The address on a normal letter typically has several lines, each of which supplies different levels of information about the exact location corresponding to the address. In a very similar way, addresses for Internet information packets have four fields that contain numbers and are separated by periods. For example,

The translation between the numbers used by the network, and the name more commonly used by people is done by a computer called a nameserver. The purpose of a nameserver is to look up addresses, so its function may be likened to directory assistance on the telephone system.

If I specify csep10.phys.utk.edu as an IP address to a network, the first thing that it does is call its "directory assistance" (that is, a nameserver) to get the number associated with that name. The network uses only the number, but by this method the humans and the computers are both kept happy: the computers in the network just have to deal with the numbers, which is what they do best, and the humans using the computers need only remember more symbolic names like mycomputer.myplace.mynet.mycountry, which is what humans do better.

Of course, the humans can also supply the IP address directly to the network as a number. Then, the nameserver isn't invoked. In fact, sometimes the nameservers, since they are machines, are down for some reason and one MUST supply the number to the network. However, that doesn't happen too often these days on modern and well maintained networks because there usually is more than one nameserver available, and if one is down another can often be found by the network when it needs a name translated.

Next

Back

Top

Home

Help