by Theresa Hadden
What is a network and why do you need one? A network consists of the components necessary for two or more computers to communicate with one another. These are just some of the advantages of networking:
A networkconsists of two types of components: hardware and software. The hardware components consist of the various pieces of equipment that connect the computers in the network. At a minimum, the required hardware consists of cables and adapter cards. Windows 98 integrates well into a broad variety of network types.
The software required to communicate across a network includes the network operating system, a network client, and one or more network transport protocols. Windows 98 functions as both a networking operating system and a desktop operating system.
NOTE: A network client allows your computer to communicate with another computer based on the type of network operating system it is using. A protocol can be thought of as the language that is spoken across your network. If two computers use different protocols, they cannot communicate. Even if both computers use the same protocol, you need to make sure they are both properly configured to communicate with each other.[dagger]
Networks are organized in one of two ways: peer-to-peer or server-based. In a peer-to-peer network, each computer acts as both a server and a client. All the information is stored on the various computers. Any computer in the network can function as a server, providing access to such resources as files stored on its local hard drive and peripherals attached to it (such as printers, fax modems, scanners, or CD-ROM drives). Every computer in the network can also function as a client to access resources on the other computers in the network.
Each computer in a peer-to-peer network can share its resources without the need for centralized administration of these resources. Each user in a peer-to-peer network is a network administrator. This alleviates the need for one person to be responsible for various network administrative tasks.
NOTE: The main advantage of a peer-to-peer network is that it doesn't require a dedicated network administrator. The main disadvantage of a peer-to-peer network is that it usually lacks a network administrator who can make sure all the network components are running smoothly. For this reason, peer-to-peer networks are suitable only for small networks.[dagger]
Windows 98's built-in networking capabilities make it an excellent choice for implementing a peer-to-peer network. As an operating system, it contains all the elements you need to allow access to local resources or to access resources located on other computers in the network. Its point-and-click interface enables you to browse the network to locate and access available resources. The same interface makes the sharing of resources extremely easy for the user even if he or she is not technically adept at network administration.
You can change a peer-to-peer network to a server-based network by adding a specialized network server. Network servers are typically computers running either the Windows NT Server operating system or a NetWare server. Network servers usually are not out in the open on people's desks; instead they are often locked away in closets or specially equipped "server rooms."
The following list outlines the basic requirements for a server-based network:
In a server-based network, the burden of controlling network administrative tasks is moved from the individual users to one or more persons who are more technically proficient in network administration. This central administrative role allows for more efficient management of the network.
The built-in network components of Windows 98 enable it to work very well in server-based networks by providing the software necessary for user authentication by the network server.
When you're planning your network, you need to consider several things. Here are some of the more important decisions you need to make before installing your network:
Use the answers to these questions to plan your network. Items that you will need to plan for include its bus topology, what equipment you will use to connect the computers on your network, and what security method you will use. Whether or not you will be creating a peer-to-peer network or a server-based network will influence your choice of operating system.
After determining the size of your network, you will need to decide on the layout and hardware components needed for installing the network. As an example, we will examine a local area network (LAN) situated in a single building because the coverage of larger networks is beyond the scope of this chapter.
When planning your network, you need to consider which topology, or design, is the most appropriate. This is especially important because the largest single cost when installing a new network can be the cost of installing the cable that connects all the computers in the network.
The most basic and easiest type of network topology to install is a bus network, which consists of two or more computers connected by coaxial cabling. Because of their ease of installation, coaxial cable networks have a very low initial cost.
NOTE: For many network installations, a major cost is actually the labor charges incurred for the physical installation of the cables that connect the various computers to one another. Pulling cables through ceilings and walls is hard work, and it always seems to take much more time than most people expect.[dagger]
Figure 42.1 illustrates a bus topology. Each computer is connected by coaxial cable using a connector called a T-piece. The coaxial cabling must be properly terminated at each end of the network with 50-ohm terminators.
FIG. 42.1 A diagram of a bus topology network.
Coaxial cable will support up to 30 computers as long as the total length of the network is no greater than 185 meters. Bus topology lends itself well to installation in areas where running cabling through the ceiling or walls is difficult or impossible. In addition, it is easy to expand a bus network by adding another length of cable and a T-piece.
Although bus networks are easy to install, they can be difficult to troubleshoot if one of the network components malfunctions. A damaged or defective cable or T-piece can disrupt the entire network. Turning off your workstation, however, will not disrupt the function of the network. If you do not want to learn the ins and outs of network troubleshooting, this type of network might end up costing you more in the long run.
TIP: If you have trouble with your network, segment your network by disconnecting the cabling in the middle. Terminate each side of the network to determine which half is not functioning. Then continue subdividing the half that does not work until you have identified the problem component.
The next type of network to consider is a star topology, in which a cable runs from each computer to a central piece of hardware called a hub. The capacity of a hub ranges from 4 to 24 computers. Figure 42.2 shows a small star topology.
FIG. 42.2 A cable runs from each computer to the hub to create a star topology.
The twisted-pair wire that is used for this topology resembles telephone wire except that it uses four pairs of wires whereas telephone wire uses only two pairs. The RJ-45 connectors that are used with twisted-pair wires look like a slightly larger version of the connectors that are used on telephone wire. In a star topology, no computer can be more than 100 meters from the hub.
TIP: Twisted-pair wire is categorized by its speed of transmission. If you plan to expand your network, you might want to invest in Category 5 cable, which supports speeds up to 100Mbps. This will prevent you from having to rewire your network later.
Star topology has the advantage of being easy to troubleshoot. A single piece of defective cabling will affect only the computer that it connects to the hub. If the hub malfunctions, however, none of the computers attached to that hub will be able to communicate on the network. This is referred to as a single point of failure.
Initially, this type of network is more expensive than a bus topology because it requires the purchase of additional equipment such as hubs. Expansion of a star topology may require the purchase of an additional hub if no more connections are available on the existing hub. As your network grows, the wiring can become untidy and unwieldy, especially if you are unable to run the wiring through the walls.
TIP: After you decide on the topology for your network, you might want to consider having the cable run by an expert in the field. This initial investment will save you considerable expense later.
Use the following list as a quick reference to the characteristics of different cable types:
| Type | Code | Cost | Total Distance | Number of Nodes per Segment |
| Twisted Pair |
10BaseT | Least $ | 100 meters | 1 |
| (UTP) | ||||
| Twisted | 10BaseT | 100 meters | 1 | |
| Pair | ||||
| (STP) | ||||
| Thin Coax | 10Base2 | 185 meters | 30 | |
| Thick Coax | 10Base5 | 500 meters | 100 | |
| Fiber | Most $ | 2,000 meters |
After the cabling has been installed, you will need a network adapter card for each computer that will be networked. A network adapter card acts as an interface between your computer and the network. Each card is specific for the bus type of your computer and the type of cabling you have. See the section "Configuring Your Network," later in this chapter, for more details on selecting your network adapter card.
The network operating system controls how communication takes place. Each computer must have a client operating system installed that allows communication over the network.
Windows 98 is an excellent choice because it serves as both the network operating system and the client operating system. In addition, its graphical user interface allows for ease of use and increased productivity. It supports a wide range of applications and has native support for mobile computing.
Network transport protocols format data for transmission over a network. Although more than one protocol may be installed on a computer, multiple protocols increase memory demands on computers and can produce excessive network traffic. It is more efficient to use only the protocol or protocols that are necessary to communicate with other computers on your network.
Windows 98 supports three primary transport protocols: NetBEUI, TCP/IP and IPX/SPX. Each of these protocols has advantages and disadvantages:
NOTE: A router is a device used to connect different networks or network segments. Networks can be subdivided to reduce traffic, improve performance, or increase the distance over which the computers in the network can communicate.[dagger]
The easiest and most common small network consists of two computers connected to each other. This setup is commonly used in small offices or home networks for the purpose of sharing a printer and/or sharing files. Three basic configurations can be used to achieve this end:
The first two configurations require that a network adapter card be installed in each computer. The third option connects two computers using a cable attached to a serial or parallel port instead of a network card. The connection is made using either a null modem or parallel cable (not a printer cable).
TIP: Consider using a network connection to transfer information between your laptop and your desktop computer at home. Transferring information between two computers via a network is much easier and faster than carrying floppy disks around in what is commonly referred to as a "sneaker net."
The selection of a network adapter card depends on the type of computer it will be installed in, the type of cabling that will be used, and whether a driver is available for your operating system.
Every PC has one or more types of adapter slots, such as PCI, ISA, EISA, VESA local bus, Microchannel, and PC Card (PCMCIA). The network card you select must be compatible with an available slot in your computer. Refer to your PC's documentation to identify the correct type of adapter for your computer.
TIP: Before installing or upgrading any adapter card in your computer (such as a modem or a video card), you need to know what type of slot is available in order to purchase the correct type of adapter card. If your computer is a Pentium (or newer), it probably has at least one PCI-compatible card slot. PCI cards provide much better performance and are easier to configure than the older ISA standard.
TIP: If you will be purchasing a network card for what will be a heavily used file server, you will gain overall productivity if you use a higher performance card. Check for performance-oriented capabilities such as bus mastering.
Network cards generally allow for connection to a network in one or more ways, depending on the type of cable connector on the card. If the network adapter card has more than one type of connector present, it is often referred to as a combo or combination card and generally is more expensive.
In addition to computer bus and cabling type, other issues include the type of memory the network card uses. Consult the product specifications when selecting the appropriate card for your use.
If you are using coaxial cable, the network card you select must have a BNC adapter. This type of adapter looks like a short metal barrel. The coaxial cable is attached to a T-piece, which is then attached to the network adapter card.
CAUTION: Never attach a coaxial cable directly to the network card. This practice will result in a nonfunctioning network because you will be unable to attach the terminators that are required for the proper function of a coaxial cable-based network.
If you are using twisted-pair cable, select a network adapter card that has an RJ-45 connector, which looks like a larger version of the connector you use to insert your telephone wire into your telephone.
In addition to a network adapter card, your also must have software that allows your computer to communicate with the card. The software you need consists of two components:
TIP: Windows 98 uses a Microsoft standard called NDIS to communicate with the network adapter card driver. This standard allows for the use of multiple protocols with a wide variety of network adapter cards.
Before you purchase your network adapter card, make sure that a driver appropriate for your operating system is readily available. Windows 98 includes drivers for most common cards. Check the most recent Hardware Compatibility List (HCL) available with the operating system or from Microsoft to ensure that you have a compatible card.
Connecting Using Coaxial Cable
After the network adapter cards have been installed in each computer, you will need to attach a piece of coaxial cable. This cable must be either RG58U or RG58A/U. Although the type of cable you use for your television looks the same, it will not work. The cabling is connected using a T-piece connector. Figure 42.3 shows a T-piece. One side of the crossbar is attached to the coaxial cable, and the other is attached to either another cable or to a terminator with a 50-ohm rating. The connector is then attached to the network card. This process is repeated for each computer.
CAUTION: An improperly terminated network or the wrong type of cable can result in network failure. Or even worse, it can lead to intermittent network problems that are extremely difficult to detect and fix.
Connecting Using Twisted-Pair Cabling
Although twisted-pair cabling is ordinarily used with a hub to connect computers, you can connect two computers using a special type of twisted-pair cable called a crossover cable. To bypass the need to use a hub, one receive pair and one send pair are crossed at one end of the cable.
Crossover Twisted-Pair CableNormally, when attaching the RJ45 connector to a twisted-pair cable, each of the eight wires must be connected to the same groove on each connector. That is, the wire attached to the first groove on one connector should be attached to the first groove on the other cable.
With a crossover cable, the send and receive wires need to be reversed at one end. Figure 42.3 illustrates how to reverse the wires to create a cable that will connect two computers without requiring the use of a hub.
A crossover cable will connect only two computers, however. If you need to connect more than two computers, you will need to use a hub.
After installing the network adapter card, you just plug the crossover cable into each of the network cards in the same manner that you plug in your telephone. If you have wired the crossover cable correctly, the two computers will be able to communicate.
TIP: Instead of making your own cable, you might want to purchase it from a cable supply company. Also, to prevent possible confusion, try to purchase your crossover cables in a different color cable than the cables you use to connect computers to your hubs.
FIG. 42.3 This diagram illustrates how to cross over the wires to create a crossover cable. The numbers on the left refer to the grooves on one connector, and the numbers on the right refer to the grooves on the connector at the other end of the cable.
Configuring Your Network
When the physical connections are made, use the Add New Hardware Wizard in Control Panel to detect and install your network adapter card. After you run the hardware wizard, verify that the correct network components are installed. You can examine the network properties either by right-clicking on the Network Neighborhood icon or by opening the Network Applet from the Control Panel. The following three items need to be installed via the Properties tab of the Network Neighborhood: Adapter. This is the network card adapter.
Protocol. This is the language that will be used to communicate on your network. Either Client for Microsoft Networks or Client for NetWare Networks. These components are sometimes referred to as redirectors; they allow you to communicate with other computers on your network. The adapter should already be listed as the one that was installed using the Add New Hardware Wizard. Thanks to Windows 98 Plug and Play capabilities, no further configuration should be necessary.
TIP: Use the Device Manager to verify that the network adapter card is functioning properly and that no resource conflicts are present.
The best protocol to install when connecting only two computers is NetBEUI. Its small overhead and the lack of configuration requirements make it the ideal choice. All you need to do is install it.
To install this protocol, open the Network Properties dialog box, select Add, and then select Protocol. The Select Network Protocol dialog box appears (see Figure 42.4). Select Microsoft on the left side of the dialog box, and then select NetBEUI on the right side.
FIG. 42.4 Select Microsoft from the left side of the dialog box. Select NetBEUI from the list of available protocols on the right side.
To add the Client for Microsoft Networks, click the Add button on the Configuration tab. Select Client, and then click on the Add button. In the Select Network Client dialog box, select Microsoft on the left side and select Client for Microsoft Networks on the right side (see Figure 42.5).
FIG. 42.5 Select Microsoft on the left side and select Client for Microsoft Networks on the right side.
When you return to the Configuration tab of the Network Properties dialog box, be sure that Client for Microsoft Networks is selected in the Primary Network Logon box, as shown in Figure 42.6.
Next, click on File and Print Sharing. Select the appropriate check boxes to allow access to your files and/or to allow others to print to your printer. Then click OK to return to the Configuration tab.
Select the Identification tab to see the options shown in Figure 42.7. Enter a name for your computer and your workgroup, and then click OK.
FIG. 42.6 Notice that Client for Microsoft Networks appears in the Primary Network Logon box.
FIG. 42.7 Here the computer name is WhiteStar, and the workgroup name is STARTREK.
NOTE: The name you select for your computer must be a unique name; it is also referred to as the computer's NetBIOS name. This name can contain up to 15 characters but cannot contain spaces or any of these characters:/ \ * , . " @You should enter the same workgroup name on each computer. Although you will still be able to communicate if the workgroup names are different, entering the same name on both computers means that both computers will be listed as members of the same workgroup when they're viewed by someone browsing the Network Neighborhood.[dagger]
You can avoid installing a network adapter card by connecting two computers directly using a null modem (serial) or parallel cable attached to a serial (COM) or parallel port on each computer.
To install the Direct Cable Connection capability, run the Add/Remove Programs applet in the Control Panel and click on the Windows Setup tab. In the Components list, select Communications, and then click the Details button. Select Direct Cable Connection and click OK.
After you've installed Direct Cable Connection and connected the two computers with either a serial or parallel cable, open the Start menu and choose Programs, Accessories, Direct Cable Connection. The Direct Cable Connection Wizard runs when you open Direct Cable Connection for the first time. One computer should be designated as a host and the other as a guest, as shown in Figure 42.8.
FIG. 42.8 Configure Direct Cable Connection with the wizard.
With Direct Cable Connection, you can gain access to shared folders on another computer, even when your computer is not wired to a network. If the other computer is connected to a network, you can also gain access to that network.
TIP: Direct Cable Connection is intended for only light use because it is much slower than a true network connection. For sustained communications between two computers, you would be much better off to buy the appropriate hardware.
A workgroup is an organizational unit. It is a loose association of computers wherein each computer tracks and controls access to local resources. These resources may be folders located on the local hard drive or hardware devices attached directly to the computer (such as a printer). Each individual computer in the workgroup controls how access to these resources occurs.
If your network is small--fewer than 10 computers--a workgroup is an easy way to configure your network and provide access to various resources. It does not require a single individual to be designated as the administrator. Rather, each user is the administrator for his own computer.
The disadvantages associated with a workgroup are related to its advantages. Because each user is his own administrator, there is no centralized control of users or resources. As your network grows, the large number of users can result in a disorganized network. Because there is no centralized location for shared documents and applications, it can become difficult for a user to locate needed information. This can result in lost productivity or the occurrence of duplicate versions of the same document.
Printing is one of the most common tasks performed on a network, and shared printing can lead to problems. Because the computer that is attached to the printer is acting as a print server, the additional load can cause degradation in performance. This can be a significant problem affecting the user working on that computer.
TIP: To improve performance of a computer that is extensively used as a file server, use the System option in Control Panel and set the typical role of the computer to Network Server. If the computer is also functioning as a print server, make sure that adequate disk space is available for spooling.
When files and folders are shared, a password is associated with the resource and the level of access that is allowed. Once file sharing has been enabled, various levels of access to a resource can be defined based on passwords. These levels are Read and Full control.
TIP: File sharing is enabled through the File and Print Sharing button on the properties sheet for the Network Neighborhood. You can elect to share files, allow others to print to your printer, or both. You must also select Share Level Access Control on the Access Control tab.
To share a folder, right-click on it in Windows Explorer and select Properties. On the Sharing tab, select Share As. Each shared folder must have a name to identify it. A suggested name will be displayed, but you can enter whatever name you want to use as long as it contains no more than 12 characters. The name that you enter here will be displayed to any user browsing your computer across the network. Figure 42.9 shows a shared folder.
TIP: Ending the share name with a dollar sign ($) creates a hidden share. Hidden shares cannot be seen by users who are browsing your computer across the network and can be accessed only by users who know the share's name.
You can also enter a comment that will be displayed to users as they browse your computer across the network. Use this field to identify the information that is contained in the share, which might make it easier for users to find needed information.
When you name your workgroup on the Identification tab of the Network Properties dialog box, one of two things happens. You either join an existing workgroup or create a new one. If the name you enter is an existing workgroup, the computer you are configuring becomes a member of that workgroup.
FIG. 42.9 The level of access that is granted may be dependent upon what password the user enters to access the resource.
However, if you enter a name that is not the name of an existing workgroup, you create a new workgroup, and that computer becomes its first member. You can then add additional computers to your workgroup simply by entering that workgroup's name on the Identification tab of the Network Properties dialog box of the computer you want to add to the workgroup.
One of the biggest problems in a peer-to-peer network that uses share level access is the number of passwords needed to control access to resources. The more resources that are available to the users, the more passwords that have to be remembered.
If a user has to keep track of multiple passwords, he might forget them or he might write them down. Either one prevents the security policy from doing what it is intended to do--which is to allow access to those you want to have access and deny access to all others.
For that reason, Windows 98 offers a password caching feature that is installed by default. This feature stores passwords in a special file. When you access a password-protected resource for the first time, make sure that Save This Password in Your Password List option is checked. Your password will be encrypted and then stored in a .PWL file in the Windows folder.
CAUTION: If you delete .PWL files, you will lose all the passwords you have stored. The next time you access a password-protected resource, you will have to retype the password.
If password caching is enabled, you can view the resources listed in the Password List File (.PWL) with the Password List Editor. The editor displays only the .PWL of the user who is logged on. You can disable password caching by using the System Policy Editor.
System Policy EditorSystem policies provide a powerful tool for controlling how users interact with their computers and the network. One way to use system policies is to enforce specific password guidelines. These include defining the minimum password length and requiring the use of an alphanumeric logon password. System policies can also be used to disable password caching so that the user must type the password every time she attempts to access a shared resource.
You can use the system policy templates to define your desired policies. You can also create a new template that contains those items you want to control. Besides establishing password requirements, you can also perform the following measures:
To use System Policies, you must install the System Policy Editor and the templates.
Here are some additional policies that you can implement to improve password security:
You can improve the security of your network by limiting the total number of shared folders. By keeping those files that require shared access located in a limited number of folders, you decrease the number of passwords a user has to remember.
Disable file and print sharing on those computers that do not contain files requiring sharing. This action not only improves security, it also enhances the performance of the computer.
A Windows 98 network is easy to implement and maintain. This native functionality results in increased efficiency. It also saves you money by allowing all your users to share hardware devices such as printers.
© Copyright, Macmillan Computer Publishing. All rights reserved.
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}