You bind an IP Manager, and thus an IP management domain, to a virtual IP interface on the host system.

Consider again the two domains, Domain A and Domain B, that are defined in the example in “IP management domains” on page 165. To discover and poll these networks, you configure a separate virtual IP interface for each domain, say Virtual A for Domain A, and Virtual B for Domain B. Then, for the IP Manager that is associated with Domain A, you bind the IP Manager to Virtual A; for the IP Manager that is associated with Domain B, you bind the IP Manager to Virtual B.

In this case, IP Manager_A will be bound to Virtual A, and IP Manager_B will be bound to Virtual B.

The domain-to-interface binding has the following effect on the IP packets that are sent and received by the IP Managers:

  • All ICMP and SNMP requests from IP Manager_A specify Virtual A as the source address. All ICMP and SNMP requests from IP Manager_B specify Virtual B as the source address. Because of this distinction, two packets that are destined for identically addressed devices in Domain A and Domain B are differentiated by the packet source address, as shown in Path of packets to the devices in Domain A and Domain B.

    1. IP

    2. Manager_B

    3. IP

    4. Manager_A

    5. Packet from IP Manager_A (Domain A)

    Figure 1. Path of packets to the devices in Domain A and Domain B
  • An SNMP response or trap whose destination address is Virtual A is interpreted in the context of Domain A. An SNMP response or trap whose destination is Virtual B is interpreted in the context of Domain B. Incoming traps from identically addressed devices in Domain A and Domain B are distinguishable by the packet destination address, as shown in Path of packets from the devices in Domain A and Domain B.

    1. IP

    2. Manager_B

    3. IP

    4. Manager_A

    5. Packet to IP Manager_A (Domain A)

    Figure 2. Path of packets from the devices in Domain A and Domain B

    Proper trap processing support requires that devices in Customer A’s network are configured to send traps to Virtual A. Devices in Customer B’s network must be configured to send traps to Virtual B. Because responses are sent to the source address of the query that initiated them, the responses are automatically sent to the right destination.