Web Sites, Hostnames and IP Addresses, Oh my.

In order to understand what all of these growth numbers mean, you need to know what all the words mean. If you already know the difference between a web site, a hostname, etc., skip to the end of this page, where I give my specific definitions for some of the less well defined terms, like a 'web site'. If you just want quick glossary form definitions, skip down to the glossary.

So, to begin with the Internet is essentially a whole bunch of computers connected together by wires which they can talk over. Now, the wires that we're interested in connect to something inside the computer called a network interface. A network interface can be an ethernet card, a modem or a number of other things.

Each network interface (usually) has a single IP address associated with it. An IP address is a set of four numbers (each between 0 and 255, with some restrictions) seperated by periods that uniquely identify that address on the Internet.

Now, to make our lives complicated, a single computer may have multiple network interfaces and therefore multiple IP addresses, but, in general, a machine has a single IP address. An IP address always is a specific interface on a specific machine (at any given moment). For the moment, it is convenient to pretend that a single IP address corresponds to a single computer in cases we look at.

Now, all direct communication between computers on the net uses these IP addresses. Human beings, however, prefer slightly more mnemonic names like 'www.mit.edu'. This is called a hostname. Some people will distinguish that 'www' is the hostname and 'mit.edu' is the domain name, but really 'www.mit.edu' is the Fully Qualified Domain Name (FQDN). Essentially though, something like 'www.mit.edu' is a hostname.

When you use a hostname, it gets looked up on a machine called a name server. Bascially, you give it a hostname and it tells you an IP address. Usually, a given hostname will always give you the same IP address back. This process of getting an IP address for a hostname is called name resolution. Often, however, multiple hostnames will end up giving back the same IP address. For example, the hostnames 'www.mit.edu' and 'anxiety-closet.mit.edu' will resolve to the same IP address, and hence, the same machine.

Additionally, it is common to refer to 'hosts' rather than hostnames. These are distinctly different. Usually when one says 'host', one means individual computers or IP addresses, as there are often multiple hostnames per host.

Now, how does a the web connect into all of this? Well, a web page is defined by a URL (which may have some inlined content such as images or Java applets, which have their own URLs). Most URLs (such as http, ftp and gopher) contain a hostname.

So, one of the hardest terms to define and keep everyone happy is the term 'web site'. Generally, a web site is a set of pages that are meaningfully linked together. Some people would prefer to define any URL as a 'web site', though 'web page' is easily a better term for this. My definition of a web site is all documents with urls beginning with a unique hostname. That is, http://www.mit.edu/people/mkgray/ and http://www.mit.edu/madlibs are part of the same site, but a document http://web.mit.edu/ is a seperate site.

Glossary

The terms listed here are fairly accepted definitions. Fuzzier, less well defined terms are defined below in the section on my terminology.
Computer
A box hooked up to the net. Usually has one or more processors.
Ethernet Card
A piece of hardware that is usually in a computer which knows how to talk the ethernet protocol on a wire. This is a kind of network interface.
Host
A fuzzy term, usually meant as a synonym for either 'computer' or 'IP address', which often doesn't matter as there is typically one IP address per computer.
Hostname
A name which can be resolved (using a name server) into an IP Address.
IP Address
A set of four numbers (each between 0 and 255, with some restrictions), seperated by periods that uniquely identifies an address on a network.
Modem
A piece of hardware that allows the computer to talk over a phone line. A modem is often used as a network interface.
Name server
A machine on the network that allows you to resolve hostnames into IP addresses.
Network Interface
A thing inside a computer that knows how to send information over a wire (or fiber, or whatever). Usually, one per computer.

My Terminology

Web Site
All documents with urls beginning with a unique hostname. That is, http://www.mit.edu/people/mkgray/ and http://www.mit.edu/madlibs are part of the same site, but a document http://web.mit.edu/ is a seperate site.

This presentations is copyright 1996, Matthew Gray
Reproduction rights are specified at the top level page.
[Web Growth Summary] [Internet Growth Summary] [Terminology] [Credits] [Contact]