background image
548 Chapter 8: WAN Protocols and Design
each next-hop Layer 3 address for each Layer 3 protocol being routed. Even with a network this
small, the configuration process can be laborious. A better solution is a dynamic protocol called
Inverse ARP.
Inverse ARP acquires mapping information in a manner opposite of IP ARP, but the information
that maps the Layer 3 and Layer 2 addresses is still found. Inverse ARP is basic; after the VC
is up, each DTE announces its network layer address to the DTE on the other end of the VC.
(Inverse ARP is enabled by default at IOS 11.2 and beyond, unless point-to-point subinterfaces
are used.) Table 8-21 summarizes what would occur in the network in Figure 8-16.
Although Table 8-21 lists a seemingly large number of messages, no static frame-relay map
commands are needed. All the necessary mapping information is learned dynamically.
Inverse ARP is enabled by default if no subinterfaces are in use. Inverse ARP is also on by
default for multipoint subinterfaces. However, Inverse ARP is not enabled on point-to-point
subinterfaces because point-to-point subinterfaces behave like a true point-to-point line in
regard to routing; if the packet needs to be sent out a point-to-point subinterface, then there is
only one possible next router and only one DLCI in use. So, Inverse ARP is not needed with
point-to-point subinterfaces.
Table 8-21
Inverse ARP Messages for Figure 8-16
Sending
Router
DLCI in
Header of
I-ARP Frame
When Sent
Receiving
Router
DLCI in
Header of
I-ARP Frame
When
Received
Information in
Inverse ARP
A
41
B
40
I am 10.1.2.1.
A
41
B
40
I am 2.0200.aaaa.aaaa.
A
42
C
40
I am 10.1.2.1.
A
42
C
40
I am 2.0200.aaaa.aaaa.
B
40
A
41
I am 10.1.2.2.
B
40
A
41
I am 2.0200.bbbb.bbbb.
B
42
C
41
I am 10.1.2.2.
B
42
C
41
I am 2.0200.bbbb.bbbb.
C
40
A
42
I am 10.1.2.3.
C
40
A
42
I am 2.0200.cccc.cccc.
C
41
B
42
I am 10.1.2.3.
C
41
B
42
I am 2.0200.cccc.cccc.
ch08.fm Page 548 Monday, March 20, 2000 5:17 PM