Difference between revisions of "Address Resolution Protocol"
(Still needs an outline of the basic protocol mechanism, but this is a good start) |
m (+cat) |
||
| Line 11: | Line 11: | ||
* [https://datatracker.ietf.org/doc/html/rfc826 An Ethernet Address Resolution Protocol] - protocol spec | * [https://datatracker.ietf.org/doc/html/rfc826 An Ethernet Address Resolution Protocol] - protocol spec | ||
| + | [[Category: Network Protocols]] | ||
[[Category: Networking]] | [[Category: Networking]] | ||
Revision as of 13:28, 6 November 2021
The Address Resolution Protocol (usually given as its acronym, ARP) is a networking protocol which is used to translate from the addresses used by the internetworking layer of a particular protocol family (e.g. Chaosnet, or the Internet Protocol of TCP/IP) to those used by a particular physical network (e.g. Ethernet), in cases where a non-trivial mapping is required between the addresses used by the internet protocol, and the addresses used by the physical network.
Background
Originally, all physical networks had addresses which were relatively short (e.g. 24 bits in the ARPANET; 8 bits in the Chaosnet LAN) and those were carried directly in the low order bits of the addresses of the internetworking layer of the particular protocol family.
With the advent of the 10 Mbit/second Ethernet, which had 48-bit physical addresses (so that all Ethernet network interfaces could be assigned a guaranteed-unique physical address at manufacturing time), this was no longer true. Use of static, manually-configured tables to hold the required mappings was clearly infeasible, so ARP was designed.
External links
- An Ethernet Address Resolution Protocol - protocol spec
