Craig Zacker - Author, Editor, Networker
Building Windows 98 Networks
Chapter 14. Network Management Tools & Tactics

A small Windows 98 workgroup network does not require a large amount of maintenance, but there are tools included with the operating system and techniques that you can use to make network administration tasks easier. For the most part, network administration on this scale is a matter of performing maintenance tasks from a remote location, so that you don't have to travel to each workstation and take over the machine. With Windows 98, you can sit at your own workstation and perform the following tasks for any system on the network:

  • Test the operational status of the TCP/IP protocol stack
  • Create desktop shortcuts
  • Manage the Start Menu
  • Monitor the files being shared
  • See who is accessing files
  • Create, delete, and manage shares

In addition to remote administration tasks, you can manually configure your network workstations to launch system maintenance utilities and other programs at regular intervals, and to support multiple users on one machine with their own individual desktops, Start Menus, and registry settings. This chapter will also cover the essential task of backing up your important data, and how Windows 98 can use one backup drive to protect all of the systems on your network.

Using PING

PING is a simple program included with all TCP/IP protocol stacks (irrespective of operating system) that you can use to determine if the network interface on a network workstation is functioning properly. The program works by sending messages called echo requests to a system that you specify, either on the local network or the Internet. If the requests reach the destination, the receiving system returns an echo reply for each request. If you receive no replies, then there is something wrong with either the sending or the receiving system.

Note: PING is a TCP/IP utility, that functions only on TCP/IP systems. If your Windows 98 machines are configured to use only the NetBEUI and/or IPX protocols, PING is not installed on the systems.

PING is a command line utility that you run from an MS-DOS window, using the following syntax:

PING address

where address is either the IP address of a system on the network, a NetBIOS name, or a DNS name. The program file, Ping.exe, is installed to the C:\Windows directory with the TCP/IP protocol stack, and is executable from any DOS prompt. When you specify a name instead of an address, the program resolves the name into the system's IP address and displays it as it transmits the echo request messages. Thus, PING is also a useful means of determining the IP address of a system.

The screen output produced by a PING command appears as follows:

c:\>ping cz3

Pinging cz3 [192.168.2.3] with 32 bytes of data:

Reply from 192.168.2.3: bytes=32 time=1ms TTL=128

Reply from 192.168.2.3: bytes=32 time<10ms TTL=128

Reply from 192.168.2.3: bytes=32 time=1ms TTL=128

Reply from 192.168.2.3: bytes=32 time<10ms TTL=128

Ping statistics for 192.168.2.3:

Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),

Approximate round trip times in milli-seconds:

Minimum = 0ms, Maximum = 1ms, Average = 0ms

The program displays a result line for each of the four echo request messages it sends by default, specifying the IP address of the recipient, the number of bytes of data transmitted in each message, the amount of time elapsed between the transmission of the request and the receipt of the reply, and the target system's time-to-live, or TTL. (The time-to-live is the number of routers that a packet can pass through before it is discarded; it has no application on a small network.) Using additional command line parameters (displayed by the PING /? command), you can modify the number and size of the echo requests transmitted by a single PING command, as well as other operational characteristics.

PING has other diagnostic uses apart from simply determining if a system is up and running. If you can successfully ping a system using its IP address, but pings sent to the system's name fail, you know that there is a malfunction occurring in the name resolution process. When you're trying to contact an Internet site, this indicates that there is a problem either with your workstation's DNS server configuration or with the DNS server itself. If you can ping systems on the local network successfully, but not systems on the Internet, then you know that there is either a problem with your workstation's Default Gateway setting, or with the connection to the Internet.

Note: Sending a PING command to a system's loopback address (127.0.0.1) tests the operability of the TCP/IP protocol stack, but is not an adequate test of the network interface. This is because traffic sent to the loopback address travels down the protocol stack only as far as the network transport layer and is redirected back up without ever leaving the computer through the network interface.

Accessing Remote System Directories

Assuming that you have shared the drive on your workstations that contains the Windows directory (C:\Windows, by default), you can access the system directories there and modify the operating environment by adding icons to the desktop and altering the Start Menu.

Creating Desktop Shortcuts

The Windows 98 desktop is a composite of layers that display icons of different types. Icons like My Computer and Network Neighborhood are system icons that cannot be removed by normal means (although the TweakUI applet, included with Windows 98, can remove them). Other icons are standard shortcuts, and just as you can add shortcuts to the desktop while working at a computer, you can add them from a remote location using a shared drive.

The Windows 98 Explorer uses the Desktop object as the root of the system's directory tree (as shown in Figure 14-1). You can place shortcuts into this object in order to make them appear on the system's desktop. However, you will also find a subdirectory called \Desktop in the C:\Windows directory that you can use for the same function.

Figure 14-1. The Desktop object appears at the root of the directory tree, as represented in the Windows 98 Explorer

The obvious difference about this directory is that it can appear on a remote system as part of a share, while the system's Desktop root object does not. This is because C:\Windows\Desktop is part of the standard structure of the C:\ drive. If you create a shortcut in this directory from another system on the LAN, the icon will appear on the workstation's desktop within a few seconds, just as if you created the shortcut locally. You can use this capability to put commonly used programs on the desktops of all of your network workstations without having to travel from machine to machine.

To create a shortcut, you click on a file or directory in Windows 98 Explorer with the secondary mouse button and drag it to another directory. The system pops up a context menu from which you select Create Shortcut Here. You can also use the Copy and Paste Shortcut commands in the Edit menu to produce the same effect. When you create a shortcut, the path that Windows 98 uses to the original file is relative. For example, if you browse to a workstation share in the Network Neighborhood and create a shortcut in the \Desktop directory that points to the Notepad.exe program in the \Windows directory, the shortcut will use a UNC name in the path to the program, as shown in Figure 14-2. This shortcut will always work, because a UNC path name is absolute, and does not change unless you modify the computer’s NetBIOS name or the share name. However, processing a UNC path, even on a local drive, is a more complex task for the system than if you used a local drive letter path.

Figure 14-2. Shortcuts created from a share in the Network Neighborhood use UNC path names to the target file

If you map a drive letter to the share and create the shortcut there, the shortcut will use a drive letter in the program's path name. The drive letter that appears in the path is that of the mapping on your workstation (where you're creating the share), not the workstation you're actually modifying. Thus, if you map the J: drive on your system to the shared C: drive on another workstation, and create a shortcut, the path will read J:\Windows\Notepad.exe.

This may seem like a problem, since the J: drive on the user's workstation, if it exists at all, will not point to the drive where the local \Windows directory is located. However, Windows 98 shortcuts are intelligent; if the file referenced by the shortcut does not exist in the specified location, the program will search for it. In the case of the J:\Windows\Notepad.exe shortcut, the first time that the user executes it, the system will find the Notepad.exe file on the local drive. If it does not exist on the J: drive, the system will launch the program, and modify the shortcut to point to the file on the C: drive. The only time a problem arises is when the file Notepad.exe actually does exist in the J:\Windows drive, as mapped on the user’s workstation. Remember, the J: drive on that workstation will be different from yours, so the user may end up running the Notepad.exe program from somebody else’s system on the network. This is usually not a serious problem, because the program will run (as long as it is the same file, and not another program that uses the same executable file name), but it will take longer to access the file on another system and produce needless additional traffic on your network.

To be sure that the shortcuts on remote systems launch the local copies of the files they reference, it is a good idea to make sure that the path names point to the correct path, using the local drive letter. There are two ways to ensure that this is the case. The first is to create the shortcut using the mapped drive letter on your system, then open the Properties dialog box for the shortcut and change the drive letter to that of the local drive. Thus, if you create a shortcut pointing to the J:\Windows\Notepad.exe file, you would simply change the J: to C: in order to make sure that the program runs locally. There is an easier way to do this, however, as long as the directory structure on your drive is identical to that on the remote system. If you create a shortcut to the C:\Windows\Notepad.exe file on your system and place it in the \Desktop directory on the remote system, the path will still point to the C: drive, but it will be the other user’s C: drive, not yours. Even if the target file is not in the same directory as on your system, the other workstation will search for it on the local drive and modify the path to the file accordingly.

Modifying the Start Menu

Windows 98 uses shortcuts for several different purposes aside from creating desktop icons. The Start Menu is also composed of shortcuts, and you can modify its contents just as you would the icons on the desktop. The \Windows\Start Menu directory on a Windows 98 system is where the folders and shortcuts that comprise the Start Menu are stored, by default. You can create new folders in this directory, that will appear as program groups in the Start Menu. You can also create shortcuts, that will appear as items in the groups within those folders.

The shortcuts in the Start Menu function just like those on the desktop, and are subject to the same path problems when you create them from a remote system. However, as with the desktop, you can build the Start Menu on another workstation by copying your own \Start Menu directory to the same directory on the other system, as long as you have the same applications installed to the same directories.

The ability to perform these types of remote network administration tasks is one of the main reasons why it is a good idea to create a standard workstation configuration and use it on all of your networked computers. Having the same applications installed in the same places on every machine makes it easier for the administrator to make changes.

Note: When you create user profiles on a Windows 98 system, each user has a separate set of system subdirectories, stored in the \Windows\Profiles directory, which contain individualized desktop and Start Menu configurations. See "Creating User Profiles," later in this chapter, for more information on remote administration when the user profiles feature is enabled.

Modifying the Send To Menu

The Send To menu is a powerful Windows 98 feature that many people do not realize can be customized. When you click on a file or directory in the Windows 98 Explorer with the secondary mouse button, the context menu that appears contains a submenu labeled Send To, which in turn contains a list of program shortcuts (see Figure 14-3). By selecting an item from the Send To menu, you feed the highlighted file name to the selected application, folder, or storage device. For example, highlighting a file and selecting 3½ Floppy from the Send To menu copies the file to the floppy drive.

Figure 14-3. The Send To menu enables you to send a file to a specific application or resource

As with the desktop and the Start Menu, the Send To menu exists as a subdirectory of C:\Windows, called \Send To. You can create shortcuts in this directory on remote systems, just as you can create shortcuts in the \Desktopand \Start Menu directories. If, for example, you create a shortcut to the Notepad.exe program in a workstations Send To menu, the user on that workstation will be able to open a file with Notepad by highlighting it in Explorer and selecting the shortcut from the Send To menu. You can also create shortcuts to directories in the \Send To directory, that will send a highlighted file to that directory when selected. However, your users must be aware that subdirectories in the Send To menu are subject to the same copying and moving rules as the Windows 98 Explorer. A file sent to a subdirectory on the same drive using the Send To menu will be moved; if the directory specified by the shortcut is on a different drive, it will be copied.

Notepad is a good choice of program to place in the Send To menu, as it enables users to open all manner of text files with extensions that may not already be associated with it. Another good choice is a graphics viewer that can quickly display image files of various formats. Note that while it is possible to select multiple files and send them to a shortcut, not every application supports the opening of multiple files from the command line. Be sure to test any application shortcuts before you add then to the Send To menus on your network workstations.

Modifying Fonts and Favorites

There are two other subdirectories in C:\Windows that you can use to administer your network workstations remotely. The \Fonts directory is where Windows stores the files that it uses to create TrueType fonts in all applications. True Type font files have a .ttf extension, and normally you install them on your Windows 98 machine from the Fonts Control Panel. However, you can also install new fonts on a Windows 98 machine by simply copying the .ttf files to the \Fonts directory. This means that you can do this from a remote system as easily as from the local one.

This capability is particularly useful if you are running an intranet web server on your network and you want to use a specific font in your web pages that your users do not have installed. Rather than provide them with the font file and instructions on how to install it, you can install it for them without having to leave your desk.

The \Windows\Favorites directory is where the Internet Explorer web browser stores the items that appear in its Favorites menu. The Favorites directory typically contains Internet shortcuts, which point to web or FTP sites on the Internet, but it can contain standard Windows 98 file system shortcuts as well. From a remote system, you can create standard shortcuts in a computer's \Favorites directory in the usual manner. For Internet shortcuts, however, you must create them on your local system first using the Add to Favorites command in Internet Explorer, and then copy them to the \Favorites directory on the remote system.

Using this technique, you can add useful web pages to the Favorites list on any machine on the network or maintain a master list of favorites and copy it to all of your users' computers on a regular basis.

Automating Tasks

Although you can easily access the file system on your network's Windows 98 computers and configure the operational environment in many ways, you cannot execute a program on another machine, except by indirect means. This ability to launch a program from another machine makes it possible to install new applications and perform routine system maintenance tasks without traveling to the computer. The following sections examine some of the workarounds that you can use to remotely trigger events on a network workstation.

Using the StartUp Group

The StartUp group has been a familiar feature since the early days of Windows. Any program placed in this group automatically launches as Windows loads. In Windows 98, the StartUp group is part of the Start Menu and contains shortcuts to applications located elsewhere on the system (or on the network).

Since it is part of the Start Menu, the StartUp group appears in the file system as a subdirectory called C:\Start Menu\Programs\StartUp. You can create shortcuts in the \StartUp subdirectory just as you would in any of the other \Start Menu subdirectories, but the programs will not execute until the next time that the user restarts or logs on to Windows 98.

Using the Task Scheduler

The Windows 98 Task Scheduler makes it possible to launch an application on a remote system without having to wait for the next system restart or logon. The application creates job packages that consist of a shortcut to an executable file and a schedule for its execution. You can configure a program to execute at a specific time and date, only once or repeatedly at given intervals. Other parameters limit the execution of tasks to particular circumstances, such as when the computer has been idle for a given period of time. The job packages used by the Task Scheduler are created by the Scheduled Task Wizard and stored in a directory called C:\Windows\Tasks as files with a .job extension. It is not possible to directly create new job packages in a remote system's \Tasks directory, but you can create jobs on your own machine and copy them to the \Tasks directory on another computer.

In order for this technique to work, the program that you schedule for execution must exist at the same location on the remote computer as on the local one. In addition, the clocks on the two computers should be set to the correct time and date, especially if you want to execute programs with precision. If, for example, you want to start the ScanDisk application to check for errors on a user's hard drive while they're waiting on the phone, the event that you've scheduled for two minutes from now won't execute promptly if the other computer's clock is an hour slow.

Using Remote Control Solutions

The only way to interactively launch an application on a remote network workstation is to use a product that enables you to take over the operation of the machine from a remote location. Microsoft's NetMeeting, for example, has a feature called Remote Desktop Sharing that can do this. Once you configure the other system to function as a NetMeeting host, you can connect to it from your workstation and exercise complete control over the machine. You can install, configure, and launch applications and perform any other task, just as if you were sitting at the workstation. For more information on using NetMeeting on your LAN, see Chapter 9, Network Applications.

Using Net Watcher

Net Watcher is a Windows 98 utility that enables you to monitor the network users that are connected to your computer, the shares that are currently being accessed, and the files that the remote users have open. You can also disconnect a user from a share, close a file that a user has open, and create new shares. Net Watcher is a useful tool for determining who is accessing files or shares at any given time. However, from a network administration standpoint, the best feature of this application is that you can connect to any computer on the network and perform these actions remotely for that system, as well as your own.

Connecting to a Remote System

When you launch Net Watcher from the Start Menu's Accessories\System Tools program group, the program displays the connections to your own machine that are currently open. To monitor the activity of another workstation on the local network, you choose Select Server from the Administer menu and specify the NetBIOS name or IP address of the computer to which you want to connect.

In order to connect to another system with Net Watcher, the other computer must have Remote Administration enabled. To do this, a user at the other machine must open the Passwords Control Panel to the Remote Administration page, fill the Enable Remote Administration of this Server checkbox, and specify a password that you will use when connecting to the workstation.

Tip: It is advisable that you use a strong Remote Administration password for systems that contain sensitive data, since a user with remote administration privileges can access all of the system's hard drives and create new shares without limitation.

When you enable Remote Administration, Windows 98 creates two administrative shares, as follows:

  • ADMIN$ - provides administrators with access to the file system, even when drives are not shared.
  • IPC$ - provides an interprocess communication (IPC) channel between the user's and the administrator's computers.

These shares are what enable you to interact with the remote system, as well as observe its networking activities.

Note: The Windows 98 Remote Administration feature is not the same as the Remote Registry service, that enables you to modify registry settings on other network systems. You can enable Remote Administration on any Windows 98 system, while the Remote Registry service requires user-level access control and a Windows NT system on the network.

Using the Connections Screen

When you first connect to another system with Net Watcher, the program displays the Connections screen, which contains a list of the users and computers that are currently accessing the system's shares, as shown in Figure 14-4. The left pane displays the number of shares and files that each user has open, while the right lists the open files in each share. You can disconnect a user from the computer (as well as any shares and files they are accessing) by highlighting a name and selecting Disconnect User from the Administer menu.

Figure 14-4. The Net Watcher Connections screen

As a security tool, Net Watcher enables you to monitor the network for unauthorized access to specific systems and shares and to take steps to prevent continued intrusion. Once you've discovered someone accessing a share without authorization, you can immediately disconnect them from the machine, and then switch to the Shared Folders screen to change the password on the share.

Note: Disconnecting a user from a system is a drastic step when the user has files open. The connection is severed with no warning to the user, and the interruption can result in data loss.

Using the Shared Folders Screen

The other two Net Watcher screens display the same information in a different format. The Shared Folders screen lists the drive shares on the system, the computers connected to them and the files opened by each computer (see Figure 14-5). From this screen, you can create and delete shares on the remote system and modify the properties of existing shares.

Figure 14-5. The Net Watcher Shared Folders screen

To create a new share, you select Add Shared Folder from the Administer menu and select the desired drive or directory from the Browse for Folder dialog box shown in Figure 14-6. This dialog box displays the existing shares on the computer, as well as the administrative shares, represented by a drive letter followed by a dollar sign (such as c$). You must select one of these administrative shares or a subdirectory of one of these shares as the root directory for your new share. Once you've made your selection, you see the standard Sharing dialog box (in which you specify a name for the share), the type of access you want to grant to users, and a password.

Figure 14-6. The administrative shares for a Windows 98 system appear as drive letters followed by a dollar sign

You can also modify the password for a share or the share name itself by selecting a share from the list and choosing Shared Folder Properties from the Administer menu. In the event of a security breech, you can even delete the share entirely to prevent all users from accessing it.

Using the Open Files Screen

The Open Files screen lists the files that are in use, and identifies who is using them (see Figure 14-7). From this screen you can close an individual file instead of disconnecting a user completely from the system by selecting Close File from the Administer menu. If, for example, a file is inaccessible because another user has left it open and walked away from their computer, you can close it without disturbing the user’s other work.

Figure 14-7. The Net Watcher Open Files screen

Creating User Profiles

User profiles are a Windows 98 feature that enable multiple users to maintain their own individualized settings on one computer. On a home or small business network where more than one user might work at the same machine, user profiles can store a desktop configuration and Start Menu for each user, as well as other configuration parameters. When you configure a system to use user profiles, you can specify exactly which parameters should be stored in each profile, and create user names and passwords for the system’s users. You cannot activate the user profile feature on a workstation from a remote location, but once the profiles are created, you can modify their contents in order to change the system configuration for specific individuals.

You create user profiles on a Windows 98 system from the User Profiles page in the Passwords Control Panel, shown in Figure 14-8. Here, you can specify whether each profile you create on the system should include desktop icons, Network Neighborhood contents, and the Start menu and program groups. When you do this, the system creates a new subdirectory in the C:\Windows directory called \Profiles, where the profiles for all of the system’s users will be stored.

Figure 14-8. The User Profiles page of the Passwords Control Panel

The system creates a subdirectory under \Profiles with the logon name of each user, as well as the following profile directories:

  • Application Data – Contains specialized information stored by the user’s applications
  • Cookies – Contains the cookie files supplied by web sites that the user has visited in Internet Explorer
  • Desktop – Contains shortcuts that will appear as icons on the user’s desktop
  • History – Contains Internet shortcuts to the sites visited by the user in Internet Explorer
  • NetHood – Contains shortcuts to shares on the network created by the user in the Network Neighborhood
  • Recent – Contains shortcuts to the user’s fifteen most recently opened documents
  • Start Menu – Contains subdirectories and shortcuts that will appear as the program groups and menu items in the user’s Start Menu

Once you’ve activated user profiles, the system will prompt every new user that logs on to the computer before creating a profile for that user in a new \Profiles subdirectory. Every time that a user logs off, any changes they have made to their settings are saved to their profile directories, instead of the common subdirectories in \Windows. You can also create profiles for the system’s users before they log on for the first time with the Users Control Panel. From the User Settings dialog box, you can create new users, specify their passwords, and manage their profiles individually. When you create a new user account, you can choose to copy the current system settings into the new \Profiles subdirectory, or create blank profile directories, forcing the user to create his or her own configuration from the default system settings.

While the User Profiles settings in the Passwords Control Panel are global, and apply to all users, the Users Control Panel enables you to specify individual profile parameters for each user. On the Personalized Items Setting screen for each user account (see Figure 14-9), you can specify which elements of the system configuration should be included in the profile. You can also add new directories to the profile that are not created by default, including the following:

  • Favorites – Contains Internet shortcuts created by the user in Internet Explorer
  • My Documents – Can be used to store the user’s personal document files
  • Temporary Internet Files – Contains the cached files stored by Internet Explorer as the user browses the web

Figure 14-9. You can create individual user profile settings from Personalized Items Settings dialog box in the Users Control Panel

When you enable user profiles, you must take the new locations of directories like \Desktop and \Start Menu into account if you plan to modify your workstation configurations from a remote system. For example, adding a shortcut to the C:\Windows\Desktop directory will have no effect when the user's desktop is stored in a profile directory. However, if you want to make changes to the desktop or Start Menu for all of a system's users, you can create shortcuts in the \Desktop and \Start Menu directories located in C:\Windows\All Users. Any shortcuts in these directories will appear on the system in addition to (not instead of) the items in the user's profile directories.

Using System Policies

Windows 98 stores most of its configuration information in the registry, a database of system settings that control all aspects of the system’s operation. When you use the Control Panel to modify system parameters, you are usually modifying settings in the registry. Windows 98 includes a program called the Registry Editor (Regedit.exe) that enables you to directly access the registry and modify its contents, but this is not recommended unless you know precisely what you are doing. The registry is composed of keys and entries with long, cryptic path names, making it difficult to find the setting that controls a specific element of the operating system. A simple typographical error in a registry modification can render an entire system unusable, and force you to completely reinstall Windows 98.

There is a safer method of modifying the registry that provides more configuration options than the Control Panel. The System Policy Editor (SPE) is an application that provides a graphical interface to the registry that enables you to make changes using standard Windows input fields and checkboxes. In addition to modifying the system’s registry directly, you can also create system policy files that contain settings for specific individuals. You can set up Windows 98 so that whenever a user logs on to the system, the policy file applies specific registry settings that configure the system as you see fit.

From a network administration standpoint, system policies are an excellent means of restricting a user's access to the workstation. Using system policies, you can prevent a user from accessing certain parts of the Control Panel and other features of the operating system, thus reducing the technical support problems that result from users experimenting with controls that they don't understand. You can even restrict a user to a collection of applications that you specify, preventing them from running games or other unauthorized software.

Installing the System Policy Editor

The System Policy Editor is not installed with the Windows 98 operating system. You must install it manually by selecting the Have Disk button in the Windows Setup page of the Add/Remove Programs Control Panel and browsing to the \tools\reskit\netadmin\poledit directory on the Windows 98 CD-ROM. The installation procedure copies the required files to the system and adds a shortcut for the System Policy Editor to the Start Menu in the System Tools program group.

The System Policy Editor relies on template files that contain the information displayed in the program's interface and the registry settings that each policy modifies. Policy templates are text files with an .adm extension that you load by selecting the Policy Template item from the Options menu. By default, the editor loads two templates, called Common.adm and Windows.adm, but the operating system includes the following additional templates:

  • Appsini.adm - Enable the use of an Apps.ini file for network applications installations (see Chapter 9, Network Applications
  • Chat.adm - Provides configuration settings for Microsoft Chat
  • Conf.adm - Provides configuration settings for NetMeeting
  • Inetresm.adm - Provides user interface configuration settings for Internet Explorer
  • Inetsetm.adm - Provides advanced configuration settings for Internet Explorer
  • Oem.adm - Provides configuration settings for Outlook Express
  • Shellm.adm - Provides configuration settings for the Windows 98 desktop shell
  • Subsm.adm - Provides configuration settings for Internet Explorer subscriptions

Other applications, such as the Internet Explorer Administration Kit and Microsoft Office, include their own policy templates that you can use to configure specific elements of those applications. After you launch the System Policy Editor, you can install any or all of the templates available to you. The editor will combine the settings of the installed templates into a single display.

Using the System Policy Editor

The System Policy Editor can run in either of two modes: registry mode or policy file mode. When you enter registry mode by selecting Open Registry from the File menu, the editor connects directly to the registry on the local system. Any changes that you make to the policies displayed in the editor will be immediately applied to the system when you select Save from the File menu.

Tip: It is possible to connect to a remote system on the LAN with the System Policy Editor and modify its registry settings directly. To do this, the remote system must be configured for user-level access control and have the Remote Registry service installed. User-level access control requires that the network have a Windows NT system function as a pass-through server. Networks composed of Windows 98 systems only must use share-level access control; you must therefore configure each system to use system policies individually.

To use the editor in policy file mode, you select New Policy from the File menu, and the settings you configure are saved to a file with a .pol extension. A single policy file can contain settings for specific users and computers on your network, as well as general settings that will be applied to every user and computer. Once you create a policy file, you can configure a Windows 98 workstation to load it whenever it boots and apply the settings contained in the file to the system.

Setting System Policies

When you launch the System Policy Editor (in either mode), you see a screen containing Default Computer and Default User objects. System policies can be applied to either a computer or a user, depending on where in the registry the setting configured by the policy is stored. In addition to the two default objects, you can create additional user and computer objects by selecting the appropriate item from the Edit menu, as shown in Figure 14-10. The policies in the two default objects are applied to all computers and users that load the policy file, except for those that have specific computer or user objects named for them. For example, if you create a user object called jdoe, the system will load the policies that it contains when John Doe logs on to the system. When any other user logs on, the system applies the Default User policies.

Figure 14-10. The System Policy Editor main screen

When you open a computer or user object in the editor, you see a hierarchical display like that shown in Figure 14-11. You can expand each of the subheadings to display the individual policies it contains.

Figure 14-11. System policies appear in the editor as part of a hierarchical display

Each policy has a checkbox that can have three possible states, as follows:

  • Selected (checkbox filled) - The policy is implemented and the associated registry settings applied
  • Cleared (checkbox empty) - The policy is not implemented, and the associated registry settings, if they exist, are removed
  • Unavailable (checkbox grayed out) - The policy is not implemented, and the associated registry settings are not changed, no matter what their value

In addition to the checkbox, a policy may have other controls in the Settings box below the expandable display. The Settings box may contain additional checkboxes, pulldown menus, fields for the entry of additional data, or just informational commentary. When you fill the checkbox for a policy, the controls in the Settings box (if any) are activated, and you can use them to further configure the policy.

Deploying Policy Files

On a workgroup network that consists only of Windows 98 machines, you must configure each workstation to use system policies individually. You can run the System Policy Editor on each workstation and modify the registry directly, or you can create a policy file and configure each workstation to load it. Since a policy file can contain settings for all of the users on your network, there is no need to create a separate file for each workstation, even if you want to use different policy settings for each user. You can create a single file on your own workstation, so that you only need to configure each workstation to load that file.

Note: The standard name for the policy file when you deploy it using a Windows NT server is Config.pol. However, for a policy file that you will deploy on a Windows 98 network, you can use any file name.

Once you have created the policy file, you can put a copy of it on each of your workstations, but it will be easier to make changes later if you use one copy of the file and store it on a network share. It's a good idea to store the policy file on a share to which your users have read-only access, so that they cannot make changes to it.

To configure each workstation to load the policy file, you launch the System Policy Editor and open the local registry. In the Default Computer object, browse to the Windows 98 Network/Update/Remote Update policy. After enabling the policy, select Manual for the Update Mode setting and specify the path and file name for the policy file you created in the Path for Manual Update field. The file can be located on a local or network drive, but for a network drive, use a UNC name rather than a drive letter, so the system will always be able to find the file.

After saving the changes you've made in the editor, you can reboot the system, and the policy file will be applied as the system restarts and the user logs on.

Restricting Workstation Access

Using system policies, you can prevent your users from accessing parts of the Windows 98 operating system that you'd rather they didn't work with themselves. The protection against tampering provided by system policies is not absolute; a clever user can bypass the policy restrictions in several ways. For the average user, though, eliminating some of the more complicated elements of the Windows 98 interface can make the overall computing experience less intimidating.

When you open a user object in the System Policy Editor and browse down to the Windows System heading, you see a number of subheadings that contain policies that you can use to restrict access to the system, including the following:

  • Shell/Restrictions – Removes specific items from the desktop, Start Menu, My Computer, and/or Network Neighborhood
  • Control Panel – Restricts access to the Display, Network, Passwords, Printer, and/or System Control Panels
  • Restrictions – Prevents users from access the registry, running unauthorized applications, and/or opening a DOS session

Using these policies, you can create a "locked down" workstation configuration in which users are able to run only specific applications and have limited access to system controls. For example, you can remove the Run command from the Start Menu, hide the drive letters in My Computer, and block access to the DOS prompt, in order to prevent users from launching unauthorized programs. By disabling the registry editing tools and restricting access to the Control Panel, you can prevent users from experimenting with system settings that might cause the computer to malfunction.

Creating a Roving Profile

As you learned earlier in this chapter, user profiles enable multiple users to work in a single computer, each with their own system configuration. Normally, the profile directories are stored on the local drive, but it is also possible to use system policies to configure a workstation to access profile directories from a network share. This way, users can travel to any workstation on the network and access their own personal icons and menus simply by logging on. This is known as a roving profile.

In order to use roving profiles in your network, all of your workstations must have the same applications installed in the same places, as you will be using one set of shortcuts to load them on different computers. Access to the share on which the profiles are stored should depend on whether you want your users to be able to modify them. If you store the profiles on a read-only share, any changes that users make to the system configuration will be lost when they log off. This is known as a mandatory profile, because users receive the exact same profile settings whenever (and wherever) they log on.

To create roving profiles, you create a system policy file with a separate user object for each user. In each user object, you browse to the Windows 98 System/Shell/Custom Folders subheading and enable the policies for the custom profile folders you want your users to be able to access over the network. For a typical scenario, you should enable the following policies:

  • Custom Desktop Icons
  • Hide Start Menu Subfolders
  • Custom Start Menu

In the fields provided, enter UNC path names to the \Desktop and \Start Menu folders for that user's profile. The UNC names will function properly even when the person logs on to the workstation where the profile directories are stored. After setting any other system policies you want to apply, save them to a policy file and deploy the file on your workstations, as discussed earlier. Whenever a user logs on to a workstation and loads the policy file, the system will access the profile directories on the network share and configure the desktop and Start Menu accordingly.

Backing Up

If you value your data, you must back it up. This is the most basic rule of computing, and many users suffer for not following it. Whether your files consist of important business documents, your kids' term papers, or your grocery list, you should back them up regularly if losing them will be an inconvenience. Having a network means that you don't have to install a separate backup drive on every computer. You can use one drive to back up the files on all of your network workstations in one single job. Windows 98 includes a backup utility that supports most of the major tape and cartridge drives on the market. With Plug and Play, getting a backup solution up and running is usually a simply task.

The first step in developing a backup strategy is to determine what data you need to back up, how much of it there is, how much time you have to back it up, and how often you should back it up. You use this information to determine what type of backup hardware you will need.

It is usually not necessary to back up the entire contents of the hard drives on your Windows 98 systems. At the very least, there are temporary files like the Windows 98 swap file used for memory paging and the files in the caches maintained by web browsers that you can omit from your backup routines, thus saving space on the tape or other medium. In many cases, a suitable backup strategy will consist of full system backups at regular intervals, such as each week, and daily jobs that back up only the files that have changed since the last full backup. On a home network, full backups can be less frequent, but you should perform frequent backups of your important data files, such as word processor documents and spreadsheets.

Selecting a Backup Device

The traditional device for performing backups is a magnetic tape drive. There are a wide range of devices available in several different tape formats. The basic criteria that you should use to select a device are as follows:

  • Interface – SCSI, which requires that you purchase and install a SCSI host adapter card, is the traditional, high-performance interface used for tape drives. However, many of the devices available today use a parallel port SCSI interface that simplifies the hardware installation process.
  • Capacity – An appropriate tape drive for your network should be able to perform your most common backup jobs using a single tape. Smaller capacity drives may be cheaper, but having to change tapes several times to complete a job will make you less likely to perform backups as often as you should.
  • Speed – Unless you have users working on the network 24 hours a day, backup speed should not be a major concern, as you can start a job and leave it to run unattended.
  • Cost – Consider not only the cost of the drive itself, but the media that it uses as well. Some types of data tape can be quite expensive. If you plan to use a SCSI drive, be sure to include the cost of the SCSI host adapter.

In addition to tape drives, there are several other types of storage devices that you can use to perform backups. Floppy drives, though universally available, are too slow and offer too small a capacity for anything other than temporary storage of a few files. Cartridge drives like the Iomega Zip and Jaz are acceptably fast, can conceivably store large amounts of data, and are becoming standard equipment in many of the computers sold today. However, the cost of the media for these drives is prohibitive. Writable CD-ROMs are also a viable alternative, since the price of blank disks has dropped to the point at which, even if you can only write to them once, using them for backups is affordable.

The problem with these alternative media is that many backup software programs do not support them. While you can still use them for informal backups, by copying files manually to the drive, you'll find that good backup software simplifies the process of selecting backup targets and locating files for restoration.

Backup Types

The most basic form of backup copies the entire contents of a drive to the tape (or other medium). This is known as a full backup. As a relatively small number of files on the average PC are modified on a daily basis, running a full backup every day is not necessary, because it copies many of the same files over and over again. However, there are certain files that you do want to back up often, such as document files that are edited on a regular basis.

Backup programs typically identify the files that need backing up using the DOS archive bit. When the program performs a full backup, it strips the archive bit from each file as it copies it to tape. Whenever a file is added or modified, the operating system adds the archive bit again. Thus, during the next backup job, the software can save time and tape by backing up only the files that have archive bits.

When performing this type of partial backup, the software can either strip the archive bits off of the files it backs up again, or leave the archive bits intact until the next full backup. These two types of jobs are called incremental and differential backups, respectively. Incremental jobs take the least amount of time and use the least amount of tape, because only the files that have changed since the last job are backed up. However, if you lose an entire drive, you must restore the last full backup job plus every incremental job performed since that full backup, because the latest version of a given file can be located on any of the incremental tapes.

When you run differential jobs, all of the files changed since the last full backup are saved every night. This requires more tape, since the files changed on the day of the first differential job have to be backed up every succeeding day until the next full backup. The benefit of this method, however, is that a full restore only requires the last full backup and the most recent differential. For a typical small network, the best strategy is to perform a full backup once a week and daily differentials.

Windows 98 Backup

Windows 98 includes a backup software program that enables you to select files, directories, or entire computers for backup and restore, anywhere on your network, using the interface shown in Figure 14-12. The program provides the ability to create differential and incremental, as well as full backup jobs, in addition to most of the standard options found in backup software products.

Figure 14-12. The Windows 98 Backup main screen

The primary feature that is lacking in Windows 98 Backup is the ability to schedule jobs to run at specified times. While you can use the Windows 98 Task Scheduler to launch the backup program at a certain time, you cannot execute a job this way. Virtually all of the third-party backup software products on the market include the ability to create rotational schedules that automatically run a series of jobs at regular intervals. Once you have created the schedule, all that you have to do is change the tape in the drive.

Backing up network drives using this software is as easy as browsing through the shares, just as you would in Windows 98 Explorer, and selecting the files or directories that you want to back up. The software creates an index of the files that comprise each backup job and writes it to the tape, so that when you have to perform a restore, the program reads the index and displays the same sort of interface, in which you can select specific items to be restored back to their original locations, or to an alternate drive.

The Future

Computer networking is a subject that you can study for a lifetime and still learn something new every day. The technology advances at an incredible rate, and sometimes just keeping up with the latest innovations can be difficult. For some users, the networking functions and applications discussed in this book will be sufficient for their home or business needs for some time to come. Other users will continue to expand and upgrade their networks. As you learn more about networking, you will probably find new applications for your LAN and you will certainly find new ways of running it more efficiently.