to load the user profile. There�s no security built into these operating systems. Any passing stranger can walk up to a Windows 98 computer, turn it on, and when the Logon dialog appears press ESC or click Cancel, and voila-he�s in the computer and can do anything he wants.
It�s possible to use logon names and profiles without requiring a password (a user without a password is said to have a null password), and this scenario is common in many home networks. However, for a business network, where outsiders may be on the premises, it�s a real security risk to permit users to omit passwords.
Network Connection ChoicesBefore you can do anything about your network-before you buy the hardware- you have to choose a connection system. The decision you make affects the type of equipment you buy and the installation tasks you perform. You have several choices, and once you make your decision, you�ll learn how to implement that decision by reading the appropriate chapter in this book.
But before you turn to the chapter on installation, you need to know which chapter to read, because each connection type is covered in its own chapter. To help you decide, I�ll present an overview of each of the following connection types:
- Ethernet cable
- Household telephone wires
- Household electrical wires
- Wireless RadioFrequency connections
The hardware devices you need to run these topologies are available from Linksys-the hardware manufacturer both authors of this book prefer (in fact both of us were Linksys fans before anyone ever thought of writing this book). In the chapters that follow, we�ll provide Linksys model names and numbers when appropriate.
During the discussion on connection choices, I�ll be talking about the speed at which each connection type can transfer data among computers. Network speeds are rated in Megabits Per Second (Mbps). A megabit is a million binary pulses, which doesn�t mean anything unless you can think about it in a familiar perspective. The easiest way to do that is to think about a dial-up modem. The fastest dial-up modems available transmit data at the rate of 56,000 bits per second (56 Kilobits or 56 Kbps). If you�ve ever used a dial-up modem to download a file from the Internet, and you watched the progress bar move rather swiftly as thousands, tens of thousands, hundreds of thousands, and finally millions of bytes were delivered to your computer, think about how fast a megabit must be.
Ethernet CablePresenting the pros and cons of choosing Ethernet cable as your network topology is easy: it�s almost all �pro� and very little �con.� Ethernet is the cable of choice for any network, because it�s fast, accurate, and almost always trouble-free; this is the connection type you find in corporate networks. (Instructions for cabling with Ethernet are in Chapter 4.) Absent some good reason, Ethernet is the topology to use for your network. However, I do admit that good reasons exist to opt for another topology, including the following common ones:
- You rent and the landlord doesn�t want you to drill holes to bring cable through the walls.
- Your computers are located in places that are difficult to reach with a physical cable.
- The notion of running cable through your home or office strikes you as �too much work.�
- The person in your household who has the final word on d�cor and aesthetic decisions says �I don�t want to see cable snaking out of the walls and into the room.�
- Your network is made up of portable computers and you never know where any individual computer is going to be used.
Ethernet can transfer data across the network at 100Mbps, as long as the NIC and the hub/switch can support that speed. Some older NICs and hubs can send data at only 10Mbps, but Ethernet NICs and hubs/switches can automatically sense the speed of the Ethernet devices on the network and drop or raise the speed to match the device�s capabilities (this feature is called autosensing).
Today, Ethernet cable is purchased in the form of 100BaseT cable, which is also called twisted pair cable and category 5 UTP cable. The �100� in the name refers to the speed at which it can transmit data (100Mbps); sometimes 100BaseT Ethernet is called fast Ethernet. Older Ethernet cable that can transmit data at 10Mbps is called 10BaseT.
The term �base� is short for �baseband signaling,� which means that only Ethernet signals are carried on the wires. This differs from telephone wire, which can handle multiple types of signals. (Your telephone service uses only a portion of the wires, and you can use other portions for other technologies, including computer networks, which is discussed in Chapter 6.)
The term �T� means twisted-pair, which is a description of the way the wires are twisted and paired through the cable. (Other wiring types, such as those for fiber-optic wires, which is called 100BaseF, also exist.)
Ethernet cable looks like telephone wire, and the connectors (also called jacks, even though the jack is the wall outlet to definition purists) look like the connectors on your telephone cable. However, they�re not the same. The wires are twisted differently and the arrangement of wires in the connector is different (a telephone jack is an RJ-11, an Ethernet jack is an RJ-45).
Household Telephone WiresHousehold telephone wires are an easy way to connect computers into a LAN, and the technology has come a long way since it was first introduced several years ago. (See Chapter 6 for instructions on creating a telephone line network.) Telephone network cable uses the wires in your telephone cable that voice communication doesn�t use, so your telephone lines are still available for normal household telephone use (including a modem and a fax). The available range of frequencies within cable is called bandwidth, and you can use the bandwidth your phone services don�t use to create a computer network.
There are pros and cons for using your household telephone wires, and they balance rather evenly. On the pro side, the only hardware you need is a NIC and length of plain telephone wire for each computer. Each computer is plugged into a regular telephone wall jack, eliminating the need to buy or install Ethernet cable. You can use your telephone wall jack for both a telephone and a network connection at the same time by installing a splitter (really called a modular duplex jack) which is a gadget you can buy at your local supermarket. The splitter has a male jack that goes into the wall jack, and it has two female jacks on the outside. Plug a telephone cable into the Phoneline network card into one jack and your telephone cable into the other jack.
On the con side, not every room in your house may have a telephone jack, so you must plan the location of your computers around jack availability. Also, if you have more than one telephone number in your house, all the computers on your network must be connected through the same number. Computers can�t communicate across different telephone numbers-but then, neither can you. If people are talking on line 1, you can�t pick up a telephone connected to line 2 and join their conversation (or eavesdrop).
If your business uses a PBX telephone system, you can�t use the jacks for your network, because the wiring is different from regular telephone jacks. PBX jacks are designed to deliver all the services that come with your phone system. In addition, if you have a DSL device for your Internet connection, you�ll face some interoperability problems, but the workarounds aren�t very complicated (and they�re covered in Chapter 6).
Another �con� for using telephone lines is that the maximum distance between any two computers is about 1000 feet-but unless you live in Windsor Castle, that shouldn�t be a problem.
Telephone networks operate at 10 Mbps (although faster devices should be available soon), which is not as fast as today�s fast Ethernet (100 Mbps) but matches the speed at which corporate networks operated until the recent introduction of fast Ethernet. In fact, plenty of networked computers still operate at 10 Mbps Ethernet, because the company�s Information Technology (IT) professionals decided they didn�t have a compelling reason to update the hardware. A connection speed of 10 Mbps is more than sufficient for almost any data transfer tasks. Actually, I don�t know whether to call the speed rating of telephone lines a pro or a con.
Household Electrical WiresYou can run a network on the unused bandwidth within your household electrical lines without interfering with any of the work those lines perform to provide power to electrical appliances. In addition, the electrical appliances don�t interfere with the transmission of data among the computers on your network. Best of all, since you�re not using the part of the bandwidth that supplies power, the network connections don�t raise your electric bill.
The NIC in each computer connects that computer to the nearest electrical outlet, and once all the NICs are installed and plugged in, you have a network. The speed of data exchange is about 14Mbps, which is more than fast enough for anything I ever wanted to do on a network. One obvious advantage to this connection type is the fact that at least one electrical outlet exists in every room, so you can put your computers anywhere you wish. Turn to Chapter 7 for instructions on installing your network over your household electrical wires.
Wireless NetworkingWireless networking uses RF (RadioFrequency) signals to communicate among the computers. The advances in wireless networking have been incredibly rapid in recent years, and the technology is becoming more and more popular as each new set of standards increases its power. Chapter 5 provides all the information you need to set up a wireless network.
The NIC has a transceiver (named for the fact that it both sends and receives data) and an antenna, and the data communication rate is about 11 Mbps for 802.11b and 54 to 72 Mbps for 802.11a (discussed more in Chapter 5). The interference problems that plagued early versions of wireless networking have disappeared, and you no longer have to worry that copying a file might open your garage door. In addition, robust security functions are built into the Linksys wireless devices, so you don�t have to worry that a neighbor who also uses wireless technology could intrude on your network and gain access to your data.
The downside of wireless communication is that the signal can be interrupted by metal. So if you store your computer under a metal desk, you�ll also have communication problems. This is also true if the walls between computers have a lot of metal pipes inside (typical of older homes, before PVC became the standard material for drain pipes). There�s also a distance maximum between computers (about 150 feet), but Linksys offers devices that extend the signal.