IPX Addressing and Address Structure
Fundamentals of the address and network structure of the IPX protocol.
Configuring IPX Addresses
Overview of the IPX addressing scheme plus address configuration examples
for different LAN and WAN interface types.
IPX Routing Configuration
Basics of IPX routing configuration using static routes and verifying
Configuring IPX Routing Protocols
Characteristics of the IPX, RIP, and NLSP dynamic routing protocols
and basic configuration examples.
Configuring IPX Filtering via Access Lists
Controlling network access and security through the use of access-list
and ipx access-group commands.
Configuring Basic IPX Dialup Services
Setting up IPX client asynchronous dialup connectivity.
Verifying IPX Connectivity and Troubleshooting
Identifying connectivity problems through the use of show,ping,
and debug commands.
Configuring IPX Type 20 Packet Forwarding
Options for configuring the IOS to forward IPX type 20 packets.
In the late 1970s, Xerox
created a network protocol called Xerox Network Systems, or XNS,
that was widely implemented by most major LAN vendors, including Novell, Inc.
Novell made some changes to the protocol in the early 1980s, renamed it Internet
Packet Exchange protocol (IPX), and incorporated it as part of NetWare. The
NetWare transport layer protocol, the Streams Packet Exchange (SPX), was also
derived from the XNS protocol suite.
NetWare is a suite
of protocols for sharing resourcesprimarily print and file servicesamong
workstations through a client-server implementation. Novell describes NetWare
as a network operating system (NOS) because it gives end users access to resources
that are available via the LAN or WAN. NetWare, a dominant corporate NOS,
is widely deployed in many internetworks.
Figure 6-1 shows the multiple
protocols commonly used in the NetWare protocol suite. We do not cover each
of these protocols in this chapter but instead concentrate on explaining the
protocols on the network and transport layers, namely IPX, IPX RIP (Routing
Information Protocol), NLSP (NetWare Link State Protocol), SAP (Service Advertisement
Protocol), and SPX. The other protocols shown in Figure
6-1 reference other internetwork technologies with which you
may be familiar.
IPX is a network layer protocol with its own proprietary addressing
structure. This section introduces the IPX address structure that each IPX
client (sometimes called a workstation by NetWare documentation)
or server must have in order to communicate with other IPX devices on an internetwork.
An IPX address has two components, a 32-bit network
component that applies to a given LAN or WAN segment and a 48-bit node
component that uniquely identifies a client or a server. These two components,
expressed together as network.node,
are written using hexadecimal format. The two-layer hierarchy of the IPX address
structure makes this addressing scheme scalable for internetworks, yet not
as scalable as the multiple hierarchies of the IP addressing structure.
The network administrator assigns the network number for an IPX network
segment in the same way that he/she selects IP subnets for given LAN and WAN
segments. All IPX clients, IPX servers, and Cisco routers on the same LAN
or WAN segment must have the same network number.
NetWare servers have internal IPX
network numbers that are different than the IPX network numbers for any LAN
or WAN interface. The internal IPX network number is used as the source network
number for NetWare services on the server. We'll discuss service advertisement
later in this chapter when we discuss SAP. A Cisco router can be configured
with an internal IPX network number using the global configuration command ipx internal-network. We discuss this internal network
number further in the section on NLSP.
Each IPX server or client needs to have a unique node number on a LAN
or WAN segment. Typically, IPX clients derive this unique number by reading
the 48-bit data link address on their LAN interface and then using that number
as their unique network layer node address. Although the LAN interface data
link address is the same as the IPX node address, do not conclude that a client
uses these two addresses in the same manner. The data link layer address is
used for data link layer encapsulation, such as ethernet or Token Ring. The
IPX node address is the second portion of the IPX network layer network.node
address for a given client.
Using the data link layer address to determine a unique 48-bit IPX node
address is not required by the IPX protocol. You can have a node address that
does not match a data link layer address as long as the node address is unique
on a given IPX network. For example, we have seen that an IOS device can have
multiple LAN interfaces. When IPX routing is enabled, the IOS device chooses
the data link layer address on the first LAN interface in the device as the
unique node address for all IPX network segments. Now imagine that the data
link layer address on Ethernet 0 of a router is 0000.0c11.12ab. If Ethernet
0 is the first LAN interface in this router and if the router is connected
to IPX network 10 and IPX network 20, the router is seen as 10.0000.0c11.12ab
on IPX network 10 and 20.0000.0c11.12ab on IPX network 20.
The IOS global configuration command ipx routing enables IPX routing in an IOS device.
The device automatically chooses an IPX node number based on the first LAN
interface when this global configuration command is configured. In the following
example, we enable IPX routing on the SF-2 router on the ZIP network:
Configuring from terminal, memory, or network [terminal]?
Enter configuration commands, one per line. End with CNTL/Z.
If your router does not have a LAN interface, you must configure a unique
IPX node address as an optional parameter to the ipx
The use of the data link layer address to determine the IPX node address
simplifies the job of the network administrator because IPX clients do not
need manual configuration. Also, this mapping of data link layer address to
network layer address can eliminate the need for a separate protocol to map
between the addresses on these two layers, such as ARP, which is discussed
in Chapter 4, “TCP/IP Basics.”