Information Technology Reference
In-Depth Information
In some cases, an Ethernet address is known, but the IP address needs to be
determined. This process is known as the
Reverse Address Resolution Protocol
(RARP)
. The details of this process are specified in
RFC 903—A Reverse
Address Resolution Protocol
.
A host that needs to resolve an Ethernet address broadcasts an ARP request
that contains the IP address as data. The message basically asks “What is the
Ethernet address of the host with this IP address?” Note that the ARP
request is an Ethernet broadcast. That is, the destination Ethernet address of
the Ethernet frame is all ones (
FF:FF:FF:FF:FF:FF
). The host that matches
the specified IP address responds with an ARP reply that contains its
Ethernet address. This address is part of the Data field of the Ethernet
frame.
The same process is used to resolve an IP address when the Ethernet address
is known. An RARP request is broadcast (destination Ethernet address of
FF:FF:FF:FF:FF:FF
) that contains an Ethernet address in the Data field. The
appropriate host responds with a RARP reply that contains the requested IP
address.
Sometimes it is necessary for a system to respond to an ARP or RARP
request on behalf of another system. This process is referred to as
proxy ARP
and some protocol stacks (typically routers) are designed with this capability.
Classes of IP Networks
In the TCP/IP network model, the Internet Protocol (IP) at the Internet
Layer is used to segregate networks into classes. The three most commonly
used IP network classes are A, B, and C. Table 10.1 lists the characteristics of
these different IP network classes.
Table 10.1
IP Address Classes
Class
First
Default
Maximum
Octet
Subnet Mask
Theoretical Hosts
A
1-126
255.0.0.0
16,777,216
B
128-191
255.255.0.0
65,536
C
192-223
255.255.255.0
256
The subnet mask is used to separate the network portion of an IP address
from the host portion.
