Inverse ARP dynamically creates a mapping between the Layer 3 address and the Layer 2 address (DLCI).  The end result of Inverse ARP is the same as IP ARP on a LAN: The router builds a mapping between a neighboring Layer 3 address and a corresponding Layer 2 address.  The process used by Inverse ARP differs for ARP on a LAN.  After the VC is up, each router announces its network layer address by sending an Inverse ARP message over that VC.

Inverse ARP announces its Layer 3 addresses as soon as the LMI signals that the PVCs are up.  Inverse ARP starts by learning the DLCI data link layer address (via LMI messages), and then it announces its own Layer 3 addresses that use that VC.  Inverse ARP is enabled by default.

The Inverse ARP messages contains the sender’s Layer 3 address, and the Frame Relay header has a DLCI in it.  These two values are placed in the Inverse ARP cache on the receiving router.