Chapter 15 - Network Adapters and Protocols

This chapter describes technical issues related to network adapters and protocols for Microsoft Windows 98 and explains how to install and configure network adapters and protocols. It also presents some technical notes and tips for networking.

In This Chapter

Overview of Network Adapter Drivers and Protocols

Overview of Protocols

Microsoft TCP/IP Protocol

IPX/SPX-compatible Protocol

Microsoft NetBEUI Protocol

Microsoft DLC Protocol

Overview of Network Adapters

Troubleshooting Network Adapters and Protocols

See Also

• For an overview of network configuration, see Chapter 14, "Introduction to Networking Configuration."

• For more information about using Windows 98 on Microsoft networks, see Chapter 16, "Windows 98 on Microsoft Networks."

• For more information about using Windows 98 on third-party networks, see Chapter 17, "Windows 98 on Third-Party Networks."

• For more information about logon, browsing, and resource sharing, see Chapter 18, "Logon, Browsing, and Resource Sharing."

• For more information about Dial-Up Networking, see Chapter 19, "Remote Networking and Mobile Computing."

• For more information about network architecture, see Chapter 29, "Windows 98 Network Architecture."

Overview of Network Adapter Drivers and Protocols

A network adapter (sometimes called a network interface card, or NIC) is a hardware card installed in a computer so it can communicate on a network. The network adapter provides one or more ports for the network cable to connect to, and it transmits and receives data onto the network cable.

Every networked computer must also have a network adapter driver, which controls the network adapter. Each network adapter driver is configured to run with a certain type of network adapter.

A networked computer must also have one or more protocol drivers (sometimes called a transport protocol or just a protocol). The protocol driver works between the upper-level network software and the network adapter to package data to be sent on the network.

In most cases, for two computers to communicate on a network, they must use identical protocols. Sometimes, a computer is configured to use multiple protocols. In this case, two computers need only one protocol in common to communicate. For example, a computer running File and Printer Sharing for Microsoft Networks that uses both NetBEUI and TCP/IP can communicate with computers using only NetBEUI or only TCP/IP. Common protocols are not necessary if the computers are using an intermediary device or application such as the gateway. For a brief explanation of gateways, see "Using Gateways for Connectivity" in Chapter 17, "Windows 98 on Third-Party Networks."

In Windows 98, you configure all network adapter drivers and protocols supporting protected-mode clients by using the Network option in Control Panel rather than by manually editing configuration files. Configuration values are stored in the registry.

Windows 98 Setup automatically configures a computer to use protocols and drivers to match network components that are running when Setup is started.

If you are setting up Windows 98 for a new networking option, you must choose which types of network adapter drivers and protocols to use. Because Windows 98 has an open architecture, you have a lot of flexibility in this decision. Windows 98 supports both network driver interface specification (NDIS) and transport driver interface (TDI) standards, allowing Windows 98 to communicate with many other networking products and making it possible to choose from a variety of network adapters and protocols.

This section describes basic issues for choosing network adapter drivers and protocols to support your networking needs.

Choosing Adapters and Drivers for Best Performance

Network adapters have become exceptionally reliable and inexpensive. The low cost of Ethernet adapters, including new Plug and Play hardware, means that usually the cost-effective way to improve network performance is to replace an older network adapter with a new model. The cost of the new hardware is offset almost immediately by savings in support time and improved performance. Also, you should buy an adapter that matches one of your computer's buses. For example, peripheral component interconnect (PCI) network adapters are available for use in PCI computers.

Choosing Protected-mode Protocols and Adapter Drivers

Microsoft recommends that you choose separate 32-bit, protected-mode protocols and drivers rather than real-mode protocols and drivers. However, Windows 98 includes mapping technology for users who must continue to use real-mode NDIS 2 or Open Datalink Interface (ODI) drivers.

Overview of Protocols

Windows 98 network protocols are implemented as 32-bit, protected-mode virtual device drivers (VxDs) that offer high performance. Windows 98 can support multiple protocols simultaneously, and protocol stacks can be shared among the installed network clients. For example, the Internet Packet Exchange/Sequenced Packet Exchange (IPX/SPX) – compatible protocol can support both Client for NetWare Networks and Client for Microsoft Networks. The following protocols are included with Windows 98:

• TCP/IP

• IPX/SPX-compatible

• NetBEUI

• 32-bit DLC

All four protocols are Plug and Play – compliant, so they can be loaded and unloaded automatically. For example, if a PC Card network adapter is removed from the computer so that the network is no longer available, the protocols are unloaded automatically after any dependent applications have been notified.

Windows 98 can support multiple network protocols and can share a protocol among the network providers that are installed. You might choose more than one protocol to ensure communication compatibility with all systems in the enterprise. However, choosing multiple protocols can cause more network traffic, more memory used on the local computers, and more network delays.

The following sections briefly describe the benefits of using each protocol and issues to consider when using each protocol.

MS TCP/IP Protocol

This is a complete implementation of the standard, routable Transmission Control Protocol/Internet Protocol (TCP/IP) protocol. Windows 98 includes only protected-mode support for this protocol. Microsoft TCP/IP provides the following benefits:

• Support for Internet connectivity and the Point-to-Point Protocol (PPP).

• Connectivity across interconnected networks with different operating systems and hardware platforms, including communication with many non-Microsoft systems, such as Internet hosts, Apple Macintosh systems, IBM mainframes, UNIX systems, and Open VMS systems.

• Support for automatic TCP/IP configuration using Dynamic Host Configuration Protocol (DHCP) servers such as Windows NT servers.

• Support for automatic IP-address-to-NetBIOS computer name resolution using Windows NT Windows Internet Naming Service (WINS) servers.

• Support for Windows Sockets 1.1 and 2.0, which are used by many client/server applications and many public-domain Internet tools.

• Support for the NetBIOS interface, commonly known as NetBIOS over TCP/IP.

• Support for many commonly used utilities, which are installed with the protocol.

Consider the following issues when using this protocol:

• TCP/IP in general has been known to require careful planning and management of the Internet Protocol (IP) address space. However, this problem is vastly reduced when you use DHCP servers to manage assignment of IP addresses for computers running Microsoft TCP/IP.

• If you want to take advantage of DHCP for private IP addressing or use WINS for name resolution on computers running Windows 98, the appropriate DHCP servers must be in place on the network.

• The Microsoft TCP/IP stack does not bind to NetWare client; therefore, you cannot use TCP/IP as a substitute for NetWare/IP when connecting to a NetWare server.

MS IPX/SPX-compatible Protocol

This protocol is compatible with the Novell NetWare Internetwork Packet Exchange/Sequential Packet Exchange (IPX/SPX) implementation. Windows 98 includes both 32-bit, protected-mode and real-mode support for this protocol. This protected-mode protocol provides the following benefits:

• Works with Microsoft Client for NetWare Networks and Novell Client for Windows 95/98.

• Supports packet-burst mode to offer improved network performance.

• Supports the Windows Sockets, NetBIOS, and ECB programming interfaces.

• Support for automatic detection of frame type, network address, and other configuration settings.

• Routable connectivity across all network bridges and routers configured for IPX/SPX routing.

Consider the following issues when using this protocol:

• This protocol is required for Microsoft Client for NetWare Networks and installed automatically with the client. When Windows 98 Setup determines that it cannot install Client for NetWare Networks on a computer running a Novell-supplied network client, Setup still tries to install this protected-mode protocol. For information about how Setup determines whether to install this protocol automatically, see Chapter 17, "Windows 98 on Third-Party Networks."

• This protocol cannot be used to configure Windows 98 to support NetWare over ArcNet. Instead, you must use real-mode IPX drivers with NetBIOS support on ArcNet network adapters.

• With this protocol, it is not necessary to load the Novell-supplied Vipx.386 driver, because the Microsoft protocol provides virtualized services to all virtual machines (VMs) and applications.

MS NetBEUI Protocol

This protocol is compatible with existing networks that use the NetBIOS extended user interface (NetBEUI), including Windows for Workgroups peer networks, Windows NT Server, LAN Manager, and other networks, and includes support for the NetBIOS programming interface. Windows 98 provides both protected-mode and real-mode support for this protocol.

NetBEUI was first introduced by IBM in 1985, when it was assumed that LANs would be segmented into workgroups of 20 to 200 computers and that gateways would be used to connect one LAN segment to other LAN segments or to a mainframe. NetBEUI is optimized for high performance when used in departmental LANs or LAN segments. Microsoft NetBEUI under Windows 98 is completely self-tuning.

One common method for setting up a network is to use NetBEUI plus a protocol such as TCP/IP on each computer that needs to access computers across a router. If you set NetBEUI as the default protocol, Windows 98 uses NetBEUI for communication within the LAN segment and uses TCP/IP for communication across routers to other parts of the WAN.

Consider the following issues when choosing this protocol:

• NetBEUI is a nonroutable protocol that cannot cross routers, although it can cross bridges and source routing bridges.

• NetBEUI is optimized for high performance only for use in departmental LANs or LAN segments.

MS DLC Protocol

The 32-bit DLC protocol provides connection and communication with mainframe computers using DLC. With the 32-bit Data Link Control (DLC) protocol, you can establish multiple connections to different IBM host and AS/400 computers over the same network token-ring, FDDI, or Ethernet adapter. Host terminal emulation programs use the 32-bit DLC to communicate directly with host computers. The 32-bit DLC protocol also allows multiple 32-bit applications to use the same network adapter at the same time to connect to different host computers. You can also use the 32-bit DLC protocol to provide connectivity to local area printers connected directly to the network.

The 16-bit DLC protocol is also available; however, in most cases, it is recommended that you use the 32-bit DLC protocol.

MS TCP/IP Protocol

Microsoft Transmission Control Protocol/Internet Protocol (TCP/IP) provides communication across interconnected networks that use diverse hardware architectures and various operating systems. TCP/IP can be used to communicate with computers running Windows 98, with devices using other Microsoft networking products, or with non-Microsoft systems such as UNIX.

Microsoft TCP/IP in Windows 98 extends the functionality that Microsoft TCP/IP offered in Windows 95. In Windows 95, Microsoft TCP/IP provided the following elements:

• Core TCP/IP protocols, including the Transmission Control Protocol (TCP), Internet Protocol (IP), User Datagram Protocol (UDP), Address Resolution Protocol (ARP), Internet Control Message Protocol (ICMP), and Domain Name System Protocol (DNS). This suite of Internet protocols provides a set of standards for how computers communicate and how networks are interconnected.

• Support for application programming interfaces (APIs), including Windows Sockets and NetBIOS.

• TCP/IP diagnostic tools to detect and resolve TCP/IP networking problems, including arp, ftp, nbtstat, netstat, ping, route, telnet, tracert, and ipconfig, plus Windows-based Telnet and IP Configuration (winipcfg) utilities.

• Client for DHCP, for automatic configuration of computers running TCP/IP on networks that have DHCP servers.

• Client for WINS, for dynamic resolution of IP addresses to computer names on networks that have WINS servers.

• Point-to-Point Protocol (PPP) for asynchronous communication, as described in Chapter 19, "Remote Networking and Mobile Computing."

With Windows 98, the following enhancements are added:

• Automatic private IP addressing by the client, as described in the section "Configuring IP Addresses Using Automatic Private IP Addressing" later in this chapter.

• Multihoming, as described in the section "Configuring Multihoming" later in this chapter.

• Windows Sockets 2 support, as described in "Enhancements to TCP/IP in Windows 98" later in this chapter.

• Support for IP Multicast (Request for Comments [RFC] 1112), including support for Internet Draft for IGMP version 2.

• Support for Internet Control Message Protocol (ICMP) Router Discovery (RFC 1256).

• Performance enhancements for certain types of high-speed, high-bandwidth networks. These include support for TCP Large Windows (RFC 1323), support for Selective Acknowledgments (RFC 2018), and support for TCP Fast Retransmit and Fast Recovery.

• DHCP enhancements, including address assignment conflict detection and longer timeout intervals.

For information about the Resource Reservation Protocol (RSVP) and Generic Quality of Service (GQoS) architecture, see Chapter 29, "Windows 98 Network Architecture." . For more information about RSVP, see also the Internet Engineering Task Force (IETF) RSVP specification at https://www.ietf.org .

For other RFCs and Internet drafts listed above, see https://www.ietf.org** **.

Planning for TCP/IP

This section gives a broad overview of IP addressing. Then it presents a few common types of networks and explains issues you might want to consider when implementing each type of network.

For information about the architecture of TCP/IP and its Windows Sockets 2 socket API, see Chapter 29, "Windows 98 Network Architecture."

For more conceptual information about TCP/IP, information about routing, and information about using TCP/IP in a large installation with Windows NT Server, see the following publications:

• Microsoft Windows NT Server Networking Guide for Microsoft Windows NT Server version 4.0.

• Network Supplement for Microsoft Windows NT Server version 4.0.

• Networking with Microsoft TCP/IP by Drew Heywood (New Riders Publishing, 1996).

See also the TCP/IP white paper "Microsoft TCP/IP and Windows 95 Networking".

For TCP/IP shareware tools such as finger, chat, whois, and nslookup, visit the following Web sites:

• https://www.download.com  

• https://www.shareware.com ** **

• https://www.tucows.com  

• https://www.windows95.com ** **

Understanding IP Addressing

Each workstation needs an IP address to communicate on a TCP/IP network such as the corporate network or the global public Internet. The section "Configuring IP Addresses" later in this chapter describes IP addresses in more detail, but for now it is just important to understand that there are two kinds of IP addresses, globally unique IP addresses and private IP addresses.

• Globally unique IP addresses, which you can use on the global Internet.

  On the Internet, the Internet Assigned Numbers Authority (IANA) assigns groups of IP addresses to organizations. The organizations can then assign IP addresses within those groups to individual computers. This prevents multiple computers from having the same IP address.

  For a computer to be visible on the Internet, it must be reachable through a globally unique IP address.

• Private IP addresses, which you cannot use on the global Internet.

  The IANA has reserved a certain number of IP addresses that will never be used on the global Internet. You can use these addresses for computers that will never be used to access the Internet.

  In some cases, you can also use these addresses for computers that need partial connectivity to the Internet but do not need to be directly reachable from the Internet. For example, if you need only Web browsing and e-mail connectivity on a computer, that computer does not necessarily have to be directly reachable from the Internet.

  However, in order to enable computers to use private IP addresses, you must deploy a firewall with proxy or Network Address Translator (NAT) capabilities. A proxy or NAT sits between a private network (a network of computers using private IP addresses) and the Internet. The firewall has a globally unique IP address that can be used on the Internet. When a computer on the private network sends a packet to the Internet, the NAT substitutes the source IP address on that packet for the NAT's own IP address. Thus, the packets appear to originate from the NAT.

  This configuration provides flexibility and a certain level of security: hosts on the Internet cannot directly send IP packets to the computer on the private network because they do not know and cannot use its IP address. The firewall can provide a level of administrative control for which machines and which applications can reach computers on the Internet, and vice versa.

  For more information about private IP addresses, see Request for Comments (RFC) 1918. For more information about NATs, see RFC 1631.

Depending on your needs, you can use either private IP addresses or globally unique IP addresses for each workstation. You could also use both a private IP address and a globally unique IP address if you have a multihomed computer (a computer that has two different adapters, each connected to a different network). For example, a workstation could be connected both to the corporate network (using an adapter that is configured with a private IP address) and to the Internet (using a dial-up adapter that is configured with a globally unique IP address). For information about multihoming and this type of special configuration, see the section "Configuring Multihoming" later in this chapter.

Regardless of the type of IP address you choose, you have two options for configuring the IP addresses:

• Dynamic IP addressing.

  With dynamic IP addressing, the IP address can be configured automatically. This method is much simpler, decreases your management time, and enables you to reuse IP addresses. It is recommended for all sizes of networks.

  In Windows 95, two types of dynamic IP addressing were available. For corporate networks, if there was a Dynamic Host Configuration Protocol (DHCP) server on the network, the computer could automatically obtain the IP address from the DHCP server. And for dial-up Internet connections, some Internet Service Providers (ISPs) automatically assigned your computer a dynamic IP address.

  Windows 98 provides a third option, automatic private IP addressing. This option is recommended for simple networks that have one LAN (subnet) and no DHCP servers. With automatic private IP addressing, if no DHCP server is available, the computer automatically assigns itself a private IP address. (If a DHCP server later becomes available, the computer obtains an IP address from the DHCP server instead.) Computers using private IP addresses can communicate only with other computers using private IP addresses, on the same subnet. They are not directly reachable from the Internet.

• Static IP addressing.

  With static IP addressing, you must manually configure the IP address. This method is time-consuming, especially on medium to large networks. It is not recommended except as a last resort.

For more information about automatic private IP addressing, see "Configuring IP Addresses Using Automatic Private IP Addressing" later in this chapter.

For more information about DHCP, see "Automatically Configuring IP Addresses with DHCP" later in this chapter.

Planning for a Dial-Up Networking Connection to the Internet

If you plan to use Dial-Up Networking to connect to the Internet, you must configure the Dial-Up Networking connection. Your ISP might automatically assign you an IP address. If not, you must configure information such as the following:

• An IP address

• A default gateway (default router)

For more information about creating a Dial-Up Networking connection, see Chapter 19, "Remote Networking and Mobile Computing."

For more information about configuring an IP address for a Dial-Up Networking connection, see Chapter 20, "Internet Access and Tools."

Planning for Small Networks

Figure 15.1 shows an example of a small (single-subnet) network that is connected to the Internet by a gateway. In this example, all the computers use globally unique IP addresses and are all visible on the Internet.

Figure 15.1 Small network with gateway 

For this type of network, you must obtain a network ID that is valid on the Internet. You must also configure a gateway to reach the Internet, and each computer needs the address of the gateway in order to reach the Internet. A Windows 98 computer cannot be configured to act as a gateway.

For addressing, you should add a DHCP server and let it automatically assign IP addresses. You could also use static IP addressing, but to do so you must manually configure each computer not only now but whenever you make a change to your network that requires new IP addresses.

Now, suppose that the network includes a NAT. You must still obtain a network ID that is valid on the Internet, configure a gateway to reach the Internet, and configure each computer with the address of the gateway. Also, you can still use either DHCP or static addressing. However, because the NAT shields the IP addresses on your network from the Internet, you can also use private IP addressing. On a simple LAN you can use either DHCP to assign private IP addresses, or you can use the automatic private IP addressing feature to let the computer assign itself a unique private address on the LAN.

Planning for Large Networks

Figure 15.2 shows an example of a large network that is divided into several subnetworks and is connected to the Internet by a gateway.

Figure 15.2 Large network with gateway 

As with the previous example, you must obtain a network ID that is valid on the Internet. You must also set up one or more gateways. If you have multiple gateways, at least one must be configured as the default gateway (the gateway that is used to connect to the rest of the network). For more information about default gateways, see the section "Configuring Multihoming" later in this chapter.

For this type of network, you should use DHCP addressing. If your network does include a NAT, you can assign private IP addresses to each computer, but you should not do so using the Windows 98 automatic private IP addressing feature because the automatic private IP addressing feature guarantees uniqueness only on a LAN-per-LAN basis. Computers that use the automatic private IP addressing feature can communicate only with computers in their immediate subnet. Thus, you should use DHCP when there are routers on the network, even if each computer uses private IP addresses. You could, however, use the automatic private IP addressing feature as a backup method if your DHCP server fails.

If your computer does not include a proxy server or a NAT firewall and you want your machines to communicate on the Internet, you must use DHCP for IP address assignment and not the automatic private IP addressing feature.

Installing MS TCP/IP

Setup installs Microsoft TCP/IP by default if you are installing Windows 98 (instead of upgrading to Windows 98) and you have a network adapter or modem in your computer. If you are upgrading to Windows 98, however, Setup generally keeps your network configuration and you will need to install Microsoft TCP/IP if you want to use it.

If your original Windows installation included a third-party TCP/IP protocol stack, Setup keeps the existing TCP/IP protocol stack instead of installing Microsoft TCP/IP. However, you might want to use Microsoft TCP/IP instead. For example, many third-party TCP/IP stacks do not yet support Windows Sockets 2. If you want to use Microsoft TCP/IP instead of your third-party stack, you must uninstall the third-party stack by using the uninstall utility provided by your network vendor, then install Microsoft TCP/IP.

To install Microsoft TCP/IP after Windows 98 Setup

1. In Control Panel, double-click Network.

2. On the Configuration tab, click Add.

3. Select Protocol.

4. Click Add.

5. In Manufacturers, select Microsoft.

6. In Network Protocols, select TCP/IP.

7. Click OK.

Configuring TCP/IP

This section discusses the following topics:

• Configuring IP addresses.

• Configuring a multihomed computer.

• Using the route table to view your IP configuration and to add routes to hosts on the network.

Configuring IP Addresses

Every computer on a TCP/IP network is identified by a unique 32-bit IP address, which also specifies routing information in an internetwork. An IP address looks like this:

172.16.94.97

This is referred to as dotted decimal notation, with each eight bits of an IP address (called an octet) separated from the next eight bits by a period. An IP address is a single value that contains two pieces of information:

• The network ID, which is the portion of the IP address that identifies a group of computers and other devices that are all located on the same logical network.

• The host ID, which identifies a particular computer within a particular network ID. A host, or node, is any device that is attached to the network and uses TCP/IP.

Each host on the network uses the network ID and host ID to determine which packets it should receive or ignore, and to determine the scope of its transmissions (only hosts with the same network ID accept each other's IP-level broadcasts).

The Internet community uses address classes to differentiate networks of various sizes. The network class can be determined from the first octet of its IP address. Table 15.1 summarizes the relationship between the first octet of an IP address and its network ID and host ID, using w.x.y.z. to designate the four octets of the IP address. As Table 15.1 shows, the value of the first octet determines which portion of the IP address will be the network ID and which portion will be the host ID.

For example, the first octet of the sample IP address 172.16.34.1 is 172. The w values portion of Table 15.1 shows that if the first octet has a value of 128-191, it is a Class B address. Thus, the sample IP address is a Class B address. Its first octet (172) is the network ID, and the remaining octets (16.34.1) are its host ID.

Note The sample IP addresses used in this chapter are private IP addresses, not IP addresses that are valid on the global Internet.

The table also identifies the total number of network IDs and host IDs for each address class that participates in the Internet addressing scheme.

Table 15.1 IP address classes 

Because the sender's IP address is included in every outgoing IP packet, the receiving computer can derive the originating network ID and host ID from the IP address field. This is done by using subnet masks, which are 32-bit values that allow the recipient of IP packets to distinguish the network ID and host ID portions of the IP address.

The value of a subnet mask can also be represented in dotted decimal notation. Subnet masks are determined by assigning ones to bits that belong to the network ID and zeroes to bits that belong to the host ID. When the bits are in place, the 32-bit value is converted to dotted decimal notation, as shown in Table 15.2.

Table 15.2 Default subnet masks for standard IP address classes 

The result allows TCP/IP to determine the host ID and network ID of the local computer. For example, if the IP address is 172.16.34.1 and the subnet mask is 255.255.0.0, then the network ID is 172.16 and the host ID is 34.1.

Subnet masks are also used to further segment an assigned network ID among several local networks. For example, a network using the Class B network address 10.100 is one of over 16,000 Class B addresses capable of serving more than 65,000 nodes each. But if this corporate network includes 12 international LANs with 75 to 100 nodes each, it is better to use subnetting to make effective use of 10.100 than to apply for 11 more network IDs. In this case, the third octet of the IP address can be used as a subnet ID, using the subnet mask 255.255.255.0, which splits this Class B address into 254 subnets: 10.100.1 through 10.100.254, each of which can have 254 nodes. Any of these network addresses could be assigned to the 12 international LANs in this example. Within each LAN, each computer is assigned a unique host ID, and they all have the subnet mask 255.255.255.0.

Note All systems connected to the same local area network must have the same subnet mask.

Host IDs 0 and 255 should not be assigned to a computer; they are used as broadcast addresses that are typically recognized by all computers.

Windows 98 provides three methods of IP addressing:

• DHCP addressing, in which DHCP servers such as computers running Windows NT Server 3.5 or later automatically assign IP addresses to Windows 98 clients.

• Automatic private IP addressing, in which the computer automatically assigns itself an IP address.

• Static IP addressing, in which network administrators manually configure IP addresses for all the computers on the network.

The following sections describe each of the three configurations.

Automatically Configuring IP Addresses with DHCP

In an effort to make implementing the TCP/IP protocol more manageable, Microsoft worked with other industry leaders to create an Internet standard called Dynamic Host Configuration Protocol (DHCP) for the automatic allocation of IP addresses. DHCP is not a Microsoft standard, but a public Request for Comments (RFC 1541) that Microsoft has implemented.

DHCP allows you to establish a range of valid IP addresses to be used per subnetwork. An individual IP address from the range is assigned dynamically to any DHCP client requesting an address. DHCP also allows you to establish a lease time that defines how long an IP address is to remain valid. Other configuration parameters can also be assigned using DHCP, such as subnet mask, DNS and WINS server identification, and so on.

A computer running Windows 98 cannot be a DHCP server. A DHCP server runs as a service on Windows NT Server 3.5 or later. If DHCP is available company-wide, users can move from subnet to subnet and always have a valid IP address. The IP Configuration utility (winipcfg) allows users or administrators to examine the current IP address assigned to the computer, the IP address lease time, and other useful data about the TCP/IP configuration.

When TCP/IP is installed, Windows 98 automatically enables the option to obtain an IP address from a DHCP server. You can disable this option if you want to manually enter IP addresses; however, in general you should not need to disable DHCP.

To disable DHCP

• You disable DHCP by configuring static IP addresses, following the instructions outlined in the section "Manually Configuring IP Addresses" later in this chapter.

If Microsoft TCP/IP is configured to obtain an IP address from a DHCP server when a DHCP server is not available on the network, an error message announces that the DHCP client could not obtain an IP address and your Windows 98 computer automatically assigns itself a private IP address (using automatic private IP addressing). When the DHCP server becomes available again, it will assign your computer an IP address.

Configuring IP Addresses Using Automatic Private IP Addressing

With Windows 98, Microsoft TCP/IP provides a new mechanism for automatic IP address assignment for simple LAN-based network configurations, called automatic private IP addressing. With automatic private IP addressing, DHCP clients can automatically assign themselves an IP address if a DHCP server is not present. This might happen, for example, on very small networks without a DHCP server, or on any size network if a DHCP server is temporarily down. The DHCP client can use B-node NetBIOS naming to assign the adapter a unique IP address from a special address space. These IP addresses must lie in the following range:

169.254.x.x 

These addresses are used only for private, internal addressing and are not valid for hosts that are visible on the global Internet.

After the adapter has been assigned an IP address, the computer can use the TCP/IP protocol to communicate with any other computer that is connected to the same LAN hub and that also uses automatic private IP addressing. However, the computer cannot communicate with computers on other subnets, or with computers that do not use automatic private IP addressing.

A Windows 98 computer that is configured for automatic private IP addressing can assign itself a private IP address if either of the following circumstances applies:

• A computer that is not configured as a laptop can assign itself a private IP address at boot time if the computer does not have a valid DHCP lease, and no DHCP server is found on the network.

• If the computer is configured as a laptop, the computer can assign itself a private IP address if no DHCP server is found on the network, regardless of whether it has a DHCP lease. This is useful, for example, for a computer that is sometimes located on a network with a DHCP server, and sometimes located on a network without a DHCP server.

In either case, if a DHCP server is later found, the computer stops using the private IP address and instead uses the IP address assigned by the DHCP service.

Automatic private IP addressing is automatically enabled. You might want to disable it in the following cases:

• Your network uses routers.

• Your network is connected to the Internet without a NAT or proxy server.

Unless you have turned DHCP messages off, DHCP messages inform you when you change between DHCP addressing and automatic private IP addressing. If you do accidentally turn DHCP messages off, you can turn them back on by changing the value of the registry entry PopupFlag from 00 to 01 in the following registry location:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VxD \DHCP 

You must reboot for the change to take effect.

You can also determine whether your computer is using automatic private IP addressing by using winipcfg, as the following procedure explains.

To determine whether automatic private IP addressing is currently enabled

• Click Start, click Run, type winipcfg, and then click More info. The resulting screen identifies your IP address and other information. Look at the box immediately beneath your adapter address. If the name of the box is "IP Autoconfiguration Address" and the IP address lies in the 169.254.x.x range, automatic private IP addressing is enabled. If, on the other hand, the name of the box is "IP Address," automatic private IP addressing is not currently enabled.

You can disable automatic private IP addressing in one of two ways:

• You can manually configure TCP/IP by following the procedure outlined in the section "Manually Configuring IP Addresses" later in this chapter. However, this method also disables DHCP.

• You can disable automatic private IP addressing (but not DHCP) by editing the registry.

You disable automatic private IP addressing but not DHCP by adding the IPAutoconfigurationEnabled registry entry with a value of DWORD 0x0 in the following registry location:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VxD \DHCP 

Use the Registry Editor to add this entry, then shut down and restart the computer.

Caution Using Registry Editor incorrectly can cause serious problems that may require you to reinstall Windows 98. Microsoft cannot guarantee that problems resulting from the incorrect use of Registry Editor can be solved. Use Registry Editor at your own risk.

You can change this registry entry for your entire network by creating an Msbatch.inf file. For more information about Msbatch.inf, see Chapter 3, "Custom Installations."

Manually Configuring IP Addresses

If you cannot use DHCP or automatic private IP addressing for automatic configuration, the network administrator must configure TCP/IP manually. Or, if custom setup scripts are used to install Windows 98, the correct values can be defined in the setup script. The required values include the following:

• The IP address and subnet mask for each network adapter installed on the computer.

• The IP address for the default gateways (hardware routers or servers running routing software with the ability to interpret IP addresses, usually used to find remote destinations).

• Whether the computer will use Domain Name System (DNS) and, if so, the IP addresses of the DNS servers on the internetwork.

• WINS server addresses, if WINS servers are available on the network.

The following procedure describes the basic configuration options for TCP/IP. If you want to configure the computer to use DNS or WINS for name resolution, see the procedures in "Using DNS for Name Resolution" later in this chapter and "Using WINS for Name Resolution" later in this chapter.

To configure the TCP/IP protocol manually

1. In Control Panel, double-click Network.

2. Double-click the instance of the TCP/IP protocol that is bound to your network adapter.

   Note If your computer has multiple network adapters, the list includes an instance of TCP/IP for each network adapter. You must configure each adapter with its own IP address, subnet mask, and gateway. There is one exception: you should configure TCP/IP settings for your dial-up adapter from your Dial-Up connection icon.

3. In the TCP/IP Properties dialog box, click the IP Address tab.

4. Select the Specify an IP address check box.

5. Type an IP address and subnet mask in the respective boxes.

   The network administrator must provide these values for individual users, based on the network ID and the host ID plan for your site.

   • The value in the IP Address box identifies the IP address for the local computer or, if more than one network adapter is installed, for the network adapter selected in the Configuration dialog box.

   • The value in the Subnet Mask box identifies the network membership and its host ID for the selected network adapter. The subnet mask defaults to an appropriate value, as shown in the following list.

     

6. To view or specify which network clients are bound to the TCP/IP protocol, click the Bindings tab.

   • To keep a network client from using the TCP/IP protocol, make sure the check mark beside the client name is cleared.

   • If the network client for which you want to use TCP/IP does not appear in this list, that client might not be installed properly. Return to the Configuration tab in the Network dialog box and reinstall that network client.

   Note The only network client provided with Windows 98 that can use Microsoft TCP/IP is Client for Microsoft Networks. Client for NetWare Networks does not use Microsoft TCP/IP.

   NetWare/IP from Novell allows the NetWare Core Protocol (NCP) request to be sent over an IP header. You can use NetWare/IP only with a Novell-provided client.

7. Click the Gateway tab. Type at least one IP address for the default gateway (IP router) on the network, and then click Add.

8. To specify an IP address for an additional gateway, type the IP address in the New gateway box, and then click the Add button.

   The first gateway in the list is the default gateway, which is used to reach destinations on remote networks. Gateway addresses can be prioritized by dragging the IP address in the list of installed gateways. Windows 98 attempts to connect to other gateways only if the primary gateway is unavailable.

9. Click OK.

10. Restart the computer for changes to take effect.

If you are using Dial-Up Networking to connect to the Internet, you can manually configure DNS and IP addresses for each connection that you define. For more information about defining IP addresses for each connection and about IP addresses on TCP/IP networks, see Chapter 20, "Internet Access and Tools." For more information about TCP/IP registry entries, see Chapter 31, "Windows 98 Registry."

Configuring Multihoming

When a computer is configured with more than one IP address, it is referred to as a multihomed system. With Windows 98, network administrators create multihomed configurations for purposes such as the following:

• To connect a computer simultaneously to the local network and (through Dial-Up Networking or PPTP) to the Internet.

• To connect to different subnets when the computer is on a single physical network that contains multiple logical IP networks.

• To allow a computer to communicate simultaneously on two otherwise isolated networks.

For example, in the most common configuration, a computer might have a LAN connection and a dial-up (PPP) or Point-to-Point Tunneling Protocol (PPTP) connection to the Internet or another IP network.

For more information about this configuration, see Chapter 19, "Remote Networking and Mobile Computing."

As another example, a computer on the local area network at a branch office that also requires a network connection to certain computers on the corporate enterprise network might require a multihomed configuration. Such a configuration might be used, for example, by financial or human-resources personnel who must access servers on a private subnet.

Figure 15.3 Multihomed Windows 98 computer connected to two separate networks

This section provides information about configuring multihomed computers running Windows 98.

Note Multihoming on a Windows 98 computer is suitable only as a limited solution for connecting a single computer to other networks. For corporate computing environments that require routing among subnetworks, you should use dedicated routers such as a Cisco routers, UNIX hosts, or a computer running Windows NT Server 3.5 or later.

Configuring TCP/IP on Multihomed Computers

You can configure multihoming by using multiple network adapters on a single computer, multiple media, or multiple IP addresses for a single network adapter.

For a multihomed computer that uses multiple network adapters for physical connections to the LAN, or a dial-up adapter for remote access, there is an instance of Microsoft TCP/IP for each adapter in the Network option in Control Panel.

Configuring multiple network adapters per physical network. 

Windows 98 places no restrictions on such configurations, so you can add as many network adapters as the computer hardware can accommodate, and assign each a separate address. However, you can only have six instances of TCP/IP and four instances of IPX/SPX installed on the system.

To configure a multihomed system using multiple network adapters

• Add TCP/IP configuration information for each network adapter in the Network option in Control Panel, either by using DHCP or by entering information for static IP addresses.

Caution You should not use this method to configure a multihomed system for a Dial-Up or virtual private networking connection. When Dial-Up Networking or virtual private networking are installed, adapters are automatically added and appear automatically in the Network option in Control Panel.

Configuring multiple networks and media types. 

You can have only six instances of TCP/IP and four instances of IPX/SPX. This means that only six adapters can bind to TCP/IP and only four to IPX/SPX. Otherwise, there are no restrictions for this type of configuration other than hardware and media support. Microsoft TCP/IP supports the following:

• Ethernet (and 802.3 SNAP).

• Token-ring (802.5).

• Fiber distributed data interface (FDDI), using NDIS 2 real-mode network adapter drivers (only a few protected-mode adapter drivers are available for FDDI supporting TCP/IP).

• WAN, using switched virtual-circuit wide-area media such as ISDN, X.25, and dial-up or dedicated asynchronous lines.

Configuring multiple IP addresses per network adapter. 

This option is not supported through the Network option in Control Panel. Additional addresses can be added directly in the registry. However, this is not a recommended method for configuring TCP/IP. NetBIOS name registration using NetBIOS over TCP/IP supports only one IP address per network adapter. Moreover, if DHCP is enabled for configuring TCP/IP, only one DHCP-assigned address can be provided per adapter.

Issues When Implementing Multihoming

If TCP/IP is configured for multiple network adapters, or for both LAN and Dial-Up connections, you must consider the following issues:

A unique IP address and subnet mask are defined for each adapter. 

For each network adapter or dial-up adapter, an instance of TCP/IP is bound to the adapter. You can choose to have IP addresses dynamically assigned by DHCP or automatic private IP addressing or defined manually as static addresses.

Domain Name System (DNS) configuration settings are global. 

The settings on the DNS Configuration tab in TCP/IP properties are used for all adapters on the computer. For example, if you change the DNS settings in the TCP/IP properties for the dial-up adapter to enable DNS, then DNS is also enabled for every LAN adapter on the computer.

However, the settings on the WINS Configuration tab are used only for the adapter you are configuring. If you enable the option named Use DHCP For WINS Resolution for a LAN adapter, this option is enabled only for that adapter, not for the dial-up adapter or for other LAN adapters on the computer.

Therefore, for a multihomed computer, you must carefully define options for DNS that are applicable for all adapters using TCP/IP. Usually, this means that you want the following:

• If you want to use DNS or the Hosts file for name resolution with any TCP/IP connection, then make sure DNS is enabled.

  Note In the case when WINS is used for name resolution on the network (including using WINS servers for host-name resolution) and when DNS is used only for dial-up connections, it is easy for a user to mistakenly disable DNS in the TCP/IP properties for the LAN adapter, without realizing that this also disables DNS for the remote connection. The result would be an inability to find remote hosts when using the remote connection.

The default gateway can be different for each adapter. 

For multiple physical connections to the WAN, you can assign a different default gateway for each network connection; however, a Windows 98 computer uses only one default gateway at a time. The default gateway for the dial-up adapter is assigned by the access provider. You can also assign different gateways for a Dial-Up Networking connection by using the Make New Connection Wizard.

Only one default gateway is used at a time. 

Although you can have a different default gateway for each adapter, Windows 98 uses only one default gateway at a time. This means that only certain hosts are reachable:

• Hosts on the local subnet

• Hosts that are reachable by the default gateway

As a result, in some cases you may lose network connectivity. For example, suppose your computer is first connected to the corporate TCP/IP network and you make a PPP dial-up connection to the Internet. Your computer stops using the default gateway that connects your computer to the corporate network and instead uses the default gateway that connects your computer to the Internet. Therefore, you can reach hosts on your local subnet, but you cannot reach other hosts on your network.

For a detailed example of a computer changing default gateways, see Chapter 19, "Remote Networking and Mobile Computing."

If you want to restore connectivity to the corporate network, you can use one of the following three methods:

• You can manually add routes to destinations on your local network. For information about adding routes, see "Using the Route Table" later in this chapter.

• In the properties for your dial-up connection, you can deselect the option named Use default gateway on remote network, then manually add routes to dial-up destination. For information about how to do this, see Chapter 19, "Remote Networking and Mobile Computing."

• If any of the routers on your corporate network uses the Routing Information Protocol (RIP), you can configure your computer to listen in on those broadcasts. For information about this configuration, see "Support for RIP Listening" later in this chapter.

Using the Route Table

The route table controls where IP packets are sent. It is maintained automatically in most cases, but in some cases you may want to manually add a route. This section describes the route table and explains how to add routes.

To see the route table for your computer, at the command prompt type route print.

Note You can reach the command prompt from the Start menu, either by clicking Run or by clicking Programs, and then clicking MS-DOS Prompt.

The following table is a sample route table from a single-homed computer.

Table 15.3 Route table sample 

Table 15.3 shows a computer with the IP address 172.26.34.232. The table contains the following seven entries:

1. The first line is the default route. This is the route to which the computer sends IP packets if the other route entries do not specify where to send them.

2. The second line is the loopback address. This is the address a host uses to send packets to itself. The loopback address is always 127.0.0.0, and the netmask is always 255.0.0.0.

3. The third line is a network route.

4. The fourth line is a host route for the local host (the route for this host computer).

5. The fifth line is the subnet broadcast address.

6. The sixth line is the IP multicast address. This is the address the computer sends packets to in order to reach an IP multicast group.

7. The seventh line is for limited broadcast address. This is the address a host uses to reach all other addresses on the subnet.

Network Address

The network address in the route table is the destination address. The network address column can have four different types of entries, listed here in the order in which they are searched for a match.

1. Host address (a route to a single, specific destination IP address).

2. Subnet address (a route to a subnet).

3. Network address (a route to an entire network).

4. Default gateway (a route used when there is no other match).

If no match is found, the packets are discarded.

Netmask

The netmask defines which portion of the network address must match in order for that route to be used. When the mask is written in binary, a 1 is significant (must match) and a 0 need not match.

For example, the mask of all 255s (all 1s) means that the destination address of the packet to be routed must exactly match the network address in order for this route to be used. For another example, the network address 172.20.232.0 has a netmask of 255.255.255.0. This netmask means that the first three octets must match exactly, but the last octet need not match.

Gateway Address

The gateway address is where the packet must be sent. This can be the local network card or the address of a gateway (router) on the local subnet.

Interface

The interface is the address of the network card over which the packet should be sent. 127.0.0.1 is the software loopback address.

Metric

The metric is the number of hops to the destination. Anything on the local subnet is one hop, and each router crossed after that is an additional hop. The metric is used to determine the best route.

Configuring Routes for a Multihomed Computer

If your computer is multihomed and has connections to two separate IP networks, such as the corporate network and the Internet, the default gateway for only one network is used. For the computer to be able to communicate with the other network, routes must be added to the route table. This can be accomplished in one of two ways:

1. If a router on the network sends RIP broadcasts and your computer is configured to listen to RIP broadcasts, routes will be added automatically. (For more information about configuring your computer to listen to RIP broadcasts, see the section "Support for RIP Listening" later in this chapter.)

2. You can manually add static routes to the route table.

   To add static routes, use the following format:

   Route add [subnet] mask [netmask] [gateway] metric [metric]
   

The following is an example route:

<pre IsFakePre="true" xmlns="https://www.w3.org/1999/xhtml">Route add 172.20.255.0 mask 255.255.255.0 172.20.234.232 metric 2

The route in this example means that to get to the subnet 172.20.255.0 with a mask of 255.255.255.0, use gateway 172.20.234.232, and that the gateway is 2 hops away.

Configuring TCP/IP Name Resolution

Computers use IP addresses to identify each other, but users usually find it easier to work with computer names. A name resolution mechanism must be available on a TCP/IP network to resolve names to IP addresses.

Windows 98 provides several different types of name resolution, including DNS, WINS, broadcast name resolution, and name resolution using Hosts or LMHosts files. Generally, a Windows 98 computer uses a combination of these name resolution types, summarized in this section.

Domain Name System name resolution. Domain Name System (DNS) is a global, distributed database based on a hierarchical naming system. DNS name resolution is used on the Internet to map friendly names to IP addresses, and vice versa. Notice that DNS replaces the functionality of the Hosts file. For more information, see "Using DNS for Name Resolution" later in this chapter.

Windows Internet Naming Service. Windows Internet Naming Service (WINS) name resolution provides static and dynamic mapping of names to IP addresses. (This contrasts with DNS name resolution, which provides only static mapping.) Computers running Microsoft TCP/IP can use WINS if one or more Windows NT Server computers configured as WINS servers are available. WINS can be used in conjunction with broadcast name resolution for an internetwork, where other name resolution methods are inadequate. Notice that WINS is a dynamic replacement for the LMHosts file. For more information, see "Using WINS for Name Resolution" later in this chapter.

Broadcast name resolution. Computers running Microsoft TCP/IP can use local broadcast name resolution, which is a NetBIOS-over-TCP/IP mode of operation defined in RFC 1001/1002 as b-node. It is restricted to only one subnet. This method relies on a computer making IP-level broadcasts to register its name by announcing it on the network. Each computer in the broadcast area is responsible for challenging attempts to register a duplicate name and for responding to name queries for its registered name.

Hosts or LMHosts files. Hosts and LMHosts files, also called host tables, are files that Windows 98 can use for local name resolution if other methods are not available. An LMHosts file specifies the NetBIOS computer name and IP address mappings. When WINS in not available, it is used as a WINS equivalent to resolve NetBIOS names to IP addresses. Likewise, a Hosts file specifies the DNS name and IP address. It is used as a local DNS equivalent to resolve host names to IP addresses. You must manually enter the name-to-IP address mappings in Hosts and LMHosts files. For more information about creating and editing Hosts and LMHosts files, see Appendix F, "Hosts and LMHosts Files for Windows 98."

Windows 98 provides support for multiple DNS servers and up to twelve WINS servers. Support for either service can be configured automatically from a DHCP server or after Windows 98 Setup by using the Network option in Control Panel.

Note To ensure that both the name and the address are unique, the computer using Microsoft TCP/IP registers its name and IP address on the network during system startup.

Using DNS for Name Resolution

This section provides an overview of DNS, then describes how to configure your Windows 98 client to use DNS.

Understanding DNS

Although TCP/IP uses IP addresses to identify and reach computers, users typically prefer to use host names. For example, users prefer the friendly name ftp.terrafirminc.tld instead of its IP address 172.16.23.55. Domain Name System, defined in Requests for Comments (RFCs) 1034 and 1035, is the naming service used on the Internet to provide standard naming conventions for IP computers.

Although DNS may seem similar to WINS, there are two major differences. First, DNS requires static configuration of IP addresses for name-to-address mapping. WINS, on the other hand can provide name-to-address mapping dynamically and requires far less administration.

Second, whereas WINS uses a flat name space (all names are located in the same domain), DNS uses a tree structure called the domain name space, where each node or domain is named and can contain subdomains. DNS is a global, distributed database based on a hierarchical naming system. The naming system was developed to provide a method for uniquely identifying hosts on the Internet.

In Windows 98, the DNS name consists of two parts — the domain name and the host name—known together as the fully qualified domain name (FQDN). For example, using the fictional domain name of Terrafirminc, an FQDN for a workstation in the nursery division could be jeff.nursery.terrafirminc.tld. Note that the DNS name can actually be multipart with a period (.) separating each part. Also note that the host portion of the name, jeff, is analogous to a NetBIOS computer name.

Note In Windows 98, a computer's globally known system name is its host name (for example, jeff), appended with a DNS domain name (for example, nursery.terrafirminc.tld). The host name defaults to the computer name (NetBIOS name) defined during Windows 98 Setup. The default name can be changed in the DNS Configuration tab when you are configuring TCP/IP properties.

The top-level domains were assigned organizationally and by country. These domain names follow the International Standard 3166. Two-letter and three-letter abbreviations are used for countries, and various abbreviations are reserved for use by organizations, as shown in Table 15.4.

Table 15.4 DNS domain names 

DNS uses a client/server model, where the DNS servers contain information about a portion of the DNS database and make this information available to clients, called resolvers, that query the name server across the network. DNS name servers are programs that store information about parts of the domain name space called zones. The administrator for a domain sets up name servers that contain the database files with all the resource records describing all hosts in their zones. DNS resolvers are clients that use name servers to gain information about the domain name space.

All the resolver software necessary for using DNS on the Internet is installed with Microsoft TCP/IP. Microsoft TCP/IP includes the DNS resolver functionality used by NetBIOS over TCP/IP and Windows Sockets connectivity applications such as File Transfer Protocol (FTP) and Telnet to query the name server and interpret the responses.

The key task for DNS is to present friendly names for users and then resolve those names to IP addresses, as required by the internetwork. If a local name server does not contain the data requested in a query, it then queries the other name servers until it finds the specific name and address it needs. This process is made faster because name servers continuously cache the information learned about the domain name space as the result of queries.

Configuring DNS

You need to determine whether users should configure their computers to use DNS. Usually you will use DNS if you are using TCP/IP to communicate over the Internet or if your private internetwork uses DNS to distribute host information. You can use DNS either instead of or in conjunction with WINS, which is described in the next section.

You also need to determine whether or not you want to use DHCP for automatic configuration. (For information about circumstances in which you might want to use DHCP, see "Planning for TCP/IP" earlier in this chapter.) If you use DHCP for automatic configuration, a DHCP server can automatically configure the list of DNS servers the client should use. If you do not use DHCP, however, you will need to configure these parameters yourself.

Using DNS with WINS Lookup

If you are already using WINS servers for local name resolution but you want to make those servers visible to the global Internet, you can do so without configuring configure DNS on each computer. Windows NT Server 4.0 and later includes a feature called DNS with WINS Lookup. If you configure a Windows NT Server 4.0 and later to use DNS with WINS Lookup, that Windows NT Server can query the WINS servers for the "friendly names" of your computers, then use the information to construct a name that can be used on the global Internet.

For example, suppose the network terrafirminc.tld contains a Windows 98 computer named Annuals. On the internal network, the Windows NT Server is using DHCP to automatically assign IP addresses and WINS to resolve those IP addresses to the computer's "friendly names." You want Internet users to be able to connect to Annuals using the fully qualified domain name annuals.terrafirminc.tld. If you have Windows NT Server 4.0 or later, you can configure Microsoft DNS server on that computer to query WINS for the IP address for Annuals. WINS returns the IP address, and DNS then successfully resolves the IP address to the FQDN annuals.terrafirminc.tld.

For more information about this configuration, see the Microsoft Windows NT Resource Kit (for Microsoft Windows NT Server version 4.0). 

This section describes how to configure a Windows 98 computer to use DNS for name resolution. You only need to follow the procedures outlined in this section if you are not using DHCP to configure name resolution. If you are using DHCP to configure name resolution, you need only configure DHCP by following the steps in "Automatically Configuring IP Addresses with DHCP" earlier in this chapter.

Figure 15.4 shows sample DNS configuration settings.

Figure 15.4 Sample DNS configuration settings 

To configure a computer to use DNS for name resolution

1. In Control Panel, double-click Network.

2. Double-click the TCP/IP protocol that is bound to your network adapter.

3. In the TCP/IP Properties dialog box, click the DNS Configuration tab.

4. If a DNS server is available, click Enable DNS. Then specify a host name and complete the other configuration information as described in the following procedure.

Tip You must enable DNS on each computer that needs to use Hosts for name resolution.

The host name is used to identify the local computer for authentication by some utilities. Other TCP/IP-based utilities can use this value to learn the name of the local computer. Host names are stored on DNS servers in a table that maps names to IP addresses for use by DNS.

To set the host name for DNS

• Type a name in the Host box.

  The name can be any combination of the letters A through Z, the numerals 0 through 9, and the hyphen (-), plus the period (.) character used as a separator. By default, this value is the Microsoft networking computer name, but the network administrator can assign another host name without affecting the computer name.

Note Some characters that can be used in computer names, especially the underscore, cannot be used in host names. This is a limitation imposed in the character set defined by the Internet Engineering Task Force (IETF) standard for DNS. For more information, see RFC 1123.

To set the DNS domain name

• Optionally, type a name in the Domain box.

  This is usually an organization name followed by a period and an extension that indicates the type of organization, such as microsoft.com. The name can be any combination of the letters A through Z, the numerals 0 through 9, and the hyphen (-), plus the period (.) character used as a separator.

This DNS domain name is appended to the host name (or short name) to create the fully qualified domain name (FQDN) for your computer. If you have not specified a domain in the Domain Suffix Search Order and you are querying for a short name, then this DNS domain name is appended to your query.

Note A DNS domain is not the same as a Windows NT or LAN Manager domain. A DNS domain is a hierarchical structure for organizing TCP/IP hosts and provides a naming scheme used on the Internet. A Windows NT or LAN Manager domain is a grouping of computers for security and administrative purposes.

You can add up to three IP addresses for DNS servers. For a given DNS query, Windows 98 attempts to get DNS information from the first IP address in the list. If no response is received, Windows 98 goes to the second server in the list, and so on. To change the order of the IP addresses, you must remove them and retype them in the order that you want the servers to be searched.

To set the DNS server search order

1. In Control Panel, double-click Network.

2. Double-click the TCP/IP protocol that is bound to your network adapter.

3. In the TCP/IP Properties dialog box, click the DNS Configuration tab.

4. In the DNS Server Search Order box, type the IP address of a DNS server that will provide name resolution. Then click the Add button to add the IP address to the list.

   The network administrator should provide the correct values for this parameter, based on the IP address assigned to the DNS server used at your site.

5. To remove an IP address from the list, select it, and then click the Remove button.

Note If you have two servers listed in this dialog box, Windows 98 checks the second server only if no response is received from the first server. If Windows 98 attempts to check a host name with the first server and receives a message that the host name is not recognized, the system does not try the second DNS server.

The Domain Suffix Search Order specifies the DNS domain suffixes to be appended to short names during name resolution. For example, with the DNS configuration settings shown in Figure 15.4, the domain name acct02.terrafirminc.tld would be appended to the short name to create an FQDN. However, if the Domain Suffix Search Order list is blank, the domain name terrafirminc.tld would be appended instead.

You can add up to five domain suffixes. Place domain suffixes in the list in the order you want them to be searched.

To set the domain suffix search order

1. In Control Panel, double-click Network.

2. Double-click the TCP/IP protocol that is bound to your network adapter.

3. In the TCP/IP Properties dialog box, click the DNS Configuration tab.

4. In the Domain Suffix Search Order box, type the domain suffixes to add to your domain suffix search list, and then click the Add button.

5. To remove a domain name from the list, select it, and then click the Remove button.

Using WINS for Name Resolution

Windows Internet Naming Service (WINS) is a service that runs on Windows NT Server to optimize NetBIOS name resolution. It provides a distributed database for registering and querying dynamic computer name-to-IP address mappings in a routed network environment. You can use WINS either alone or in conjunction with DNS.

WINS reduces the use of local broadcasts for name resolution and allows users to locate computers on remote networks automatically. Furthermore, when dynamic addressing through DHCP results in new IP addresses for computers that move between subnetworks, the changes are updated automatically in the WINS database. Neither the user nor the network administrator needs to make manual accommodations for name resolution in such a case.

WINS consists of two components: the WINS server, which handles name queries and registrations, and the client software (NetBIOS over TCP/IP), which queries for computer name resolution. A WINS server is a Windows NT Server 3.5 or later computer with WINS server software installed. When Microsoft TCP/IP is installed under Windows 98, WINS client software is installed automatically.

On a Windows-based network, users can browse transparently across routers. To allow browsing without WINS, you must ensure that the users' primary domain has Windows NT Server computers on both sides of the router to act as master browsers. These computers need to contain correctly configured LMHosts files with entries for the domain controllers across the subnet.

With WINS, such strategies are not necessary, because the WINS servers and proxies provide the support necessary for browsing Windows NT domains across routers. For a technical discussion of how WINS works and how it can be set up on the network, see Windows NT Server 4.0 TCP/IP in the Windows NT Server 4.0 documentation set.

If there are WINS servers installed on your network, you can use WINS in combination with broadcast name queries to resolve NetBIOS computer names to IP addresses. If you do not use this option, Windows* *98 can use name query broadcasts (b-node mode of NetBIOS over TCP/IP) plus the local LMHosts file to resolve computer names to IP addresses. Broadcast resolution is limited to the local network, as described earlier in this section.

If DHCP is used for automatic configuration, these parameters can be provided by the DHCP server. Otherwise, you must configure information about WINS servers manually. WINS configuration is global for all network adapters on a computer.

Configuring WINS

The following procedure describes how to configure WINS and how to enable DHCP.

To configure a computer to use WINS for name resolution

1. In Control Panel, double-click Network.

2. Double-click the TCP/IP protocol that is bound to your network adapter.

3. In the TCP/IP Properties dialog box, click the WINS Configuration tab.

4. If a DHCP server is available that is configured to provide information on available WINS servers, select the Use DHCP For WINS Resolution check box.

   – Or –

   If a WINS server is available but not a DHCP server, select Enable WINS Resolution and type the IP addresses of the Primary and Secondary WINS servers. These values should be provided by the network administrator, based on the IP addresses assigned to these Windows NT Server computers.

5. If WINS is enabled, in the Scope ID box, type the computer's scope identifier, if required on an internetwork that uses NetBIOS over TCP/IP.

   Usually this value is left blank. Scope IDs are used only for communication based on NetBIOS over TCP/IP. In such a case, all computers on a TCP/IP internetwork must have the same scope ID. A scope ID can be assigned to a group of computers if those computers communicate only with each other and not with computers outside the group. Such computers can find each other if their scope IDs are identical.

Enabling WINS Persistent Connection Attempts

Windows 98 includes an enhancement to the WINS client (called NetBIOS over TCP/IP, or NetBT), a session-layer network service that performs name-to-IP address mapping for name resolution. (For more information, see the following section, "WINS Technical Notes.") In Windows 95, NetBT queried only one WINS server, and returned a failure for a setup session attempt if the first IP address it tried to use failed to establish a session. This caused a problem because replication delays between two or more WINS servers on the network can cause WINS servers to return stale IP addresses when queried.

You can configure Windows 98 to work around this problem by configuring NetBT to continue querying multiple WINS servers if it failed to establish the initial session. Thus, it first queries the primary WINS server for an IP address and tries to establish a connection using that IP address. If this fails, it tries to get another IP address from the next WINS server, or by using broadcast name resolution. It will continue to query WINS servers until either all of the WINS servers specified have been queried or a connection is established.

To enable WINS persistent connection attempts, set the registry entry TryAllNameServers (String data type) to a value of 1. TryAllNameServers is found in the following registry key:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \Vxd \MSTCP 

WINS Technical Notes

The WINS protocol is based on and is compatible with the protocol defined for WINS server in Requests for Comments (RFCs) 1001 and 1002, so it is interpretable with any other implementations of these RFCs.

Microsoft TCP/IP uses NetBIOS over TCP/IP as specified in RFCs 1001 and 1002, which define a software interface that supports name resolution for NetBIOS client and server programs in the WAN environment.

RFCs 1001 and 1002 define the following four node types:

• B-node, which uses broadcasts to resolve names.

• P-node, which uses point-to-point communications with a NetBIOS server to resolve names.

• M-node, which uses broadcasts first (b-node), then name queries (p-node) if broadcasts are not successful.

• H-node, which uses name queries first (p-node), and then uses broadcasts (b-node) if the name server is unavailable or if the name is not registered in the WINS database.

If WINS is enabled on a Windows 98 computer, the system uses h-node by default. Without WINS, the system uses b-node by default. Non-WINS clients can access WINS through proxies, which are WINS-enabled computers that listen to name query broadcasts and then respond for names that are not on the local subnet or are h-nodes.

To see which node type is configured on a Windows 98 computer

1. Click Start, click Run, and then, type winipcfg.

2. Click the More info button.

3. In Host Information, look at the Node type box.

Viewing Your WINS and DNS Configuration

Regardless of how you configure your computer to use name resolution, you can view your settings using the winipcfg utility. For instructions on using winipcfg, see "Using Diagnostic Utilities to Troubleshoot," later in this chapter.

Technical Notes on TCP/IP

This section discusses technical information about TCP/IP. First it describes the enhancements to TCP/IP in Windows 98. Next, it lists the supported standards. Finally, it lists and briefly describes TCP/IP command-line utilities.

For information about using these utilities to troubleshoot TCP/IP problems, see "Troubleshooting TCP/IP" later in this chapter.

Enhancements to TCP/IP in Windows 98

Microsoft TCP/IP in Windows 98 includes many enhancements over the version of Microsoft TCP/IP that was included in Windows 95. It includes performance enhancements, support for Windows Sockets 2 and multimedia applications, and greatly improved support for IP multicast. Additionally, it includes RIP listening, a feature that can improve network connectivity for multihomed computers in large corporate networks. This section describes some of the enhancements to Microsoft TCP/IP in Windows 98.

Support for Windows Sockets 2

In Windows 98, Windows Sockets 2 is the preferred application programming interface (API) for TCP/IP socket programming.

Microsoft TCP/IP contains the following enhancements to support Windows Sockets 2:

• Support for Resource Reservation Protocol (RSVP), and Quality of Service (QoS). For more information, see the next section.

• Support for IP multicast, as described in Request for Comment (RFC) 988. This includes support for the Internet Draft for IGMP version 2 and for RFC 1256, ICMP Router Discovery. For more information, see "Support for IP Multicast" later in this section.

• Support for raw sockets, enabling application developers to write applications that directly access Windows Sockets 2. Raw sockets can be used if an application does not want to use either UDP or TCP but instead wants to encapsulate its data in an IP packet to send it.

For more information about Windows Sockets 2, see Chapter 29, "Windows 98 Network Architecture." For information about troubleshooting setup problems associated with Windows Sockets 2, see Chapter 5, "Setup Technical Discussion." For information about incompatibilities between Windows Sockets 2 and third-party networking clients, see Chapter 17, "Windows 98 on Third-Party Networks."

Support for Quality of Service and Resource Reservation Protocol

Microsoft TCP/IP supports the Generic Quality of Service (GQoS) and Resource Reservation Protocol (RSVP) APIs.

The GQoS APIs let applications request certain characteristics for a network connection. The GQoS APIs can request attributes such as the following:

• Peak bandwidth (average or peak bit rate available).

• Latency (the maximum acceptable delay between transmission of a bit and its receipt by the receiver).

• Delay variation (the difference between a packet's minimum and maximum delay).

RSVP is a signaling protocol that is used to establish connections with the requested GQoS characteristics. It handles QoS requests, reserving network bandwidth when possible and when requested, then ensuring that the network can provide that bandwidth.

GQoS is available only if you have the most recent version of Microsoft TCP/IP (which includes Windows Sockets 2).

For more information about the RSVP and QoS architecture, see Chapter 29, "Windows 98 Network Architecture." . For more information about RSVP, see also the Internet Engineering Task Force (IETF) RSVP specification at https://www.ietf.org .

For more information about installing Microsoft TCP/IP, see the section "Installing Microsoft TCP/IP" earlier in this chapter.

Support for IP Multicast

With Windows 98, Microsoft TCP/IP supports RFC 1112 IP Multicast, or the transmission of IP packets to a group of zero or more hosts in a multicast group. Windows 98 computers can create, join, and leave multicast groups, and they can send IP packets to groups they belong to.

Microsoft TCP/IP also supports the Internet Draft for IGMP version 2. IGMP version 2 specifies a way for hosts to quickly report termination of multicast group membership. This is useful for low-bandwidth connections. For example, suppose a user is listening to a radio broadcast over a slow dial-up link and wants to change the channel to listen to a different radio station. With IGMP 2, the computer can send a message to the upstream multicast router to stop forwarding the group's packets.

Support for ICMP Router Discovery

Microsoft TCP/IP supports Internet Control Message Protocol (ICMP) Router Discovery, described in RFC 1256. ICMP Router Discovery enables hosts attached to broadcast networks to learn IP addresses of neighboring routers.

Performance Enhancements

With Windows 98, Microsoft TCP/IP includes several features that improve network performance. These features include the following:

• Support for TCP Large Windows and time stamps

• Support for Selective Acknowledgments

• Support for Fast Retransmission and Recovery

Support for TCP Large Windows

Windows 98 TCP/IP supports TCP Large Windows (TCPLW) and time stamps as documented in RFC 1323. Time stamps enable computers to measure round trip times and to reject old duplicate packet segments. TCPLW and time stamps are useful for networks that have high bandwidth and high delay, such as high-speed transcontinental connections or satellite links.

Support for TCPLW and time stamps are enabled by default. By default, TCPLW will be used if an application requests a Windows Sockets 2 socket to use a buffer size greater than 64 KB or if a DefaultRcvWindow size of more than 64 KB is used in the registry. To change the default value, you must add the registry entry Tcp1323Opts to the following location:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VXD \MSTCP \Parameters 

The registry entry Tcp1323Opts is a string value type. The values for the key are:

0 - No window scaling and time stamp options1 - Window scaling options but no time stamp options2 - Time stamp but no window scaling options3 - Window scaling and time stamp options

Support for Selective Acknowledgments 

Windows 98 TCP supports Selective Acknowledgments (SACK) as documented in RFC 2018. Selective Acknowledgments allows TCP to recover from IP packet loss without resending packets that were already received by the receiver. Selective Acknowledgments is most useful in combination with TCPLW and time stamps. SACK support is enabled by default. To disable SACK support, you must add the registry entry SackOpts to the following location:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VXD \MSTCP \Parameters 

The registry entry SackOpts is a string value type. The values for the entry are:

0 - No SACK options1 - SACK option enabled

The default value is 1 - SACK option enabled.

Support for Fast Retransmission and Fast Recovery 

Windows 98 TCP/IP supports Fast Retransmission and Fast Recovery on TCP connections that are incurring IP packet loss. These mechanisms allow a TCP sender to quickly infer a single packet loss and resend the packet after receiving duplicate acknowledgments for a previously sent and acknowledged TCP/IP packet. This mechanism is useful when the network is intermittently congested and intervening routers or switches are dropping packets.

By default, the sender must receive three duplicate acknowledgments before resending the last unacknowledged TCP/IP packets. You can change this default value by adding a registry entry to the following location:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VXD \MSTCP \Parameters 

The registry entry MaxDupAcks is data type DWORD and accepts integer values from 2 to N. The value data of MaxDupAcks indicates the number of duplicate acknowledgments the sender receives before resending the last unacknowledged TCP/IP packet.

Support for RIP Listening

In some large networks, a computer might be configured as a multihomed system and connected to two different networks. Figure 15.3 (earlier in this chapter) shows this configuration.

To communicate with both networks at the same time, the computer must be able to communicate with both routers. However, as the section "Configuring Multihoming" described, Windows 98 computers can use only one default gateway at a time.

There is a solution to this problem. If one or both of the routers uses the Routing Information Protocol (RIP) to send routing information, the computer can be configured to "listen in" to the RIP messages. Your computer can learn other routes on the network, by listening to RIP messages, and then add their IP addresses to the route table. Thus, you do not need to manually add routes to the route table. This process is called RIP listening or silent RIP. Figure 15.5 shows an example of a multihomed host that uses RIP listening.

Figure 15.5 Multihomed host using RIP listening 

Router 2 sends RIP messages, and the Windows 98 computer listens in on those messages. Router 1 does not send RIP messages, so the Windows 98 computer is configured to use Router 1 as the default gateway. Thus, the Windows 98 computer can communicate with hosts on both networks.

The following procedure describes how to enable RIP listening.

To enable RIP listening

1. In Control Panel, double-click Network, and then click Add.

2. In Select Network Component Type, select Service, and then click Add.

3. Click Have Disk, and then type the location of the RIP listening files. They are located on the Windows 98 CD, in the \Tools\RIP directory. Select the file Irip.inf, and then click OK.

You must restart your computer for the changes to take effect.

Supported Standards

TCP/IP standards are defined in Requests for Comments (RFCs) published by the Internet Engineering Task Force (IETF) and other working groups. Table 15.5 lists the RFC standards supported by Microsoft TCP/IP. To find these standards, visit https://www.ietf.org/ . 

Table 15.5 Request for Comments standards supported 

TCP/IP Utilities

The TCP/IP utilities offer network connections to non-Microsoft hosts such as UNIX system computers. You must have the TCP/IP network protocol installed to use the TCP/IP utilities. The tools listed in Table 15.6 are installed automatically when you install Microsoft TCP/IP.

Table 15.6 Tools installed with TCP/IP 

Important The FTP and Telnet utilities rely on password authentication by the remote computer. Passwords are not encrypted before being sent over the network. This allows another user equipped with a network analyzer on the same network to steal a user's remote account password. For this reason, it is strongly recommended that users of these utilities choose different passwords for their workgroups, computer, or domain from the passwords used when connecting to computers that are not on Microsoft networks.

The following describes most of the TCP/IP commands included with Windows 98. For information about the syntax of those commands, use the help provided in the MS-DOS virtual machine.

For information about winipcfg, see "Using Diagnostic Utilities to Troubleshoot," later in this chapter.

To get help on TCP/IP utilities

• At the command prompt, type the command name followed by a space and -?. For example, type arp -? to get help on the arp command. For some commands, you do not need to type the dash and question mark.

Arp

This diagnostic command displays and modifies the IP-to-Ethernet or IP-to-token-ring address translation tables used by the Address Resolution Protocol (ARP).

Ftp

This connectivity command transfers files to and from a computer running a File Transfer Protocol (FTP) service. The ftp command can be used interactively or by processing ASCII text files.

To use the ftp command

• At the command prompt, type ftp plus any desired switches and press ENTER.

  For example, you might type ftp -s:myfile.scr 

To get help with the ftp command

• From within FTP, type help command, where command is the name of the command you need help with. For a list of commands, simply type help.

Table 15.7 shows the FTP commands available when Microsoft TCP/IP is installed on a computer.

Table 15.7 FTP commands in Microsoft TCP/IP 

Nbtstat

This diagnostic command displays protocol statistics and current TCP/IP connections using NetBIOS over TCP/IP.

Notes on Nbtstat 

The column headings generated by the nbtstat utility have the following meanings.

Input

Number of bytes received.

Output

Number of bytes sent.

In/Out

Whether the connection is from the computer (outbound) or from another system to the local computer (inbound).

Life

The remaining time that a name table cache entry will live before it is purged.

Local Name

The local NetBIOS name associated with the connection.

Remote Host

The name or IP address associated with the remote host.

Type

This refers to the type of name. A name can either be a unique name or a group name.

<03> 

Each NetBIOS name is 16 characters long. The last byte often has special significance, because the same name can be present several times on a computer. This notation is the last byte converted to hexadecimal. For example, <20> is a space in ASCII.

State

The state of NetBIOS connections as shown in Table 15.8

Table 15.8 NetBIOS connection states 

Netstat

This diagnostic command displays protocol statistics and current TCP/IP network connections.

Notes on Netstat 

The netstat utility provides statistics on network components. Table 15.9 outlines these statistics.

Table 15.9 Netstat statistics 

Ping

This diagnostic command verifies connections to one or more remote hosts.

Notes on Ping 

The ping command verifies connections to a remote host or hosts, by sending Internet Control Message Protocol (ICMP) echo packets to the host and listening for echo reply packets. The ping command waits for up to 1 second for each packet sent and prints the number of packets transmitted and received. Each received packet is validated against the transmitted message. By default, 4 echo packets containing 64 bytes of data (a periodic uppercase sequence of alphabetic characters) are transmitted.

You can use the ping utility to test both the host name and the IP address of the host. If the IP address is verified but the host name is not, you may have a name resolution problem. In this case, be sure that the host name you are querying is in either the local Hosts file or in the DNS database.

Route

This diagnostic command manipulates network routing tables.

Telnet

This connectivity command starts terminal emulation with a remote system running a Telnet service. Telnet provides Digital Equipment Corporation (DEC) VT 100, DEC VT 52, or TTY emulation, using connection-based services of TCP.

To provide terminal emulation from a Windows 95 computer, the foreign host must be configured with the TCP/IP program, the Telnet server program or daemon, and a user account for the computer running Windows 95.

The Telnet application is found in the Accessories program group if you install the TCP/IP connectivity utilities. Telnet is a Windows Sockets – based application that simplifies TCP/IP terminal emulation with Windows 98. HyperTerminal can also be used as a Telnet client.

Note Microsoft does not provide the Telnet server daemon (telnetd).

Tracert

This diagnostic utility determines the route taken to a destination by sending Internet Control Message Protocol (ICMP) echo packets with varying Time-To-Live (TTL) values to the destination. Each router along the path is required to decrement the TTL on a packet by at least 1 before forwarding it, so the TTL is effectively a hop count. When the TTL on a packet reaches 0, the router is supposed to send back an ICMP Time Exceeded message to the source system. The tracert command determines the route by sending the first echo packet with a TTL of 1 and incrementing the TTL by 1 on each subsequent transmission until the target responds or the maximum TTL is reached. The route is determined by examining the ICMP Time Exceeded messages sent back by intermediate routers. Notice that some routers silently drop packets with expired TTLs and are invisible to tracert.

Notes on Tracert 

When you type tracert destination, where destination is the host you wish to reach, four columns are displayed. The first column is the hop number, which is the TTL value set in the packet. Each of the next three columns contains the round-trip times in milliseconds for an attempt to reach the destination with that TTL value. An asterisk (*) means that the attempt timed out. The fourth column is the host name (if it was resolved) and IP address of the responding system.

IPX/SPX-compatible Protocol

The Microsoft Internetwork Packet Exchange/Sequenced Packet Exchange (IPX/SPX) – compatible protocol (nwlink.vxd) supports the 32-bit Windows Sockets 1.1 programming interface, so that any Win32-based Windows Sockets 1.1 application can run on IPX/SPX with Windows 98. (There are no 16-bit Windows Sockets applications using IPX/SPX.)

The IPX/SPX-compatible protocol can be used by Client for NetWare Networks to communicate with NetWare servers or computers running File and Printer Sharing for NetWare Networks.

This protocol can also be used by Client for Microsoft Networks to communicate with computers running Windows for Workgroups 3.11 or Windows NT that are also running IPX/SPX.

The IPX/SPX-compatible protocol uses the nwnblink.vxd module to support computers that use NetBIOS over IPX and to support the NetBIOS programming interface. This protocol can also use NetWare servers configured as routers (and other IPX routers) to transfer packets across LANs.

Configuring the IPX/SPX-compatible Protocol

The Microsoft IPX/SPX – compatible protocol is installed automatically when Client for NetWare Networks is installed. You can also install this protocol to support other network clients, including Client for Microsoft Networks.

To install the IPX/SPX-compatible protocol

1. In Control Panel, double-click Network, and then click Add.

2. In the Select Network Component Type dialog box, click Protocol and then click the Add button.

3. In Manufacturers, click Microsoft, and then in Network Protocols, click IPX/SPX-compatible Protocol, and then click OK.

4. Click OK. Setup copies the files you need.

When you install the IPX/SPX-compatible protocol, Windows 98 automatically detects and sets appropriate values for the frame type, network address, and other settings. However, in some cases you might need to configure settings for this protocol manually.

To configure the IPX/SPX-compatible protocol

1. In Control Panel, double-click Network.

2. Double-click the instance of IPX/SPX-compatible Protocol that is bound to your network adapter.

   Note If the computer has multiple network adapters, the list will contain an instance of the IPX/SPX-compatible protocol for each network adapter. You must configure each adapter with its own settings. You can only have four instances of IPX/SPX on your system, so if you have more than four adapters are installed, you should bind IPX/SPX only to the adapters that will use this protocol.

3. Click the Advanced tab.

4. Most values have correct defaults in typical installations. If you need to change a value for a particular purpose, select the item in the Property list and specify a setting in the Value list based on the information in the Table 15.10.

   Table 15.10 IPX/SPX configuration values 

   Property	Value
   Force even-length IPX packets	Enabled only for Ethernet 802.3 on monolithic implementations that cannot handle odd-length packets.
   Frame type1	Specifies the frame type based on detection. This value is used for network adapters that support multiple frame types. The possible values are: Auto-detect (recommended) Ethernet 802.2 (default for NetWare 3.12 and later) Ethernet 802.3 Ethernet II Token-ring Token-ring Subnetwork Access Protocol (SNAP)
   Maximum connections	Specifies the maximum number of connections that IPX will allow. Configured dynamically.
   Maximum sockets	Specifies the maximum number of IPX sockets that IPX assigns. Configured dynamically.
   Network address	Specifies the IPX network address as a four-byte value. Configured dynamically.
   Source routing2	Specifies the cache size to use with source routing. This parameter is used only on token-ring networks, where it is used to turn on source routing. Important Cache size is specified by entry count, not byte count. The recommended value of 16 entries is the most efficient and best setting for most installations.

5. 1 Each time the computer starts, Windows 98 detects the frame type by sending a general RIP request in each frame format. Based on the responses received from routers, Windows 98 determines the most prevalent frame type used and sets that as the default frame type.

   2 Source routing is a method of routing data across bridges. For NetWare networks this means forwarding NetWare frames across an IBM token-ring bridge. With NDIS protocols, source routing is done by the protocol. With ODI-based protocols, source routing is configured with the network adapter driver or using the NetWare route.com utility.

You should not need to change bindings in most circumstances. However, you can disable the bindings for a protocol if you do not want other computers using that protocol to see this computer. At least one protocol, however, must be bound to the network client for the computer to communicate with the network.

To change bindings for the IPX/SPX-compatible protocol

1. In Control Panel, double-click Network.

2. Double-click the instance of IPX/SPX-compatible Protocol that is bound to your network adapter.

   Note If the computer has multiple network adapters, the list will contain an instance of the IPX/SPX-compatible protocol for each network adapter. You must configure each adapter with its own settings. You can only have four instances of IPX/SPX on your system, so if you have more than four adapters are installed, you should bind IPX/SPX only to the adapters that will use this protocol.

3. Click the Bindings tab.

4. Click any network component to change its bindings.

   If the option is checked, it is bound to the protocol. If it is not checked, that network component is not using the IPX/SPX-compatible protocol. For more information, see "Configuring Network Adapters" earlier in this chapter.

   Note Microsoft Client for NetWare Networks is always bound only to the IPX/SPX-compatible protocol. This network client cannot use another protocol.

Using NetBIOS over IPX

NetBIOS is an interface used by network applications to communicate with other NetBIOS-compliant applications. The NetBIOS interface is responsible for the following:

• Establishing logical names on the network.

• Establishing connections (called sessions) between two computers by use of their logical names on the network.

• Transmitting data between networked computers.

Windows 98 provides a 32-bit, protected-mode driver to support NetBIOS services over IPX (Vnetbios.386). This implementation is compatible with the Novell NetBIOS support driver. Performance enhancements include acknowledgment of previous frames in response frames (called PiggyBackAck), plus a "sliding window" acknowledgment mechanism.

These NetBIOS enhancements are used only when the computer is communicating with other computers using IPX over NetBIOS, such as other computers running Windows 98, Windows NT, or NetWare when running Lotus Notes or other NetBIOS applications. NetBIOS over IPX is not necessary for computers running Windows 98 to be able to communicate with each other. The redirector and server networking components in Windows 98 communicate with the IPX protocol directly without NetBIOS.

Novell provides a terminate-and-stay resident (TSR) NetBIOS driver named Netbios.exe, which is a Level 1 NetBIOS provider that consumes about 40K of conventional memory. This driver acknowledges each frame received, thus increasing the amount of traffic generated when NetBIOS is used. With the Microsoft implementation of NetBIOS over IPX, you can remove the real-mode Netbios.exe TSR.

Note A Windows 98 computer that uses IPX without NetBIOS can connect to a Windows NT 3.5 or later server that uses IPX without NetBIOS. However, the Windows NT 3.5 or later computer service can only connect to a Windows 98 computer running File and Printer Sharing for Microsoft Networks when the Windows 98 computer is using NetBIOS over IPX.

To use the IPX/SPX-compatible protocol with NetBIOS on a computer

1. In Control Panel, double-click Network.

2. Double-click the instance of IPX/SPX-compatible Protocol that is bound to your network adapter.

   Note If the computer has multiple network adapters, the list will contain an instance of the IPX/SPX-compatible protocol for each network adapter. You must configure each adapter with its own settings. You can only have four instances of IPX/SPX on your system, so if you have more than four adapters are installed, you should bind IPX/SPX only to the adapters that will use this protocol.

3. Click the Bindings tab.

4. Click the NetBIOS tab, and then click I want to enable NetBIOS Over IPX/SPX.

Technical Notes on IPX/SPX on NetWare Networks

• There is no need to enable source routing on token-ring networks if the communication is on the same ring, even if one computer has it enabled.

• SPX-II is a protocol definition for windowing and transmitting large packets over SPX. The IPX/SPX-compatible protocol included with Windows 98 (Nwlink.vxd) does not support SPX-II. Some third-party dynamic-link libraries (DLLs) provide SPX-II support when you are using Nwlink.vxd. However, if you are using a third-party DLL that provides SPX-II support and a program communicates directly with Nwlink.vxd, SPX-II support will not be available for that program.

• When you install Windows 98 with the IPX/SPX-compatible protocol, parameters in Net.cfg and Shell.cfg will be moved to the registry if they are not already there. (If you are upgrading over an installation of Windows 95 that included the IPX/SPX-compatible protocol, those parameters will already have been moved to the registry.)

• To determine the network address in IPX packets, Windows 98 checks the wire for RIP packets and chooses the most likely address. The network address is dynamic and changes when a new network address becomes more prevalent.

• All transport layer interface (TLI) libraries can run on the IPX/SPX-compatible protocol in Windows 98. TLI is similar to TDI in Microsoft networking as a layer between the protocol and network adapter driver; this implementation is similar to STREAMS and provides a STREAMS environment for NetWare, but Windows 98 uses Windows Sockets instead.

MS NetBEUI Protocol

Windows 98 provides the NetBIOS extended user interface (NetBEUI) protocol for compatibility with existing networks that use NetBEUI. Because NetBEUI is nonroutable and was designed for smaller LANs, you should use the TCP/IP or IPX/SPX-compatible protocol for enterprise-wide networks that require a routable protocol.

NetBEUI in Windows 98 provides two types of traffic:

1. Unreliable connectionless traffic, in which the sender sends packets to the receiver without setting up a connection and with no guarantee that the packets will arrive. NetBEUI connectionless traffic are used for name resolution, datagrams, and miscellaneous traffic.

2. Reliable connection-oriented traffic, in which the sender and receiver establish a reliable connection before sending data. NetBEUI connection oriented traffic is used for commands such as net use, net view, and net start.

NetBEUI in Windows 98 supports a NetBIOS programming interface that conforms to the IBM NetBEUI specifications and includes several performance enhancements. The NetBEUI module, Netbeui.vxd, is accessible through the NetBIOS interface.

If Windows 98 Setup detects NetBEUI during installation, it installs support for Microsoft NetBEUI automatically. If you are upgrading from a computer that did not have NetBEUI, Windows 98 does not automatically install it; however, you can add it at any time.

To install NetBEUI

1. In Control Panel, double-click Network.

2. On the Configuration tab, click Add.

3. In the Select Network Component Type dialog box**,** click Protocol, and then click the Add button.

4. In Manufacturers, click Microsoft, and then in Network Protocols, click NetBEUI, and then click OK.

5. Click OK. Setup copies files from your original installation media and prompts you to restart your computer.

The Advanced properties for NetBEUI affect only real-mode NetBEUI. These values are set dynamically for protected-mode NetBEUI.

To configure real-mode NetBEUI manually

1. In Control Panel, double-click Network.

2. On the Configuration tab, double-click the instance of NetBEUI that is bound to your network adapter.

   Note If your computer has multiple network adapters, an instance of NetBEUI appears for each network adapter. You must configure each adapter with its own settings.

   The Bindings tab shows which clients and services are currently using the NetBEUI protocol. For information about configuring bindings, see "Configuring Network Adapters" later in this chapter.

3. Click the Advanced tab to modify settings for Maximum Sessions and NCBs for the real-mode NetBEUI. The following list outlines the options in NetBEUI advanced configuration.

   Option	Description
   Maximum Sessions	Used to identify the maximum number of connections to remote computers that can be supported from the redirector. This is equivalent to the sessions= parameter formerly specified in Protocol.ini.
   NCBs (network control blocks)	Used to identify the maximum number of NetBIOS commands that can be used. This is equivalent to the ncbs= parameter formerly specified in Protocol.ini.

4. Click OK. Then shut down and restart the computer.

MS DLC Protocol

Windows 98 includes both a protected-mode DLC driver (Dlc.vxd), which supports 32-bit and 16-bit DLC applications, and a real-mode DLC driver (Msdlc.exe), which supports only 16-bit applications. Microsoft DLC is used primarily to access IBM mainframe computers and Hewlett-Packard network-ready printers. Microsoft recommends that you use the protected-mode driver in most cases. However, some 16-bit DLC applications will work only with the real-mode DLC driver. For more information about applications that require the 16-bit DLC, see "Using 32-bit DLC for Connectivity" later in this chapter.

The following sections how to use 32-bit, protected-mode DLC protocol and the 16-bit, real-mode DLC protocol.

Using 32-bit DLC for Connectivity

The 32-bit DLC protocol provides connection and communication with mainframe computers using DLC. With the 32-bit Data Link Control (DLC) protocol, you can establish multiple connections to different IBM host and AS/400 computers over the same network token-ring, FDDI, or Ethernet adapter. Host terminal emulation programs use the 32-bit DLC to communicate directly with host computers. The 32-bit DLC protocol also allows multiple 32-bit applications to use the same network adapter at the same time to connect to different host computers.

You can also use the 32-bit DLC protocol to provide connectivity to local area printers connected directly to the network. For example, you can use the DLC protocol to print to a printer that uses an adapter to connect directly to the network rather than to a port on a print server. The DLC protocol must be installed and running on the print server for the printer. The protocol does not need to be installed on the computers that send print jobs to the print server. To take advantage of the DLC protocol device driver, you must create a network printer in the printers folder by using a third-party utility such as Hewlett Packard's Jet Administration Utility.

The 32-bit DLC protocol software supports the following 32-bit and 16-bit DLC programs:

• Windows-based 32-bit programs that use Command Control Block 2 (CCB2).

• 16-bit programs that use Command Control Block 1 (CCB1).

Microsoft recommends that in most cases you use the 32-bit DLC protocol rather than the 16-bit DLC protocol or IBM's LAN Support drivers, because it supports both 32-bit applications and most 16-bit applications, provides better performance, and can access protected memory.

However, the 32-bit DLC protocol does not support certain 16-bit DLC applications. Some 16-bit DLC applications load TSRs to check the existence of DLC before Windows 98 loads. If you need to run any of those applications, you may want to use the 16-bit DLC protocol. For more information about the 16-bit programs the 32-bit DLC protocol does not support, see "Running 16-bit Applications" later in this chapter. The 32-bit DLC protocol also does not support 32-bit programs created to run under OS/2.

Client for Microsoft Networks does not use the 32-bit DLC protocol to communicate with the Microsoft network. Also, the 32-bit DLC protocol does not have a NetBIOS interface. 32-bit terminal emulation programs usually call the protected-mode DLC protocol by loading a TSR that acts as an interface to the DLC protocol. A 16-bit terminal emulation program can call the 32-bit DLC protocol by using the Int 0x5C (NetBIOS) interrupt vector. The 32-bit DLC protocol can coexist with other protocols, and it conforms to the network driver interface specification (NDIS) 3.1 and later.

Windows 98 supports the Ethernet multivendor standard DIX 2.

The following sections describe how to install the Microsoft 32-bit DLC protocol and how to configure it with Novell NetWare ODI drivers and IBM LAN Support.

For information about the architecture of Microsoft DLC, see Chapter 29, "Windows 98 Network Architecture."

Installing the 32-bit DLC Protocol

Before you install the 32-bit DLC protocol, you must perform the following tasks:

• Make sure you have access to your original Windows 98 setup files.

• If you are installing the 32-bit DLC protocol over the IBM LAN Support program (DXM mode drivers), first remove the existing IBM DLC from the Network option in Control Panel.

• If you are installing the 32-bit DLC protocol over an existing Madge Smart.exe (token-ring), you must first remove the existing ODI driver from the Network option in Control Panel. Then you must add the NDIS 3.1 or later Madge driver and the Microsoft 32-bit DLC protocol stack.

To install the 32-bit DLC protocol

1. In Control Panel, double-click Network.

2. On the Configuration tab, click Add.

3. Click Protocol, and then click Add.

4. In the Manufacturers box, click Microsoft. 

5. In the Network Protocols box, click Microsoft 32-bit DLC.

6. Click OK.

7. Setup prompts you to shut down and reboot your computer.

Windows 98 Setup removes any previous configuration settings for 16-bit Microsoft DLC, and adds an entry in the Autoexec.bat file for Dlchlp.exe, the file that supports 16-bit DLC programs. If no 16-bit DLC programs are used, you can save real-mode and 32-bit DLC protected-mode memory by disabling 16-bit support. For more information about disabling 16-bit support, see "Disabling 16-bit Support" later in this chapter.

Windows 98 setup also associates a unique CCB adapter number to the adapter that is bound to the DLC protocol.

For information about changing the CCB adapter number, see "Configuring the 32-bit DLC Protocol" later in this chapter.

For information about using multiple CCB adapter numbers, see "Adding a Second Network Adapter" later in this chapter. After you have installed the 32-bit DLC protocol, you can verify that the protocol is bound to the adapter by using the following procedure.

To verify bindings

1. In Control Panel, double-click Network, and then click the name of your adapter.

2. Click Properties. 

3. Click the Bindings tab.

4. Make sure Microsoft 32-bit DLC is selected.

Note You cannot check the bindings by clicking the 32-bit DLC protocol, because only clients and services can be bound to protocols.

Configuring the 32-bit DLC Protocol

After you have installed the 32-bit DLC protocol, you can configure adapter settings and settings for the 32-bit DLC protocol.

For information about configuring adapter settings, see Chapter 15, "Network Adapters and Protocols."

This section covers the following topics:

• Configuring settings for the 32-bit DLC protocol.

• Configuring Service Access Points (SAPs) and link stations.

• Configuring timer values.

• Adding a second network adapter.

• Disabling 16-bit support.

Configuring Settings

This section describes the 32-bit DLC settings you can manually configure.

Note In most cases, you do not need to change the default values. However, if you are loading multiple terminal emulation sessions, you might want to increase values for SAPs and link stations.

To configure 32-bit DLC settings

1. In Control Panel, double-click Network.

2. Click the Microsoft 32-bit DLC protocol.

3. Click Properties.

4. Click the Advanced tab.

Table 15.11 shows the default settings you can change. Applications can also change these settings.

Table 15.11 DLC configuration settings 

Choosing 16-bit DLC Protocol Settings

Table 15.12 shows 16-bit DLC protocol settings that are still available with the 32-bit DLC protocol. Note that the names have changed.

Table 15.12 Available 16-bit DLC protocol settings 

Determining Links and SAPs Settings

Each program requires a certain number of Service Access Points (SAPs) and link stations. Because each SAP or link station takes up memory, you should provide just enough for your program to run.

If you do not know the number of SAPs and link stations your program requires, and you want to minimize the memory your terminal emulation programs use, start with large values and gradually reduce them until the program no longer works.

The Microsoft 32-bit DLC protocol uses defaults of 5 SAPs and 20 link stations. This should be sufficient for most programs. If needed, you can increase the number of SAPs and link stations. For more information about configuring 32-bit DLC settings, see "Configuring Settings" earlier in this chapter.

For information about which SAP and link station settings to use with your DLC programs, see the documentation for those programs.

Configuring Timer Settings

The Microsoft DLC protocol uses three timers:

• t1 (retransmission)

• t2 (acknowledgment)

• ti (inactivity)

Each timer has a "short tick" rate and a "long tick" rate that individual commands use to determine timer values. A command such as Dlc.Open.Sap specifies a timer value with a number range of 1 through 10 units of milliseconds.

For example:

• When the number is in the range of 1 through 5 units of milliseconds, the actual timer value is:

  (number selected) * (short-tick value) * 40 milliseconds
  

• When the number is in the range of 6 through 10 units of milliseconds, the actual timer value is:

  (number selected - 5) * (long-tick value) * 40 milliseconds
  

Some network programs adjust these timer entries automatically. The Dlc.Open.Adapter command overrides the default value.

Adding a Second Network Adapter

If you add a second DLC32-compatible network adapter to your computer, a new copy of the 32-bit DLC protocol is created and automatically bound to your secondary network adapter. When you restart your computer, the CCB adapter number setting for that copy of the 32-bit DLC protocol automatically changes to a unique number.

Note If you remove your primary network adapter, the CCB adapter number for the second adapter does not change.

Disabling 16-bit Support

If your users will use only 32-bit network programs, you can disable support for 16-bit programs. This speeds up network access and saves real-mode and protected-mode memory.

To disable 16-bit support

1. In Control Panel, double-click Network, and then click the Microsoft 32-bit DLC protocol for which you wish to disable 16-bit support.

2. Click Properties.

3. Click the Advanced tab.

4. Click the Support CCB1 setting, and set it to 0. The reference in your Autoexec.bat file to Dlchlp.exe is removed.

Caution Do not manually remove Dlchlp.exe from your Autoexec.bat file.

Switching from IBM LAN Support

The 32-bit DLC protocol cannot be upgraded over IBM LAN Support. If you are adding the 32-bit DLC protocol to a computer that used IBM LAN Support, you must first remove the IBM DLC protocol from the Network option in Control Panel, then add the Microsoft 32-bit DLC protocol. Otherwise, you may not be able to connect to your host, and the Windows 98 computer could stop responding.

Note Before upgrading over IBM LAN Support, record the parameters for the DXM driver line(s) so that you can configure the equivalent Microsoft 32-bit DLC protocol settings. If you are using an NDIS configuration with IBM LAN Support, you should also record the parameters for your Protocol.ini file.

Windows 98 Setup does not migrate the IBM LAN Support program settings will to the Microsoft 32-bit DLC protocol. You will need to make sure that the settings for the Microsoft 32-bit DLC properties correlate to the prior settings for IBM LAN Support. Table 15.13 shows how to convert the Xmit_swap setting in Dxme0mod.sys to the Swap Addr Bits and Ethernet DIX settings for the 32-bit DLC protocol.

Table 15.13 IBM LAN Support program settings 

Note You can configure the alternate adapter by using the CCB Adapter Num setting. For more information about the CCB Adapter Num setting, see "Adding a Second Network Adapter" earlier in this chapter.

Running 16-bit Applications

If you are running 32-bit Microsoft DLC with 16-bit applications that use DLC, you will need to update your Autoexec.bat file. Make sure that Net.exe is loaded before Dlchlp.exe.

Applications that require a terminate-and-stay-resident (TSR) program to connect to a host computer before starting Windows 98 will not work with the Microsoft 32-bit DLC protocol stack. The following applications require TSRs:

• IBM PC SUPPORT (versions V2R1 - V2R3)

• RELAY GOLD version 5.0 and 6.0a

• EICON ACCESS 5250 version 3.21

Contact your vendor for updated information about these applications.

Technical Notes for the 32-bit DLC Protocol

Table 15.14 describes the files included with the 32-bit DLC protocol.

Table 15.14 32-bit DLC files 

Note If you do not need support for 16-bit DLC programs, do not remove Dlchlp.exe from your Autoexec.bat file manually. Instead, disable support for 16-bit DLC programs by using the Support CCB1 setting.

Using 16-bit DLC for Connectivity

The 16-bit DLC protocol provides connection and communication with mainframe computers using DLC. To provide connection and communication, the DLC protocol must be installed on the client computer that is running Windows 98. The Microsoft DLC protocol works with either token-ring or Ethernet network adapter drivers.

Microsoft DLC is also used to provide connectivity to local area printers connected directly to the network. For example, DLC can be used for printing to a printer such as an HP LaserJet 4Si that uses an HP JetDirect network adapter to connect directly to the network (rather than to a port on a print server). The DLC protocol must be installed and running on the print server for the printer. Computers sending print jobs to a print server for a DLC network printer do not need the DLC protocol — only the print server requires DLC. To take advantage of the DLC protocol device driver, you must create a network printer in the Printers folder.

In addition to the 16-bit DLC protocol, Windows 98 includes a 32-bit DLC protocol. In most cases, Microsoft recommends that you use the 32-bit DLC protocol. It provides better performance and can access protected memory. However, certain applications require Microsoft's 16-bit DLC because they require a TSR program to connect to a host computer before starting Windows 98. For more information about these applications, see "Running 16-bit applications" earlier in this chapter.

The following sections describe how to install the Microsoft DLC protocol and how to configure it with Novell NetWare ODI drivers and IBM LAN Support.

Installing and Configuring Real-Mode MS DLC

This section describes how to install the real-mode NDIS 2 Microsoft DLC protocol to bind with an NDIS 2 network adapter driver. This is an NDIS 2 protocol that, when installed, is bound to an NDIS 2 network adapter driver. The Windows 98 protected-mode protocols use the Ndis2sup.vxd module to coexist with and run over the real-mode NDIS 2 network adapter drivers.

When you install Windows 98, Windows 98 Setup detects whether the computer already has the Microsoft DLC protocol installed; if so, Setup migrates the protocol and its settings to Windows 98. Otherwise, you can add Microsoft DLC as a protocol after Windows 98 is installed.

Note You must install the version of Microsoft DLC provided with Windows 98. You cannot use the Microsoft DLC INF file from Windows for Workgroups 3.x to install this protocol; Windows 98 uses a new INF file format.

You can also install Microsoft DLC using setup scripts, as described in Appendix D, "Msbatch.inf Parameters for Setup Scripts."

To install the Microsoft DLC protocol on a computer running Windows 98

1. In Control Panel, double-click Network.

2. On the Configuration tab, click Add. 

3. In the Select Network Component Type dialog box, double-click Protocol.

4. In the Select Network Protocol dialog box, click Microsoft in the Manufacturers list, and then click Microsoft DLC in the Network Protocols list. Then click OK.

5. Shut down and restart the computer for the changes to take effect.

6. Make sure that there is only one net start entry in Autoexec.bat, because Setup automatically adds an entry to support DLC, even if such an entry already exists.

The properties in Table 15.15 are set by default for real-mode Microsoft DLC. You can use the Network option in Control Panel to change these default values, or to set values for other parameters that you might use, in the Advanced properties for Microsoft DLC protocol.

Table 15.15 DLC properties 

The default values for SWAP and UseDIX are appropriate for most token-ring LAN environments. If the computer has an Ethernet adapter, then you should set the correct values for these parameters in the Advanced properties for Microsoft DLC. If you previously used the IBM Dxme0mod.sys driver, use Table 15.16 to map the Xmit_swap parameter to set values for the two Microsoft DLC parameters.

Table 15.16 Mapping Xmit_swap parameters 

Table 15.17 shows some typical settings in Autoexec.bat and Protocol.ini for Microsoft DLC with an Intel EtherExpress PRO LAN NDIS 2 network adapter driver on an Ethernet network. Notice that the [msdlc$] section is added automatically by Windows 98 when the protocol is installed.

Table 15.17 Sample configuration file settings for Microsoft DLC with NDIS 2 adapters 

net init
msdlc.exe
net start

[netbeui$]
DriverName=NETBEUI$
Lanabase=0
sessions=10
ncbs=12
Bindings=EPRO$
[nwlink$]
DriverName=nwlink$
Frame_Type=4
cachesize=0
Bindings=EPRO$
[epro$]
DriverName=EPRO$
INTERRUPT=10
ioaddress=0x300
[protman$]
priority=ndishlp$
DriverName=protman$
[ndishlp$]
DriverName=ndishlp$
Bindings=EPRO$

[data]
version=v4.00.000
netcards=EPRO$,*PNP8132
[msdlc$]
DriverName=msdlc$
stations=20
saps=3
swap=0
usedix=1
Bindings=EPRO$

Configuring Real-mode MS DLC with ODI Drivers

For computers that are running Microsoft DLC with ODI drivers using the Novell-supplied Odinsup.exe file, Windows 98 Setup installs over this configuration and leaves entries for ODINSUP and MSDLC in Autoexec.bat. Microsoft does not provide direct support for Microsoft DLC used with Odinsup.exe.

If you must run a real-mode network redirector or TSR (Netx.exe, Vlm.exe, and so on), you need to configure Microsoft DLC by binding the Microsoft DLC protocol to an ODI network adapter driver. Otherwise, install Microsoft DLC over NDIS 2 with the protected-mode Microsoft Client for NetWare Networks. For information about these configurations, see "Architecture for DLC" in Chapter 29, "Windows 98 Network Architecture."

Table 15.18 shows settings used to configure Microsoft DLC with ODI drivers.

Note This manual editing of Protocol.ini will remove "Microsoft DLC" from the Network Configuration dialog. The usage of Odinsup.exe prevents Windows 98 from correctly identifying the "Microsoft DLC" as being used by the Ethernet adapter and therefore will not be displayed even though the protocol will still function.

Table 15.18 Configuration file settings for real-mode Microsoft DLC with ODI drivers 

lsl ;Novell-supplied component
mlid_driver.com ;Novell-supplied component
odihlp.exe ;Windows 98 component
odinsup.exe ;Novell-supplied component
msdlc.exe ;Windows 98 component
net start netbind ;Windows 98 component

Protocol ODINSUP
Bind EPROODI
BUFFERED
Link Driver EPROODI
Port 300
Frame Ethernet_802.2
Frame Ethernet_802.3
Frame Ethernet_II
Frame Ethernet_Snap

[protman$]
priority=ndishlp$
DriverName=protman$
[ndishlp$]
DriverName=ndishlp$
Bindings=
[data]
version=v4.00.000
netcards=
[nwlink$]
Frame_Type=4
cachesize=0
DriverName=nwlink$
[msdlc$]
DriverName=msdlc$
xstations0=0
xstations1=0
stations=20
saps=3
xsaps0=1
xsaps1=1
swap=0
usedix=1
Bindings=EPROODI
[EPROODI]
Drivername=EPROODI
INTERRUPT=10
ioaddress=0x300

Upgrading Existing IBM LAN Support Installations

If you are using DLC support supplied by IBM to connect to host computers, Windows 98 Setup can detect IBM DLC; it leaves the installation intact and configures Windows 98 to run over that configuration. Although support for IBM DLC can be installed using the Network option in Control Panel, the required components must be provided by your network vendor.

This section describes two typical IBM LAN Support configurations, using Dxmc0mod.sys and Dxme0mod.sys.

Dxmc0mod.sys, the monolithic IBM DLC driver. For this configuration, Windows 98 Setup does one of two things:

• If the computer is running NetWare, Setup keeps the Dxmc0mod.sys driver and related settings, and installs the Generic ODI driver plus Microsoft Client for NetWare Networks, or keeps the real-mode client.

• If the computer is running Dxmc0mod.sys and no other networking components, Setups keeps the Dxmc0mod.sys driver and does not install any Windows 98 networking components.

Table 15.19 shows a sample configuration for IBM Dxmc0mod.sys with Microsoft Client for NetWare Networks.

Table 15.19 Example of configuration settings for Dxmc0mod.sys with ODI drivers 

lsl ; Novell-supplied component
lansup ; Novell-supplied component
odihlp.exe ; Microsoft component

device=path\dxmaood.sys ;IBM-supplied component
device=path\dxmcomod.sys ;IBM-supplied component

Dxme0mod.sys, the NDIS driver for IBM DLC. For this configuration, Windows 98 Setup does one of three things:

• If the computer is running Dxme0mod.sys and no other networking components, Setup keeps the Dxme0mod.sys driver and does not install any Windows 98 networking components.

• If the computer is running Dxme0mod.sys and Novell NetWare, Setup installs an NDIS network adapter driver plus Microsoft Client for NetWare Networks, and leaves the Dxme0mod.sys driver intact.

• If the computer is running Dxme0mod.sys with the IBM MS-DOS LAN Requestor, Setup installs an NDIS 2 adapter driver, keeps the Dxme0mod.sys driver, installs Client for Microsoft Networks, and removes the IBM MS-DOS LAN Requestor redirector components.

Table 15.20 shows a sample configuration for IBM Dxme0mod.sys with Client for Microsoft Networks using an NDIS 2 adapter driver. The same basic kinds of settings are used for a computer running Microsoft Client for NetWare Networks with an NDIS 2 adapter driver.

Table 15.20 Example of settings for Dxmce0mod.sys with Client for Microsoft Networks 

device=c:\windows\protman.dos /i:c:\windows
device=c:\windows\epro.dos ;ndis2 driver
device=c:\lsp\dxma0mod.sys ;IBM-supplied
device=c:\lsp\dxme0mod.sys ,,3 ;IBM-supplied
device=c:\windows\ndishlp.sys ;Windows 98 

[protman$]
priority=ndishlp$
DriverName=protman$
[ndishlp$]
DriverName=ndishlp$
Bindings=EPRO$
[data]
version=v4.00.000
netcards=EPRO$,*pnp8132
[netbeui$]
DriverName=NETBEUI$
Lanabase=0
sessions=10
ncbs=12
Bindings=EPRO$
[nwlink$]
DriverName=nwlink$
Frame_Type=4
cachesize=0
Bindings=EPRO$
[EPRO$]
DriverName=EPRO$
INTERRUPT=10
ioaddress=0x300

[DXMAIDXCFG]
dxme0_nif=dxme0.nif
dxmj0mod_nif=dxmj0mod.nif
smcdosjp_nif=smcdosjp.nif
smcdosjp2_nif=smcdosjp.nif
smcdosat_nif=smcdosat.nif
smcdosat2_nif=smcdosat.nif
smcdosmc_nif=smcdosmc.nif
smcdosmc2_nif=smcdosmc.nif
[ETHERAND]
DriverName=DXME0$
Bindings=EPRO$

Overview of Network Adapters

Windows 98 Setup automatically detects most network adapters, installs the appropriate driver for the adapter, and provides appropriate default settings to configure the adapter. If you add a new network adapter, its driver is bound automatically to all network driver interface specification (NDIS) – compatible protocols currently running on the computer. If any protocols are added later, they will also be bound automatically to the network adapter driver. TCP/IP can be bound to a maximum of six network adapters.

This section provides technical details for configuring network adapters, setting LAN adapter numbers, and other technical notes.

For specific information about PC Card adapters, see Chapter 24, "Device Management."

Note For information about specific network adapters, see the Windows 98 Adapter Card Configuration Help File on the Microsoft Windows 98 Resource Kit compact disc.

Understanding the Benefits of NDIS Adapter Drivers

The network driver interface specification (NDIS) describes the interface that network adapter drivers use to communicate with underlying hardware, with overlying protocol drivers, and with the operating system. All network adapter drivers and protocols provided with Windows 98 conform to NDIS. Windows 98 provides NDIS 5.0, an extension of NDIS 4.0 that supports NDIS versions 2.x 3.1, 4.0, and 5.0 protocol and adapter drivers. Windows 98 also provides a replacement for version 3.0 drivers, which are incompatible with Windows 98.

You should use NDIS 3.1 or later drivers whenever possible with Windows 98. If you are using a Novell-supplied network client, you should use ODI-based client software rather than monolithic IPX drivers.

NDIS 5 drivers add to the functionality provided by NDIS 3.1 drivers. For example, by using NDIS 5 drivers, Windows 98 can support a wide range of network media, including Ethernet, fiber distributed data interface (FDDI), token-ring, asynchronous transfer mode (ATM), and WAN technology. The NDIS 5 specification accommodates Plug and Play features, so that in many cases network adapters can be added and removed dynamically while the computer is running. NDIS 5 drivers also provide performance improvements, and compatibility with Windows NT Server 5.0.

This section summarizes the related benefits.

For information about NDIS architecture, see Chapter 29, "Windows 98 Network Architecture." See also the introduction to NDIS at https://www.microsoft.com/whdc/hwdev/tech/network/ndis5.mspx .

Plug and Play support for network protocols and adapters. Windows 98 can automatically determine the adapters to which each protocol should bind. Additionally, if you are using a PC Card or CardBus adapter, and a Plug and Play event occurs, such as the removal of an adapter from a portable computer, the NDIS protocols and network adapters can detect the event and remove themselves from memory automatically.

NDIS miniport driver model. In releases of NDIS before 3.1, adapter drivers implemented not only the media access functionality that was specific to the network adapter, but also the media access functionality that is common to all NDIS drivers.

For NDIS 3.1 and later adapter miniport drivers, the Windows 98 NDIS wrapper implements the half of the media access functionality that is common to all NDIS drivers. Thus, the miniport driver provided by the adapter manufacturer must implement only the half of the media access layer that is specific to the network adapter. These include specific details such as establishing communications with the adapter, turning on and off electrical isolation for Plug and Play, providing media detection, and enabling any value-added features the adapter might contain.

The Windows 98 miniport drivers are binary-compatible with Windows NT 3.5 and later miniport drivers, which means they can be used on either operating system without being recompiled. (You can recognize a miniport driver by its .sys file name extension; other drivers have .vxd extensions.)

Real-mode NDIS 2 support. An NDIS 2.x protocol under Windows 98 must use an NDIS 2.x network adapter driver. Both the protocol and network adapter drivers must load and bind in real mode before Windows 98 runs. Values in Protocol.ini are used to load the real-mode NDIS drivers, as described in Chapter 16, "Windows 98 on Microsoft Networks." However, you still use the Network option in Control Panel to configure NDIS 2 adapters.

When you run a real-mode network, Windows 98 uses NDIS 2 versions of NetBEUI and IPX/SPX protocols. These protocols are not intended for everyday use, since Windows 98 supplies faster protected-mode versions of these protocols.

Windows 98 also supports existing ODI drivers with Novell NetWare – compatible network clients. For information, see Chapter 17, "Windows 98 on Third-Party Networks."

Support for Windows Management Infrastructure. Windows Management Interface (WMI) collects a wealth of information about the entire system as well as device configuration. This information is stored in the registry and made available through extensions to the Registry API.

NDIS 5 supports the Windows Management Infrastructure (WMI) for Web-Based Enterprise Management (WBEM) of NDIS miniports and their associated adapters, providing the architecture for user-mode management of NDIS drivers and network adapters. This will enable network administrators and support technicians to remotely monitor and control systems using NDIS 5 compliant drivers.

For more information about WMI, see https://www.microsoft.com/whdc/hwdev/tech/network/ndis5.mspx .

Support for broadcast media. NDIS 5 includes extensions to support Broadcast Architecture. It supports high-speed unidirectional broadcast media such as services provided by DirectTV, PrimeStar, and Intercast. The extensions include definitions for receiver tuning, multiple media stream negotiation, fast data streaming, and support for UDP/IP multicast packets using a Microsoft LAN Emulation driver.

For more information about Broadcast Architecture, see Chapter 13, "WebTV for Windows 98."

Support for connection-oriented media. In addition to supporting connectionless media such as Ethernet, token-ring, FDDI, Wireless WAN, and infrared, with NDIS 5 Windows 98 now supports raw access to connection-oriented media such as ATM, integrated services digital network (ISDN), X.25, and Frame Relay. This support is provided only if you are using miniport drivers.

For more information about NDIS support for connection-oriented networks, see Chapter 29, "Windows 98 Network Architecture."

Plug and Play Networking

Plug and Play is an independent set of computer architecture specifications that hardware manufacturers use to produce computer devices that can be configured with no user intervention. With Windows 98, you can install Plug and Play – compliant devices such as network adapters simply by plugging in the device and turning on the computer. If Windows 98 has support for the device it automatically adds its configuration information to the registry, ensures that the correct files are installed, and ensures that the configuration options are set properly.

The networking components in Windows 98 are designed for dynamic Plug and Play operation with most ISA, EISA, PCI, IBM Micro Channel, CardBus, and PC Card network adapters. To take advantage of these features, the computer must be running all protected-mode networking components, including client, protocols, and network adapter drivers.

NDIS 3.1 and later supports adding and removing Plug and Play network adapters dynamically while the computer is running. For example, if you undock a portable computer (called hot undocking), the Windows 98 protocols can remove themselves from memory automatically.

Additional Plug and Play networking benefits are available when you use 32-bit socket services with PC Card cards. You can click the PC Card icon on the taskbar to remove the card without shutting down Windows 98 or turning off the computer. Clicking the PC Card icon causes the operating system to perform an orderly shutdown of the affected network components. Windows 98 notifies applications that the network is no longer available and automatically unloads any related drivers or protocols.

To help mobile users who might need to change adapters in their hardware, Windows 98 uses 32-bit Card and Socket Services to support hot removal and insertion of PC Cards, including network adapters. Support for hot docking means that users do not have to restart their computers each time they make a change to their hardware configuration by docking or undocking their computers. For information about using and configuring PC Cards, including how to enable 32-bit Card and Socket Services, see Chapter 30, "Hardware Management."

Network Plug and Play support in Windows 98 includes application-level support. An application created for Windows 98 might be designed with the ability to determine whether the network is available. Therefore, if a network adapter is removed, for example, the application automatically puts itself into "offline" mode to allow the user to continue to work, or it shuts down.

Configuring Network Adapters

This section describes how to configure network adapters for Windows 98 computers. First, the section "Installing and Configuring Network Adapters" provides a brief overview of how to install both Plug and Play and legacy network adapters. Next, the section "Configuring Network Adapter Settings in Windows 98" explains how to configure network adapter driver properties from within Windows 98. Finally, the section "Binding Network Adapter Drivers to Protocols" describes how to bind network adapter drivers to protocols.

Installing and Configuring Network Adapters

How you install your network adapter depends on what type of network adapter you have.

If you have a Plug and Play network adapter, Windows 98 automatically detects the adapter and performs most of the required configuration.

If you have a non-Plug and Play network adapter (also called a legacy network adapter), the Add New Hardware wizard starts and should automatically configure the network adapter. However, you might need to perform software or hardware configuration. For example, with some legacy network adapters you must set jumpers or switches on the hardware itself. With others, you must run a vendor-supplied software configuration program from within an MS-DOS window.

Note that this configuration is separate from the configuration you must do from within Windows 98. As the section "Configuring Network Adapter Settings in Windows 98" explains, you must configure settings from the Network option in Control Panel to match the settings you have configured using jumpers, switches, or a software configuration program.

Configuring Network Adapter Settings

This section discusses how to configure properties for network adapter drivers.

The properties you must configure from within Windows 98 depend on whether you are installing Plug and Play or non-Plug and Play devices.

If you install a Plug and Play network adapter, Windows 98 should automatically configure it. You should not need to change most of the settings, including any of the settings described in "Configuring Network Adapter Resource Settings" later in this chapter. However, if your network configuration is not working properly, you might want to review this section.

If you install a non-Plug and Play network adapter, the Add New Hardware Wizard should automatically configure it. When asked whether you want Windows to search for your new hardware (that is, to perform hardware detection), make sure to click Yes. This helps Windows 98 choose the correct driver and resource assignments. If hardware detection fails, you might need to configure Windows 98 settings to match the settings for the network adapter itself. For example, if you used a software configuration utility to set an Interrupt Request (IRQ), you must configure Windows 98 to use the same IRQ for the network adapter.

If you experience problems with the settings for a network adapter, you should begin troubleshooting by removing the network adapter driver from the Network option in Control Panel. Then reboot, and if the card is not detected, use the Add New Hardware option to reinstall support for the network adapter, using hardware detection to ensure that Windows 98 determines the correct adapter driver and standard settings for that network adapter. If you experience problems after installing a Plug and Play device or after automatically detecting a non-Plug and Play device from the Add New Hardware option, you might need to manually configure properties for network adapter drivers.

Specifying the Driver Type for a Network Adapter

Generally, you will not need to specify the driver type for a network adapter. In fact, some network adapters do not let you specify the driver type. However, if you need to use a real-mode driver you must specify the driver type.

To specify the driver type for a selected network adapter

1. In Control Panel, double-click Network.

2. On the Configuration tab, select the name of the network adapter driver from the list of installed components.

3. Click the Properties button.

4. In the properties for the network adapter, click the Driver Type tab.

5. Click one of three options (if available for the specific adapter), as described in the following list.

   Network adapter driver type	Description
   Enhanced mode (32-bit and 16-bit) NDIS driver	Installs an NDIS 3.1 or later – compliant driver. This is the preferred driver type for use with 32-bit, protected-mode network clients.
   Real-mode (16-bit) NDIS driver	Installs an NDIS 2.x – compliant driver.
   Real-mode (16-bit) ODI driver	Installs a real-mode driver created to support ODI for Windows 3.1 on NetWare networks.

Configuring Network Adapter Resource Settings

Windows 98 can determine hardware settings for Plug and Play network adapters. Thus, for Plug and Play network adapters you should let Windows 98 go through its Plug and Play installation process. You should not change any settings described in this section unless you are absolutely sure they are incorrect. In fact, you cannot change network adapter resource settings for most Plug and Play network adapters.

For non-Plug and Play network adapters, you should use the Add New Hardware option in Control Panel, using detection to determine the correct driver and resource settings. You should accept the proposed settings unless you are absolutely sure they are incorrect.

Most ISA devices cannot share IRQ settings, memory buffer addresses, or ROM addresses. Where possible, Windows 98 identifies and resolves conflicts. However, if one of the supported devices does not seem to work, the problem may be the particular hardware configuration. To make sure there are no conflicts among network adapters or other peripherals, or between the system board and adapters, check the settings in Device Manager.

For information about checking the settings in Device Manager, see Chapter 24, "Device Management."

If Device Manager shows that any of these settings conflict, you can use the Network option in Control Panel to make sure that the settings match the settings for your network adapter.

To configure resources in a network adapter's properties

1. In Control Panel, double-click Network.

2. In the list of installed components, click the network adapter and then click the Properties button. Make sure to click the network adapter, not the protocol binding for that adapter.

3. In the properties for the network adapter, if a Resources tab appears, click it.

   Note If a Resources tab does not appear, you cannot configure the network adapter's resources from the Network option in Control Panel.

4. Select a configuration from the Configuration type box. Confirm values for the listed settings by comparing the proposed settings with the values recommended in the documentation for the adapter.

5. To select from the available values for a setting, click the arrow beside the setting's current value.

   • A hash (#) character appears by current settings.

   • An asterisk (*) appears beside settings that conflict with another device in the system. You should avoid this setting or reconfigure the other devices to use different settings.

The settings available depend on the type of network adapter. For example, for Intel adapters, you cannot set the IRQ using the adapter's properties. Table 15.21 describes a few typical settings. Each setting must match the adapter's settings, as specified in the documentation for the adapter.

Table 15.21 Example hardware resource settings for a network adapter 

If the settings in the Resources tab do not match the adapter's settings, you must either change the settings in the Resources tab or, in some cases, change the adapter's settings. To determine how to change the adapter's settings, refer to the documentation for your network adapter.

Note For some legacy adapters, it is possible that the adapter uses resources not listed with the Resource properties. For these adapters, the NDIS driver determines the resource settings directly from the adapter itself. Even though these resources do not appear in the list, they can still conflict with other devices. For example, the resource list for the IBM token-ring adapter shows only the I/O settings, but this adapter also uses IRQ and Memory resources.

Configuring Advanced Properties for Network Adapters

You can configure Advanced properties for both Plug and Play and non-Plug and Play devices. The options available in the Advanced tab of the Properties dialog box vary, depending on the type of network adapter. For information about specific settings that appear for a selected network adapter, see the documentation provided by the manufacturer for the adapter and driver. The manufacturer can also provide guidelines for when to change the default values for advanced configuration options.

To specify advanced settings for the network adapter

1. In Control Panel, double-click Network.

2. In the list of installed components, click the network adapter and then click the Properties button.

3. In the properties for the selected network adapter, click the Advanced tab.

   The following figure shows the advanced options for an Intel EtherExpress network adapter.

    

4. To change these values, select an item in the Property list, and then select a setting in the Value box.

5. Click the OK button.

6. Restart the computer.

Table 15.22 shows some typical settings for general types of network adapters. Network adapters that use the new, fast Ethernet technology might have many more settings. Examples of such adapters include SMC EtherPower 10/100 (9332) PCI Ethernet Adapter, DEC Etherworks 435, and Intel EtherExpress PRO/100. For an explanation of specific settings for a network adapter, see the documentation provided by the adapter manufacturer. You should not change the default settings unless you know they are incorrect.

Table 15.22 Typical network adapter settings 

Binding Network Adapter Drivers to Protocols

For a protocol to communicate with each network adapter on your computer, the network adapter driver must be bound to the protocol. (This applies to both Plug and Play and non-Plug and Play network adapters.) The bindings define the relationships between networking software components. Windows 98 automatically binds the appropriate protocols to the network adapter.

You might want to change the bindings if you have multiple network adapters and you do not want to use a specific protocol with a particular network adapter. Or, if the computer is on a local area network and is also connected to the enterprise internetwork, you might not want the computer's shared resources to be seen on the internetwork. In that case, you can disable the binding between the related protocol and the adapter that connects the computer to the internetwork.

To configure bindings between a network adapter and installed protocols

1. In Control Panel, double-click Network.

2. In the list of installed components, click the network adapter and then click the Properties button.

3. In the properties for the selected network adapter, click the Bindings tab.

   In the list, the protocols that are bound to the selected adapter are checked. If a particular protocol does not appear in the list, check that it is installed correctly by returning to the Configuration tab of the Network dialog box and reinstalling it.

4. If you do not want this network adapter to be bound to a particular protocol, clear the check box beside the protocol.

Setting LAN Adapter Numbers

NetBIOS defines the interface between the network client and the protocol layers using a set of function calls that allow an application to use the protocol services. Because many network applications use NetBIOS to send commands to the protocol driver, the NetBIOS interface is supported by all protocols provided with Windows 98.

Each combination of a NetBIOS network protocol and a network adapter forms a logical network over which computers can communicate with each other. For example, a computer can have a token-ring adapter and an Ethernet adapter, and might use NetBEUI on the token-ring network and both NetBEUI and TCP/IP on the Ethernet network. In this case, the computer is connected to three logical networks, each of which is assigned a NetBIOS LAN adapter (LANA) number that Windows 98 uses for communication.

When Windows 98 uses multiple protocols, it transmits data first using one protocol, then again using the next protocol, and so on. When multiple protocols are installed on a computer, the first protocol to be used is called the primary protocol.

On a computer running Windows 98, each binding of a protocol to a network adapter has a LAN adapter number assigned to it. (For example, one protocol bound to two network adapters requires two LAN adapter numbers; two protocols each bound to two adapters requires four LAN adapter numbers.)

In Windows 98, LANA numbers are assigned dynamically in sequence of binding order for the protocols, beginning with 7 and then 0, 1, and so on. This accommodates dynamic Plug and Play events such as the removal of a network adapter while the computer is running. If you are running Windows 98 in such a dynamic environment, Windows 98 cannot guarantee that a given protocol will receive the same LANA number each time the system is started. If the computer's network hardware never changes, the LANA numbers might not change at each startup. However, the default protocol is always LANA 0.

You need to change a LAN adapter number only if you have a NetBIOS application that needs to know the LANA number. For example, Lotus Notes requires that you enter the LANA number that Lotus Notes will use. To configure Windows 98 to use Lotus Notes, set the default protocol to be the NetBIOS-based protocol on which you want to run Lotus Notes. (Setting the default protocol makes it LANA 0.) This protocol can be NetBEUI, IPX/SPX-compatible with NetBIOS support, or TCP/IP.

To select a default protocol for LANA settings

1. In Control Panel, double-click Network.

2. Double-click the protocol you want to be the default.

3. In the protocol's Properties dialog box, click the Advanced tab.

4. Click the option named Set this protocol to be the default protocol so that the check box is checked. Then click OK, and shut down and restart the computer for the changes to take effect.

Configuring Asynchronous Transfer Mode and LAN Emulation

Transport protocols such as TCP/IP and IPX are connectionless. That is, they do not need to establish a connection before transmitting data; computers using those protocols simply send packets out to the network. However, asynchronous transfer mode (ATM) is connection-oriented and must establish a connection before transmitting any data. LAN Emulation provides the bridge between the connectionless protocols and ATM, enabling them to function transparently over ATM networks.

This section describes how to configure LAN Emulation on Windows 98 computers.

For more information about the ATM and LAN Emulation architecture, see the section "Architecture for NDIS" in Chapter 29, "Windows 98 Network Architecture."

The following procedure describes how to configure a computer as a LAN Emulation client, so it can communicate with other devices on the ATM network.

Note Windows 98 Setup automatically installs TCP/IP only if it detects a network adapter. However, it does not automatically install TCP/IP on computers that have only an ATM card. Thus, if you want to use TCP/IP over the ATM network, you must also install TCP/IP by following the procedures outlined in "Installing Microsoft TCP/IP" earlier in this chapter. If you want to use another network protocol instead, you must install it by following the procedures outlined in the section for that protocol, earlier in this chapter.

To configure a Windows 98 computer as a LAN Emulation client

1. In Control Panel, double-click Network.

2. On the Configuration tab, click the Add button.

3. In the Select Network Component Type dialog box, click Protocol, then click the Add button.

4. In the Manufacturers box, click Microsoft.

5. In the Network Protocols box, click ATM Call Manager, and then click OK.

6. Repeat steps 2 through 4 one time.

7. In the Network Protocols box, click ATM Emulated LAN, and then click OK.

8. Repeat steps 2 through 4 one time.

9. In the Network Protocols box, click ATM LAN Emulation Client and then click OK.

The previous procedure configures the computer to be part of the default emulated LAN. An emulated LAN is a virtual network that acts like a traditional LAN. ATM networks can consist of one or more emulated LANs: for example, a network administrator might want to create different emulated LANs for users in the Marketing and Accounting departments. However, all ATM networks have one default emulated LAN, which is the emulated LAN that all computers belong to unless they have been specifically configured as part of a different emulated LAN. Thus, after you have configured your computer as a LAN Emulation client, it automatically belongs to the default emulated LAN and can communicate with other devices on the ATM network.

In some circumstances, the ATM network administrator might configure additional emulated LANs. If so, and if you are given the name of an emulated LAN to which you should belong, you can configure your computer to be part of that emulated LAN.

Note When you are using ATM LAN Emulation, any network protocol that you will use on the emulated LAN must be installed on your computer. If you have not yet installed the network protocol you will use, you should do so now. For more information about installing network protocols, see the section for that protocol, earlier in this chapter.

To configure your computer as part of an Emulated LAN other than the default LAN

1. In Control Panel, double-click Network.

2. In the list of installed components, select the network adapter binding for ATM LAN Emulation Client.

3. In the Value box, enter the name of the emulated LAN.

Technical Notes on Network Adapters

NDIS 3.0 network adapter drivers that worked with Windows for Workgroups 3.11 do not work under Windows 98. You must use an NDIS 2.x real-mode driver, an ODI driver, or an updated NDIS 3.1 or later protected-mode driver for the network adapter. The driver must have a Windows 98 INF file. Many real-mode drivers, updated protected-mode drivers, and supporting INF files are included with Windows 98.

Troubleshooting Network Adapters and Protocols

This section contains information about troubleshooting problems related to network protocols. For general information about troubleshooting the network installation, including how to use net diag, see Chapter 14, "Introduction to Networking Configuration." For information about troubleshooting procedures and tools provided with Windows 98, see Chapter 27, "General Troubleshooting."

Correcting Problems with Network Adapters

This section describes specific problems you might encounter using network adapters with Windows 98.

For information about specific network adapters, see the Windows 98 Adapter Card Configuration Help File on the Windows 98 Resource Kit compact disc.

For general troubleshooting steps, see the "Troubleshooting Basic Networking Configuration" section in Chapter 14, "Introduction to Networking Configuration."

Plug and Play card does not function with 16-bit network drivers. 

If you are running Plug and Play with a 16-bit, real-mode driver, your Plug and Play network adapter might not function properly. If this happens, disable Plug and Play support for that network adapter.

To disable Plug and Play support for a network adapter

1. Run the software setup utility that comes with your Plug and Play card.

2. Set the card to non-Plug and Play Mode.

3. In Control Panel, double-click System, and then click the Device Manager tab.

4. Expand the Network adapters tree, and then select your network card.

5. Click Remove.

6. In Control Panel, double-click Add New Hardware.

7. Manually reinstall the network adapter.

   Note For some network adapters, you will also need to install a non-Plug and Play driver.

If you later need to re-enable Plug and Play support, follow these steps:

To re-enable Plug and Play support for a network adapter

1. In Control Panel, double-click System, and then click the Device Manager tab.

2. Remove the network adapter.

3. Run the software setup utility that comes with your Plug and Play card.

4. Set the card to Plug and Play Mode.

5. Reboot the computer. Windows 98 automatically detects the network adapter.

Correcting Problems with TCP/IP

This section describes TCP/IP utilities that you can use to diagnose TCP/IP connectivity problems. Next, it discusses general troubleshooting methods to solve TCP/IP connectivity problems. Finally, it lists common problems you might experience.

Using Diagnostic Utilities to Troubleshoot

Use the TCP/IP diagnostic utilities included with Microsoft TCP/IP to diagnose connectivity problems. The following list describes which MS-DOS utility helps to identify various problems. For a more detailed description of these utilities, see "TCP/IP Utilities" earlier in this chapter.

Table 15.23 Using TCP/IP diagnostic utilities 

For more information about these utilities, see "TCP/IP Utilities" earlier in this chapter.

You can also use the graphical IP Configuration utility winipcfg to display, update, or release TCP/IP configuration values, including DHCP values.

Winipcfg can also show you whether or not your computer is using automatic private IP addressing for its IP address. For more information about automatic private IP addressing, see "Configuring IP Addresses Using Automatic Private IP Addressing" earlier in this chapter.

To use winipcfg

1. Click Start, click Run, and then type:

   winipcfg 

2. The resulting screen identifies your IP address and the IP address of your default gateway.

3. Click More info.

4. The resulting screen tells you your IP address, subnet mask, and default gateway for each of your network interfaces. It also shows your DNS and WINS settings.

To test TCP/IP using ping

• Check the loopback address by typing ping 127.0.0.1 and pressing ENTER at the command prompt. The computer should respond immediately. To determine whether you configured IP properly, use ping with the IP address of your computer, your default gateway, and a remote host. (If you are using DHCP, use winipcfg to find the IP address.)

If you cannot use ping successfully at any point, verify the following:

• The computer was restarted after TCP/IP was installed and configured.

• The local computer's IP address is valid and appears correctly in the TCP/IP Properties dialog box.

• The IP address of the default gateway and remote host are correct.

• IP routing is enabled on the router, and the link between routers is operational.

• The local computer's registry includes an entry for **lmhosts=c:\**directory that correctly indicates the location of LMHosts.

If you can ping other computers running Windows 98 on a different subnetwork but cannot connect using Windows Explorer, net use \\server\share,* *or net view \\server\share, verify the following:

• The host is functioning.

• The correct host computer name was used.

• Using tracert, make sure that a router path exists between your computer and the target computer.

• LMHosts contains correct entries, so the computer name can be resolved.

• The computer is configured to use WINS, the WINS server addresses are correct, and WINS servers are functioning.

You can use winipcfg to verify your TCP/IP configuration settings. For information about using winipcfg, see "Using Diagnostic Utilities to Troubleshoot," earlier in this chapter.

Troubleshooting TCP/IP Gateway Problems

If a host computer on one subnet has problems communicating to computers on another subnet when using TCP/IP, the following information can help you determine whether the problem is with the gateway. For these troubleshooting steps, it does not matter whether the gateway that provides routing capabilities is a hardware router or a Windows NT or UNIX server configured to act as a router.

First, to troubleshoot gateway problems, you need a network map, plus the IP addresses and subnet masks for the host computer with problems, the near side of the hardware or software router that acts as the gateway, the remote side of the gateway, and the destination computer (node). For example, Figure 15.6 shows two subnets connected by a router.

Figure 15.6 Two subnets connected by a router 

The following example of troubleshooting steps uses the IP addresses from this illustration.

To troubleshoot possible gateway (router) problems

1. Use ping to access the host computer that is having problems communicating outside the subnet. For example:

   ping 172.22.4.66
   

If this works, this host is probably healthy at the IP level.

If this does not work, use the usual methods to check the IP configuration and network connections on this host.

2. If the problem is not solved, use ping to access the near side of the router (that is, the default gateway). For example:

   ping 172.22.3.1
   

If this works, this side of the router is healthy.

If this does not work, use the usual methods to check the actual IP configuration and network connection for the near side of the router. Adjust the gateway settings on the problem host computer, if required.

Notice, however, that if you can use **ping** to get a response from this address, it does not necessarily mean that this is actually a router.

3. If the problem is not solved, use ping to access the far side of the router. For example:

   ping 172.22.4.25
   

If this works, the router is working.

If this does not work, have another user use the same **ping** command from the destination node (172.22.4.66 in the example).

If this works, the router is not working correctly.

4. If the problem is not solved, use ping to access the remote host. For example:

   ping 172.22.4.66
   

If this works and all problems in the previous steps have been resolved, TCP/IP should be working fine.

If this does not work, check the IP configuration and network connections on the destination computer. Typically, at this point the problem is that the remote computer has not route configured back to the original host computer. That is, the remote computer's routers or routing table does not contain the information necessary to send packets back to the original host computer.

When troubleshooting router connections, note the following:

• Do not use the host name when you are testing the router; instead, use the IP address. This will avoid any problems related to the Hosts or LMHosts files, DNS server, WINS server, or any other methods of name resolution.

• In most cases, the subnet mask should be the same for all hosts on the same side of the router.

• There could be two routers at separate sites performing the same job as described above; if this is the case, treat this situation the same as above, keeping in mind that each router could have a near and far side, depending on the configuration.

• If multiple routers exist between the source and destination, use tracert to see an ordered list of routers used. Tracert can also help you determine how many hops there are between the source and destination, and how much time it takes a packet to travel from the source to the destination.

Tip for using SNMP for routing The TCP/IP utilities use the public interface provided by inetmib1.dll in both Windows NT and Windows 98. This API can be used with Simple Network Management Protocol (SNMP) for actions such as setting and getting routing information programmatically on Windows 98 computers. For information about the management information base (MIB) object types provided for Microsoft networking, see the Microsoft Windows NT Server Resource Kit (for Microsoft Windows NT Server version 4.0).

Notice, however, that you cannot use a Windows 98 computer to run the SNMPUTIL tool provided with the Windows NT Resource Kit utilities.

Troubleshooting Other TCP/IP Problems

This section describes how to troubleshoot other problems with TCP/IP.

Windows 98 does not retain primary WINS server IP address. 

If the setting for your primary WINS server IP address is not retained when you reboot, check that a secondary WINS server IP address is also configured. If not, add a secondary WINS server IP address.

Windows 98 does not send DHCP request packet. 

If the DHCP key in the registry contains eight MAC address entries, Windows 98 cannot create a new entry for the current session and will not send a DHCP request packet. If this happens, use the Registry Editor to remove all keys except for the Dhcpinfo00 key from the following registry entry:

HKEY_LOCAL_MACHINE \System \CurrentControlSet \Services \VxD \DHCP 

Then restart the computer.

Windows 98 detects an IP address conflict. 

If your Windows 98 computer uses static IP addressing (described in the section "Manually Configuring IP Addresses" earlier in this chapter) and you receive a message that Windows 98 has detected an IP address conflict, another computer on the network is using the same IP address and you must change the IP address.

To change your IP address

1. In Control Panel, double-click Network.

2. Click your TCP/IP protocol.

3. Click the Properties button.

4. Click the IP Address tab.

5. Configure the protocol to use a different IP address that is not already in use on the network.

Cannot log on to Windows NT domain. 

If you cannot log on to a Windows NT domain, check that your computer uses NetBIOS name services. Windows 98 needs NetBIOS name services to log on to a Windows NT domain.

IP address connects but host names do not. 

Verify that the Hosts file and DNS settings have been configured for the computer by checking settings on the DNS Configuration tab.

• If you are using a Hosts file, verify that the DNS host name of the remote computer is identical—especially in terms of spelling and capitalization—to the name in the file and the application using it.

• If you are using DNS, verify that the IP addresses of the DNS servers are correct and in proper order. Use ping with the remote computer, and type both the host name and IP address to determine if the host name is resolved properly.

Use the netstat -a command to show the status of all activity on TCP and UDP ports on the local computer. A good TCP connection is usually established with 0 bytes in the send and receive queues. If data is blocked in either queue or if the state is irregular, there might be a problem with the connection. If not, you are probably experiencing network or application delays.

Connect times are long after adding to LMHosts. 

You might experience long connect times with a large LMHosts file if there is an entry at the end of the file. If so, mark the entry in LMHosts as a preloaded entry by following the mapping with the #PRE tag, or place the mapping higher in the LMHosts file. Then use the nbtstat -R command to update the local name cache immediately. The LMHosts file is parsed sequentially to locate entries without the #PRE keyword. You should place frequently used entries near the top of the file, and place the #PRE entries near the bottom**.** 

You see a message stating computer is unable to connect to a server. 

This message appears if name resolution fails for a particular computer name. If the computer is on the local subnetwork, confirm that the target server name is spelled correctly and that the target server is running TCP/IP. If the computer is not on the local subnetwork, be sure that its name and IP address mapping are available in the LMHosts file or the WINS database. If all TCP/IP elements appear to be installed properly, use ping with the remote computer to be sure that its TCP/IP software is working.

Use the nbtstat -n command to determine what name (or names) the server registered on the network. The nbtstat command can also display the cached entries for remote computers from either #PRE entries in LMHosts or recently resolved names. If the remote computers are using the same name for the server, and the other computers are on a remote subnetwork, be sure that they have the computer's mapping in their LMHosts files.

Correcting Problems with NetBEUI and IPX/SPX

This section describes problems that might occur with the NetBEUI and IPX protocols.

You cannot connect using NetBEUI

• Use net diag to test for NetBIOS connectivity over the LANA that NetBEUI is using. If it fails, check the transceiver type, cabling, and adapter.

• Check the NetBEUI protocol bindings.

• Verify that routing is not involved.

A NetBIOS application fails to start

This might be because the application is hard-coded to use the protocol on LANA 0 (such as Lotus Notes). You can force a particular protocol to always occupy LANA 0 by selecting it as the default protocol, as described in "Setting LAN Adapter Numbers" earlier in this chapter.

You cannot connect using the IPX/SPX-compatible protocol

Verify that both computers trying to connect are using the same frame type and that other settings are correct for this protocol.

• Verify the following in the Advanced tab of the IPX/SPX-compatible Protocol Properties dialog box, as described in "Configuring the IPX/SPX-compatible Protocol" earlier in this chapter:

  • The correct frame type is set. The recommended setting is Auto, but this frame type only checks SAP broadcast traffic on the network and might be selecting an incorrect frame type in a mixed frame-type environment.

  • Source Routing is enabled and a cache size is set if needed.

  • The option named Force Even Length Packets is set properly. NetWare servers with older NetWare Ethernet drivers or older IPX routers may require even-sized packets. If required, change this setting to make sure the computer transmits only even-length IPX frames.

• On the IPX routers, check the setting for Type 20 Packets (NetBIOS packets). When using NetBIOS over IPX, the IPX packet type is set to 14h (decimal 20). Manufacturers of routers might consider all NetBIOS traffic as being nonroutable LAN traffic even when carried over the routable IPX protocol, and so, by default, will not pass Type 20 NetBIOS IPX packets. To use NetBIOS over IPX connectivity, Type 20 packet passing must be enabled on the router.

• Use net diag to test for IPX connectivity over the related LANA number used by NetBIOS over IPX.

• Use System Monitor to view statistics for the IPX/SPX-compatible protocol. Then retry network operation and check the activity. If there is none, remove and reinstall the protocol, and then retry and retest the operation.

Correcting Connection Problems with MS 32-bit DLC

This section describes common problems you might encounter while using the 32-bit DLC protocol.

Windows 98 reports a Dlchlp error message. 

This error message appears if you have removed or renamed your Autoexec.bat file, or removed the reference to Dlchlp.exe from your Autoexec.bat file. To correct this error, add the Dlchlp.exe entry to your Autoexec.bat file.

Setting 32-bit DLC as the default protocol does not change the LANA number for NetBIOS protocols. 

The 32-bit DLC protocol does not use a LANA number, so there is no reason to set DLC as the default protocol.

The 32-bit DLC protocol does not bind to the Dial-Up adapter. 

Currently, the Microsoft 32-bit DLC protocol cannot bind to the Dial-Up adapter. When the Dial-Up adapter is installed and you then install the 32-bit DLC protocol, the 32-bit DLC protocol appears to be bound to the Dial-Up adapter. However, after you restart your computer, Windows 98 removes the binding. This is by design.

You cannot use Extra SAPs and Extra Stations settings. 

Manual configuration of Extra SAPs and Extra Stations is no longer necessary with the Microsoft 32-bit DLC protocol.

After removing an adapter driver, programs cannot connect to the host. 

When your computer is configured with multiple network adapters, removing an adapter driver may prevent the DLC program from connecting properly to the host. The connection fails when the removed network adapter driver is the adapter for which the DLC program is configured. To prevent this problem, if the DLC program is configured for the primary adapter, make sure the CCB Adapter Num is set to 0. If the DLC program is configured for the alternate (or secondary) adapter, make sure the CCB Adapter Num is set to 1. The 32-bit DLC protocol allows CCB2 adapter numbers 0 through F.

You cannot connect to the host over Ethernet after adding the 32-bit DLC protocol. 

The most common issues you will face when using the DLC protocol on Ethernet are problems with bit swapping and Ethernet DIX. Check to see whether the host uses frame type 802.3 or Ethernet DIX 2.0, and whether address swapping is required on your network. For more information about configuring the 32-bit DLC protocol, see "Configuring the 32-bit DLC Protocol" earlier in this chapter.

After you have installed the 32-bit DLC protocol on a token-ring network adapter, Windows 98 stops responding. 

Sometimes a token-ring network adapter has an RcvBufsize setting that is too small for the size of the frame being sent across the wire. If this happens, restart Windows 98 in Safe Mode, and increase the RcvBufSize setting for the network adapter driver. For more information about the appropriate settings for your network adapter, see the documentation for your network adapter. For information on getting to Safe Mode, see Chapter 27, "General Troubleshooting."

Correcting Connection Problems with MS 16-bit DLC

If you encounter problems using the real-mode Microsoft DLC protocol, check the following items:

• When adding the Microsoft DLC protocol, make sure that there is only one net start entry in Autoexec.bat. Setup adds an entry for Microsoft DLC, even if an entry already exists. Also, make sure that Autoexec.bat is configured properly, as described in "Installing and Configuring Real-Mode Microsoft DLC" earlier in this chapter.

• Do not make direct entries or changes in Protocol.ini for Microsoft DLC. Instead, make all changes in the Advanced properties for Microsoft DLC by using the Network option in Control Panel. If you make changes directly in Protocol.ini, the next time you change any values by using the Network option in Control Panel, all settings in Protocol.ini for Microsoft DLC will be overwritten.

• Some terminal emulation applications use TSRs to communicate with the Microsoft DLC protocol. If your emulation application uses a TSR that runs from Autoexec.bat, make sure that the entry for the TSR still exists (usually, the TSR entry occurs after the msdlc and net start lines). Windows 98 Setup removes or comments out many TSRs.

• The option named Set This Protocol To Be The Default Protocol in the Advanced properties for Microsoft DLC does not provide any functionality. This option should not be checked, because Microsoft DLC does not use LANA settings.

• Real-mode Microsoft DLC is an NDIS 2 protocol, so the network adapter must have an NDIS 2-compatible network adapter driver for use with Windows 98. Windows 98 includes many compatible drivers, but some Windows 98 drivers for certain PCI and PC Cards do not have a corresponding NDIS 2 driver to allow loading real-mode Microsoft DLC.

Additional Resources