9-1 Novell IPX Routing

• Novell IPX is the protocol used by Novell servers (prior to Version 5.0) for communication between clients and servers.

• Novell IPX addresses are 80 bits in length. They consist of a 32-bit administrator-assigned network address and a 48-bit node address.

• For LAN interfaces, the node number is the device's MAC address.

• Novell LAN interfaces can use a variety of Layer 2 encapsulations . The encapsulation must be defined on the router to correctly match that of the server.

• Novell IPX has three possible routing protocols: IPX RIP, IPX EIGRP, and NLSP.

Configuration

1. (Required) Enable the IPX process:

    (global)  ipx routing  [  nodenumber  ] 

   This command enables the IPX process and the IPX RIP routing protocol. The optional node number allows a 48-bit node number to be specified with the process. This node number is written in hexadecimal form (0000.0000.0001). It becomes the IPX node number for any router interface that does not have a Layer 2 MAC address. If you do not select a node number, the router automatically uses the node number of the first LAN interface on the router.

2. (Recommended) Configure IPX for load balancing:

    (global)  ipx maximum-paths   number  

   This command specifies how many equal-cost paths will be used by the routing protocol for load-balancing purposes.

3. (Required) Configure the interfaces with IPX network numbers , and select the encapsulation:

    (interface)  ipx network   number  [  encapsulation   encapsulation-type  ] 

   This command places a network number on the interface, which enables IPX on that interface. IPX numbers are assigned by the Novell server administrator when the server is installed. Each IPX network number in a campus network must be unique. As soon as you have specified the network, you can also specify the encapsulation type. The encapsulation must match that of the server or any other routers on the segment. Table 9-1 lists the Novell encapsulations for LAN interfaces, their default encapsulations, and the Novell server names .

Table 9-1. Novell Encapsulation Types

Interface Type	Cisco Encapsulation Options	Novell Server Encapsulation Equivalents
Ethernet	novell- ether [1]	Ethernet_802.3
	sap	Ethernet_802.2
	snap	Ethernet_Snap
	arpa	Ethernet_II
Token Ring	sap [1]	Token-Ring
	snap	Token-Ring_SNAP
FDDI	snap [1]	FDDI_Snap
	sap	FDDI_802.2
	novell-fddi	FDDI_RAW

^[1] Indicates default encapsulation

NOTE

It is possible for a single network wire to have multiple encapsulation types and networks running over that wire. If you need to configure multiple networks on a single LAN interface, use subinterfaces and configure the appropriate network and encapsulation on each interface. See IPX Example 1 for subinterface configuration details.

4. (Optional) Configure NLSP as the routing protocol.

   The following steps are required to configure NLSP routing. This assumes that IPX routing has already been configured, as described in Steps 1 and 3.

   1. (Required) Configure an internal IPX network number on the router:

       (global)  ipx internal-network   network-number  

      Each NLSP router uses its internal network number to identify itself and its routing updates. The internal IPX network number must be unique and is a 32-bit hexadecimal address. An example of an internal network number is 2B. Any leading bits of 0 are not needed when identifying IPX network numbers.

   2. (Required) Enable the IPX NLSP routing process:

       (global)  ipx router nlsp  [  tag  ] 

      The tag option allows you to specify a unique NLSP process. The tag is optional if the router has only one process. A maximum of three NLSP processes may be configured on the router at the same time. The tag can be any combination of printable characters .

      NOTE

      NLSP is a link-state routing protocol that defines areas to create a routing hierarchy. There are two versions of NLSPversion 1.0 and version 1.1. If a router is a part of only one area, only a single routing process needs to be defined, regardless of the version. If multiple areas are connected to a router for hierarchical purposes, you must define multiple NLSP processes to discover, select, and maintain route information about the areas they interconnect. These processes collectively compute the router's routing table. Tags are used to define NLSP processes.

   3. (Required) Specify the IPX networks that are to be a part of the NLSP process:

       (global)  area-address   address mask  

      This command specifies which networks are in a given area by giving an address and mask. Depending on the mask, the address defines the prefix of the networks in an area. The address is the prefix or beginning address for all the networks in a given area. The mask is 1s represented in hex for the bits of the address that should mask. Therefore, if you wanted all the addresses from 11110000 to 1111ffff, you would use an address of 11110000 with a mask of FFFF0000. If you wanted to define all networks on the router in a given area, you would use an address of 0 and a mask of 0.

      NOTE

      As a routing protocol, the concept behind NLSP was to allow a router to belong to three different areas and then communicate with different levels of routers. When Novell embraced IP as a mechanism for communicating between servers, IPX became a legacy protocol, and little future development was done for IPX and NLSP. Although it is possible to use masks to define different areas, it is more common for all the networks on the router to belong to a single area.

   4. (Required) Enable NLSP on the router interface:

       (interface)  ipx nlsp  [  tag  ]  enable  

      This command turns on the NLSP routing process on a given interface. The tag option is used to bind the interface to a specific NLSP process, as shown in Step 4b.

      NOTE

      When you enable the NLSP routing process, all RIP-learned and -connected IPX routes are automatically redistributed in the IPX NLSP tables and are advertised out the interfaces configured to run NLSP in Step 4d.

   5. (Optional) Control RIP updates on an interface:

       (interface)  ipx nlsp  [  tag  ]  rip  [  on   off   auto  ] 

      This command controls how RIP updates are handled on an interface running NLSP. Table 9-2 shows the settings for each option. The default value is auto. The tag option is used to bind the interface to a specific NLSP process.

   6. (Optional) Control SAP updates on an interface:

       (interface)  ipx nlsp  [  tag  ]  sap  [  on   off   auto  ] 

      This command controls how SAP updates are handled on an interface running NLSP. Table 9-3 shows the settings for each option. The default value is auto. The tag option is used to bind the interface to a specific NLSP process.

   7. (Optional) Set the NLSP priority on an interface:

       (interface)  ipx nlsp  [  tag  ]  priority   priority-number  

      This command controls the priority of an interface for the selection of a designated router on a LAN segment. The priority range is 0 to 127; the default is 44. The tag option is used to bind the interface to a specific NLSP process.

   8. (Optional) Set the NLSP hello timer on an interface:

       (interface)  ipx nlsp  [  tag  ]  hello-interval   seconds  

      This command controls the period in which hellos are sent out on an interface. The range is 0 to 1600. The default is 20 for a nondesignated router and one-half the configuration amount for a designated router (10 seconds by default). The tag option is used to bind the interface to a specific NLSP process.

Table 9-2. NLSP RIP Settings

Setting	Meaning
On	Always generates and sends RIP periodic traffic.
Off	Never generates and sends RIP periodic traffic.
Auto	Sends RIP periodic traffic only if another RIP router is sending periodic RIP traffic.

Table 9-3. NLSP SAP Settings

Setting	Meaning
on	Always generates and sends SAP periodic traffic.
off	Never generates and sends SAP periodic traffic.
auto	Sends periodic SAP traffic only if another SAP router is sending periodic SAP traffic.

5. (Optional) Configure EIGRP as the routing protocol.

   The following are the steps required to configure EIGRP routing. This assumes that IPX routing has already been configured as described in Steps 1 and 3.

   1. (Required) Configure an internal IPX network number:

       (global)  ipx router eigrp   autonomous-system-number  

      This configures the EIGRP process to route for IPX traffic. The autonomous-system-number is specified to determine which EIGRP routers will exchange information. Only EIGRP routers with matching AS numbers will communicate. This command also moves the user to ipx-router configuration mode for the entry of network numbers.

   2. (Required) Enable the IPX EIGRP process for specified networks:

       (ipx-router)  network  {  network-number   all  } 

      This command specifies which networks will use IPX EIGRP. When a network number is specified, EIGRP begins sending updates on the interface that is configured with that network number. Use the all option if you want to run EIGRP on all the interfaces.

      NOTE

      When you enable the EIGRP routing process, all RIP-learned and -connected IPX routes are automatically redistributed in the IPX EIGRP tables and are advertised out the interfaces associated with the network statement in Step 5b.

   3. (Recommended) Disable IPX RIP on EIGRP interfaces:

       (global)  ipx router rip  (ipx-router)  no network   network-number  

      This command specifies which networks will not use IPX. After you enable IPX routing and specify network numbers for an interface, IPX RIP runs by default on all interfaces that have been configured. On interfaces without IPX clients or servers (such as WAN interfaces), there is no need for IPX RIP. Use the no network command to disable IPX RIP routing on those interfaces.

   4. (Optional) Configure the percentage of bandwidth used by EIGRP for updates:

       (interface)  ipx bandwidth-percent eigrp   as-number percent  

      This command specifies how much of the interface bandwidth EIGRP can use, based on the interface bandwidth statement, for routing updates. The default is 50 percent.

   5. (Optional) Adjust the period that EIGRP uses to send out hello packets:

       (interface)  ipx hello-interval eigrp   autonomous-system-number seconds  

      This command specifies how often EIGRP sends out hello packets for a specified autonomous system. The defaults are 60 seconds for NBMA networks with T1 speeds or lower and 5 seconds for all other networks.

   6. (Optional) Adjust the period that EIGRP uses to send out hello packets:

       (interface)  ipx hold-time eigrp   autonomous-system-number seconds  

      This command specifies how long an EIGRP sender is considered valid. If no hellos are received in the period of the hold-time value, the sender is no longer valid, and all the routes received from that sender are no longer valid. The default hold time is 15 seconds. It should be set to at least 3 times the value of the configured hello time interval.

   7. (Optional) Specify how long a lost route is placed in the hold-down state:

       (interface)  ipx hold-down eigrp   autonomous-system-number seconds  

      The default is 5 seconds.

   8. (Optional) Disable the split horizon feature for IPX EIGRP routing:

       (interface)  no ipx eigrp-split-horizon  

      This command prevents the IPX EIGRP routing protocol from exercising split horizon. This is required when you are using EIGRP on a multipoint NMBA medium such as Frame Relay so that routes learned on the multipoint interface can be sent out to other routers out of that interface.

IPX Example 1

Figure 9-1 shows a network diagram for this example. Router 5 is running RIP on one LAN interface connecting to one IPX segment. The IPX routing process has been configured with a node number of 0000.0000.00005 to allow the serial interfaces to use a node number that corresponds to the router number. NLSP is configured to run on the other Ethernet interface and across the WAN links. The priority for NLSP has been adjusted on the Ethernet 1 interface to allow this router to become the designated NLSP router for that segment. The router has been configured to load-balance up to two equal-cost paths. The serial interfaces have also been configured to never send IPX RIP or SAP updates.

Figure 9-1. Network Diagram for IPX Example 1

  ipx routing 0000.0000.0005   ipx maximum-paths 2   ipx internal-network C001   !   interface ethernet 0   ipx network ACE   !   interface ethernet 1   ipx network BAD   ipx nlsp enable   ipx nlsp priority 80   !   interface serial 0   ipx network CAB   ipx nlsp enable   ipx nlsp rip off   ipx nlsp sap off   !   interface serial 1   ipx network 5DC0   ipx nlsp enable   ipx nlsp rip off   ipx nlsp sap off   !   ipx router nlsp   area-address 0 0  

IPX Example 2

Figure 9-2 shows a network diagram for this example. Router 6 is running IPX RIP on one of the two LAN interfaces. The IPX routing process has been configured with a node number of 0000.0000.00006 to allow the serial interfaces to use a node number that corresponds to the router number. EIGRP has been configured for operation on the WAN links. IPX RIP has been disabled on the WAN links, and split horizon has been disabled on a multipoint Frame Relay link. EIGRP has been configured to never use more than 45 percent of the s0 link's configured bandwidth.

Figure 9-2. Network Diagram for IPX Example 2

  ipx routing 0000.0000.0006   !   interface ethernet 0   ipx network A001   !   interface ethernet 1   ipx network B001   !   interface serial 0   ipx network 6108   bandwidth 56   ipx bandwidth-percent eigrp 101 45   !   interface serial 1   encapsulation frame-relay   interface serial 1.1 point-to-multipoint   ipx network 986   frame-relay map ipx 986.0000.0000.0002 622   frame-relay map ipx 986.0000.0000.0003 623   no ips eigrp-splithorizon   !   ipx router eigrp 101   network 6108   network 986   !   ipx router rip   no network 6108   no network 986