Chapter 4 - Planning for Managing IP Interfaces
When you manage IP interfaces, your primary task involves using the TCP/IP configuration program to assign an IP address to each IP interface.
This chapter helps you determine addresses for the IP interfaces on your server, and shows you how to record them and other pertinent information on the IP Interfaces Worksheet. Chapter 8 explains how to use the TCP/IP configuration program to enter this information in the TCP/IP configuration. This chapter discusses these topics:
![]() |
Deciding Which Interfaces Need an IP Address |
![]() |
Assigning IP Addresses |
![]() |
Address Resolution and Broadcast Address Planning |
![]() |
Filling Out the IP Interfaces Worksheet |
![]() |
Filling Out the ARP Entries Worksheet |
Deciding Which Interfaces Need an IP Address
IP addresses can be assigned to the following interfaces:
![]()
Ethernet ![]()
IBM Token-Ring interface ![]()
ProNET-10 ![]()
The logical VINES interface, which is used only for PCs running the VINES Transport version of PC/TCP
Assign an IP address to interfaces that connect your server directly to foreign hosts or foreign host gateways. Examples of these hosts include mainframes and minicomputers.
PCs running the Ethernet-only version of PC/TCP must be connected directly to an IP gateway through an Ethernet interface. This gateway can be either a Foreign Host or a VINES server running the TCP/IP Routing option. An IP address must be assigned to the server's Ethernet interface.
For PCs running the VINES Transport version of PC/TCP, the server that provides them with routing services must have a unique IP address assigned to the VINES interface. The PCs must have IP addresses containing the same network and subnetwork numbers as those in the IP address assigned to the VINES interface.
The network manager for Server 1 in Figure 4-1 must assign an IP address to both Ethernet interfaces. VINES handles traffic on the HDLC interface.
Since a PC equipped with PC/TCP requires the services of the Routing option on Server 1, Server 1's network manager must also assign an IP address to the VINES interface.
Assign a unique IP address to each interface that IP will use to access a physical medium, and to the VINES interface, if necessary. These unique addresses simplify the job of managing routing. See Chapter 5 for more information.
For example, in Figure 4-2, Server 1 has four interfaces to physical media:
![]()
Two Ethernet interfaces ![]()
An HDLC Server-to-Server interface ![]()
A Token-Ring interface
The Token-Ring network and the HDLC Server-to-Server interface make up a small VINES network. Any IP devices in the VINES network are reached through the logical VINES interface.
The network manager must assign a unique IP address to three interfaces:
![]()
A unique IP address must be assigned to Server 1's VINES interface, since the workstations running the VINES Transport version of PC/TCP are using Server 1 as their default gateway. ![]()
Ethernet 1 needs an IP address, so that packets can reach Foreign Host 2. ![]()
Ethernet 2 needs an IP address, so that packets can reach Foreign Host 1.
Server 1's IP addresses are shown in Figure 4-2.
Network and Subnetwork Numbers for LAN Interfaces
Each host on a LAN segment should have a unique address. However, all hosts on a LAN segment should have the same network and subnetwork numbers in the IP addresses of their interfaces on that segment. Assigning the same network and subnetwork numbers enables the TCP/IP configuration program to define direct routes to the networks or subnetworks that the numbers identify. This eliminates the need to define these routes manually.
You may have to contact the network managers of other hosts on the segment to obtain these numbers. Record them on the IP Interfaces Worksheet. The section "Filling Out the IP Interfaces Worksheet," later in this chapter, provides information.
In Figure 4-3, the network managers of Server 1, Server 2, and Foreign Host 2 use the same network number/subnetwork number combination in the IP addresses they assign to their Ethernet interfaces: 160.134.1.
Using the IP address assigned to Server 1's Ethernet interface, and assuming an 8-bit subnetwork mask, the TCP/IP configuration program on Server 1 automatically defines a direct route to 160.134.1:
If Server 1's IP address did not contain the same network or subnetwork numbers as the IP addresses of the other hosts, the TCP/IP configuration program could not automatically create a useful direct route to 160.134.1.
In Figure 4-4, the TCP/IP configuration program on Server 1 has no knowledge of 160.134.1. It assumes that 125.125.1 is the network/subnetwork combination in the IP addresses of the other hosts, and automatically defines a route to 125.125.1:
IP on Server 1 cannot use the entry to route packets to the other hosts on the Ethernet segment. When IP on Server 1 receives packets for Server 2 or Foreign Host 2, IP sees that the addresses contain 160.134.1, resulting in a mismatch. In this event, Server 1 must have a direct route entry that defines a route to 160.134.1.
For example, the network manager of Server 1 would have to use the Manage Adjacent Networks option in the TCP/IP configuration program to manually create a route to 160.134.1:
Address Resolution and Broadcast Address Planning
The TCP/IP configuration program allows you to manage address resolution and IP broadcast addresses for interfaces to LANs. For address resolution, the TCP/IP configuration program allows you to:
![]()
Enable and disable Address Resolution Protocol (ARP) functionality for your server. ARP is enabled by default.
When ARP is disabled, the server does not dynamically resolve IP addresses to physical addresses. The server uses any information that you manually configure on the server to resolve addresses.
You should disable ARP only when no other hosts on a LAN implement ARP. If direct mapping is enabled, ARP is disabled automatically.
Note: When ARP is enabled, direct mapping is not automatically disabled.
![]()
Enable and disable direct mapping. This is allowed only for ProNET-10 interfaces. ![]()
Create ARP entries for hosts if your server is on a LAN where address resolution cannot be done dynamically. These entries consist of IP address/LAN address pairs. ![]()
Implement proxy ARP on your server.
For information on address resolution, see "Address Resolution" in Chapter 2.
Record all ARP entries that must be manually entered on the ARP Entries Worksheet. To obtain the IP addresses and LAN addresses, you may have to consult with network management personnel about other hosts on the LAN. LAN addresses tend to be either hard-coded on LAN cards, or set by jumper switches. The section "Filling Out the ARP Entries Worksheet" later in this chapter provides information.
The TCP/IP configuration program also enables you to change IP broadcast addresses. For each IP interface, the default outgoing IP broadcast address is composed of the network/subnetwork numbers and a host field of all 1s (signified by 255). If your server will be communicating with hosts running Berkeley's 4.2 BSD TCP/IP, you must specify 0 for the host field.
For interfaces that are assigned subnetted IP addresses, the configuration program lets you change the IP broadcast address so that only the network number is used in broadcast packets. Make this change if other hosts on the IP network attached to the interface do not understand subnetting. Otherwise, use the entire network and subnetwork numbers.
If all hosts on the attached network have the same network number but different subnetwork numbers, the broadcast address should contain only the network number.
If the attached network contains hosts with multiple network numbers, you must set the IP broadcast address to all 1s for the network and host fields (255.255.255.255).
Filling Out the IP Interfaces Worksheet
Use a blank IP Interfaces Worksheet to record the IP address for each interface on your server. You will need this information for the task of managing IP interfaces, as described in Chapter 8. This worksheet is shown in Appendix A of this guide.
Record the following information on the IP Interfaces Worksheet:
Server Name - The server to which the information on the worksheet applies.
Interface Name - EtherLink Plus card that provides an interface to Ethernet.
Slot - This is the physical slot number in which the LAN card is installed.
IP Address - The IP address of the interface.
Subnetwork Mask - The subnetwork mask for the IP address. Refer to Table 4-1 if you need help determining the subnetwork mask.
Broadcast Address - The interface's broadcast address. In most cases, the broadcast address consists of the interface's network/subnetwork number, and all 1s. The VINES interface does not use the broadcast address.
Enabled - Choose Y if you want the IP interface to be enabled immediately when added to the TCP/IP configuration. If you do not want the interface to communicate right away, choose N. The VINES interface is always enabled.
A sample IP interfaces worksheet is shown in Figure 4-5. This worksheet is for Server 1 in Figure 4-2.
Filling Out the ARP Entries Worksheet
Appendix A provides a blank ARP Entries Worksheet. On the ARP Entries Worksheet, you record the IP address/LAN address pairs you assign to hosts on LANs that do not support the automatic creation of ARP entries. Make a copy of the ARP Entries Worksheet for each interface for which you are creating ARP entries. You will need this information to manage IP interfaces, as explained in Chapter 8.
Record the following information:
Server Name - The server to which the information on the worksheet applies.
Interface Name - The name that describes the interface for which you are creating ARP entries.
IP Address (Interface) - The IP address of the interface.
Slot - This is the physical slot number in which the LAN card is installed.
ARP Enabled? - Choose Y (the default) if you want to enable ARP. Choose N if you want ARP to be disabled.
Direct Mapping Enabled (ProNET-10 only)? - Choose Y or N to indicate whether you want to enable direct mapping for ProNET-10 LANs. If direct mapping is enabled, ARP is disabled automatically.
IP Address (host) - The IP address of the host on the LAN.
LAN Address - The host's LAN address.
Published? - Choose Y if the entry is published. Choose N if the entry is not published.
When you publish an entry, you initiate proxy ARP. Proxy ARP is a technique in which a gateway answers ARP requests intended for another machine by supplying its own physical address. By using proxy ARP, the gateway then accepts responsibility for routing packets from the source machine to the machine it is imitating. This allows the server to route IP packets from a host on one segment to a host on another.
For example, if your server has an Ethernet segment and a ProNET-10 segment, hosts connected on one segment may need your server to give them the LAN addresses of hosts on the other. In that case, ARP entries should be published.
A sample ARP Entries Worksheet is shown in Figure 4-6.