URL is a web address - like http://www.wordmagazine.co.uk, in easy human-readable form.

I/P address is the address on a computer network of a particular device, or web page, which can be accessed in a browser (sometimes) as http://192.168.1.1

The same "thing" can be represented by a URL (more human readable, more memorable) and by an IP Address.

DNS stands for Domain Name System - it's the store of references that match URLs and IP Addresses.

The wiki page is techy but helpful :

http://en.wikipedia.org/wiki/Domain_Name_System

Gracias el hombre malo

Has a unique number (its IP Address). Unhelpfully, it's ultimately in binary, so to make it a bit more manageable, it's translated into octets, or groups of eight bits. So
11111111111111111111111111111111 = 255.255.255.255 (I can explain. It'll bore you)

Octets look a bit more pleasing to the eye, but aren't super easy to remember, so DNS acts as a phone book, translating these octets into meaningful names, purely for our benefit. Our computer would quite happily work with the numbers, but that's an issue for us fleshbags.

Each URL you type into the browser is sent to a DNS server somewhere which uses a tree-like structure to do the translation, only it looks at in in reverse. So www.dell.com/support is treated thusly. Ignore everything after '/' - that's a page on the site and not our issue. It's a .com, so ignore all the .uk .fr addresses etc and look in the list under .com. Ok here's Dell.com, I've got a number for that. lets see if I have one for www.dell.com. No? I'll send a request to another server and see if they know. And of it goes, until www.dell.com is translated into an IP address (yes, WWW is the name of a server - typically a logical 'alias' for it's real name), which is then passed back to your computer.

Once your computer has an IP address, it will attempt to find that address, by seeing if it exists on it's network (Say your local machine is 192.168.1.1). It doesn't - in fact it can't as 192.168.x.x addresses are for private use only and not allowed on the internet. So the information is passed to your "default gateway" - typically your router in a home LAN. This will send the request up to your ISP, effectively saying, "Hey guys, do you know where 12.1.3.12 is?" This request hits a router which will either have a known route to the server, or will take a guess at an appropriate route and pass the request to another router elsewhere for it to have a go. If you issue the command

tracert www.bbc.co.uk

You can see the path its taking. It's like a computer version of "The 7 degrees of separation from Kevin Bacon" - in a surprisingly short time, one router will have a definitive answer. Once the path is known, the information will be passed back to the requesting PC and all routers will update their tables. Now you have a path between computer and server, so your computer can say "Hi", the server can choose to say "Hello, here's what I have to offer" or not. At this point, communication is with a "web server" and depending on what you are trying to do, an appropriate channel of communication can be established. Your computer and server might encrypt their conversation (HTTPS), by mutually agreeing a key based on a publically-available certificate - which is authenticated by a trusted third party. God help us if Verisign is hacked...Anyway, remember '/support'? This is what you were after - can the webserver translate this to a page or a directory with a default response? It can! Joy! Send the information back, in the form of HTML to have it interpreted by the browser on your PC which will render it into pretty pictures.

The clever bit is that if one path on the route goes down, the routers will transparently figure out a new way. There's a reason the military came up with all this - blow up New York? Fine - all traffic will be routed via Buffalo.

Of course, all this takes place at light speed, and is, frankly fucking miraculous.