Inverse Address Resolution Protocol (Inverse ARP or InARP), is a protocol used for obtaining Network Layer addresses of other nodes from Data Link Layer addresses. For example, in Ethernet networks InARP would primarily be used to get IP addresses when MAC addresses are known.

It is primarily used in Frame Relay and ATM networks, in which Layer 2 addresses of virtual circuits are sometimes obtained from Layer 2 signaling, and the corresponding Layer 3 addresses must be available before these virtual circuits can be used. In Frame Relay networks InARP is used to get Data Link Connection Identifier (DLCI) mappings to Virtual Circuits and is enabled automatically by default.

InARP operates essentially the same as ARP with the exception that InARP does not broadcast requests but sends them by unicast directly to the destination.  This is because the hardware address of the destination host is already known. The source host that sends the request simply formats it by inserting its source hardware and protocol addresses and the known destination hardware address. It then zero fills the target protocol address field.  Finally, it will encapsulate the packet for the specific network and send it directly to the target station. The diagram below gives us an illustration of how Inverse ARP works

1. Host A wants to send a packet destined to a host with MAC address CC-CC-CC-CC-CC-CC but does not know the IP address of the device
2. Host A then sends a unicast ARP request to the specified destination host
3. Switch A looks at the destination MAC address in the packet and forwards the packet to Host C
4. Host C updates its ARP table and sends a reply back to Host A
5. Host A receives the reply and updates its ARP table and then sends the packet unicast to Host C with its destination IP address

Upon receiving an InARP request, a station may put the source protocol address/hardware address mapping into its ARP cache as it would any ARP request.  Unlike other ARP requests, however, the destination host may assume that any InARP request it receives is destined for it. For every InARP request, the destination host may format a proper reply using the source addresses from the request as the target addresses of the reply.  If the host is unable or unwilling to reply, it ignores the request.

When the source station receives the InARP reply, it may complete the ARP table entry and use the provided address information. The information learned via InARP or ARP may be aged or invalidated under certain circumstances.

InARP (Layer 2 known, Layer 3 unknown) is the inverse of ARP (Layer 3 known, Layer 2 unknown) as they both do just the opposite. In addition, InARP is actually implemented as a protocol extension to ARP. InARP uses the same packet format as ARP; but has different operation codes.

The Address Resolution Protocol (ARP) is a computer networking protocol to find out the MAC address (Physical address) of a device when the IP address (Logical address) is known. This is predominantly used in Local Area Network (LAN) environments as well as routing data traffic based on IP addresses when the next hop router must be known. The diagram below illustrates how ARP functions

1. Host A wants to send a packet destined to 192.168.1.9
but does not know the MAC address of the device
2. Host A then sends an ARP request to all hosts (BROADCAST)
which is forwarded to the Switch A connected to all hosts
3. Switch A looks at the destination MAC address in the packet and broadcasts the packet to all hosts on the LAN except Host A
4. Host B and Host D receive the packet and just update their ARP table with the IP address of Host A and its MAC address
5. Host C updates its ARP table and notices its IP address in the destination IP address field and sends an ARP reply which is unicast back to the source MAC address of Host A
6. Host A receives the ARP reply and updates its ARP table and then sends the required packet unicast to Host C

Originally part of RFC 826 defined in 1982. Although ARP has been used in many types of networks such as IP, DECNET, Token Ring, FDDI and other Ethernet Technologies, today it is predominantly used to translate IP addresses to Ethernet MAC addresses because of the prevalence of IPV4 and Ethernet in general. It is generally reference under layer 2 (data link) and layer 3 (network) of the OSI networking model.

ARP functions as a low level request and response protocol that is sent across the media access level of the underlying network. In case of Ethernet systems, ARP disguises itself in the payload of the Ethernet packet (see below)

ARP announcements 

Sometimes ARP may be used for announcement purposes. For example, if the IP address or MAC address changes due to a changed Network Card, DHCP lease expiring, etc. The host signals that to the rest of the network by sending out a Gratuitous ARP message broadcast. This type of an announcement is just for the information of the other hosts so that they can update their ARP table and is not send with the intention of receiving a reply. It is also used during MAC address change in Network Interface card (NIC) Teaming, High availability clusters.

ARP probe

A host broadcasts ARP probe packets all around the network before it begins using an IPV4 address (manually assigned, DHCP assigned). This is done to test if the address is already in use and is accomplished by sending out an ARP request constructed with a 0.0.0.0 (all-zero) source IP address.