Principles Digital Communication System & Computer Networks (Charles River Media Computer Engineering)

21.4 REVERSE ADDRESS RESOLUTION PROTOCOL

Machines that do not have secondary storage (diskless machines) do not know their IP addresses. At the time of bootup, the machine has to get its own IP address. However, the physical address of the machine is known, because it is a part of the hardware, the network card. As you can see, this problem is the opposite of the earlier problem, and hence reverse address resolution protocol (RARP) was developed.

In reverse address resolution protocol, a server stores the IP addresses and the corresponding network addresses of the nodes. This protocol is used when the nodes are diskless machines and do not know their IP addresses.

In RARP, the machine that wants to find out its IP address broadcasts a packet with its own network address. A RARP server, which stores the IP addresses corresponding to all the network addresses (in secondary storage), receives the packet and then sends a reply to the machine with the information on its IP address. A requirement for RARP to succeed is that a designated RARP server should be present that contains the lookup table. Sometimes, for providing a reliable service, primary and secondary RARP servers will be installed in LANs. If the primary RARP server fails, the secondary RARP server will take over and provide the IP addresses to the nodes on the LAN.