NWLink IPX/SPX Configuration in Windows 2000

This article provides an overview of the NWLink IPX/SPX/NetBIOS Compatible Transport Protocol in Microsoft Windows 2000 and examines the specific roles of its configuration parameters. Readers will learn the operational purpose of this legacy transport stack, including why it was implemented, how manual and automatic frame type detections function, and the technical necessity of parameters like internal network numbers and frame-specific network numbers.

NWLink was Microsoft’s native 32-bit implementation of Novell’s Internetwork Packet Exchange/Sequenced Packet Exchange (IPX/SPX) protocol suite. In Windows 2000, NWLink primarily enabled seamless connectivity with Novell NetWare servers and client networks. It was also widely used to run client-server applications designed specifically for IPX/SPX, as well as legacy multi-node network applications and games that lacked native TCP/IP support.

While NWLink functioned out of the box with default settings in small environments, enterprise networks required precise configuration to prevent communication failures and routing conflicts.

Configuring NWLink in Windows 2000 involves three critical options located in the protocol's properties sheet under Local Area Connection settings:

1. Internal Network Number

The Internal Network Number (also called the Virtual Network Number) is an eight-digit hexadecimal number assigned to identify a specific computer internally.

  • Purpose: It provides a unique address for routing when a computer hosts services that act as network servers (such as File and Print Services for NetWare or Microsoft Routing and Remote Access Service), or when the system has multiple network adapters installed.
  • Functionality: Unlike external network numbers assigned to physical wire segments, the internal network number belongs to the operating system itself. Setting this value to 00000000 (the default) indicates that the host relies on standard routing, but assignation of a unique non-zero hex value is mandatory if the Windows 2000 server routes traffic between multiple IPX networks.

2. Frame Type Selection

The frame type defines the exact structure and packaging of data packets transmitted across the physical Ethernet medium. Windows 2000 supports several Ethernet frame types:

  • Ethernet 802.2 (standard for NetWare 3.12 and later)

  • Ethernet 802.3 (raw, common in NetWare 3.11 and earlier)

  • Ethernet II (standard for TCP/IP and older NetWare variants)

  • Ethernet SNAP (Subnetwork Access Protocol)

  • Purpose: Network nodes on the same physical wire must share the exact same frame type to exchange IPX packets. If a Windows 2000 machine transmits using Ethernet 802.2 while a NetWare target listens on Ethernet 802.3, the machines cannot communicate.

  • Configuration Modes:

    • Auto Frame Type Detection: Windows 2000 scans the physical wire during startup. If it detects Ethernet 802.2, it defaults to it; otherwise, it assigns the first frame type detected. If no IPX traffic is seen, it defaults to 802.2.
    • Manual Frame Type Detection: Allows administrators to bind multiple frame types to a single network card simultaneously. This is necessary when communicating across a mixed NetWare infrastructure with diverse server versions.

3. Network Number (External)

Each frame type manually added to a network adapter requires an associated Network Number—an eight-digit hexadecimal address.

  • Purpose: The external network number identifies the specific physical subnet to which the adapter is attached.
  • Functionality: While Auto-Detect attempts to extract this number from existing broadcast traffic on the wire, entering it manually ensures that packets are delivered with the correct subnet header. Mismatching this value with the network segment’s existing NetWare router tables causes routing loops, unreachable subnets, and address conflicts.