How does the Internet Work?


eye icon This icon indicates that there is a screen capture or animation describing the concept.

speaker icon This icon indicates that there is a sound clip describing the tutorial.


Tutorial Overview

Topics covered in this tutorial (click link to jump to topic):

1. Introduction 4. Internet Hardware
2. Transport 5. Internet Maps
3. Domain Names    


To understand the Internet you need to understand three main areas:

  1. the fundamental transport protocol i.e. how data is moved around the Internet
  2. domain names and domain name resolution i.e. how things are named or referenced on the Internet
  3. servers, routers and clients i.e. the hardware of the Internet

Transport Protocol - TCP/IP

The fundamental means of moving data around the Internet is controlled by a protocol called TCP/IP, or transmission control protocol/internet protocol. TCP/IP is also used on private networks, like your office LAN or home network. As the name suggests, TCP/IP is the combination of TCP and IP protocols working together.

Under TCP/IP a file is broken into smaller parts called "packets" by the file server. Each packet is assigned an IP (Internet protocol) address of the computer it has to travel to. As the packet moves through the global network it is "switched" by a number of servers and routers along the way toward its destination... the requesting computer or "client" computer.

The IP address tells those servers which way to switch the packet. Each time the packet is switched a "wrapper" is added to the packet – this way we can tell how many computers and which computer handled the file while it was in transit. You can do this with the command "tracert" at the command line of your computer (e.g. tracert Tracert stands for "trace route".

In Australia, a file coming from the United States can be switched up to 15-20 times, that is fifteen computers passing on the data like a hot potato toward the destination computer.

The packets do not necessarily travel together on the Internet. Packets from the same file may travel via different paths through different servers, but toward the same destination.

Packaging technology allows us to use limited bandwidth most efficiently. It means parts of a file can be shared across a number of phone lines instead of having to find one phone line to put a large file into. It is also hard to break the network, as the data will be routed around the trouble spot. In this respect TCP/IP can be likened to a group of 10 hitchhikers (packets) who can not get a lift all together, but easily get lifts if they break up, going by different cars and maybe by different roads… but agree to meet up at a particular point in the future.

Further information on TCP/IP.

This tutorial is brought to you courtesy of Dynamic Web Solutions. Please contact Peter or Michael for a quote on a dynamic, mobile friendly web site with content management, Google analytics, hosting and web marketing support: (+61) 2 66993800.

Return to top

Domain Names

A domain name is a business' or an organisation's name on the Internet. It is the part of the email address after the "@" sign or part of the web address after the "www.". e.g. is the domain name of this web site. It is also an abbreviation of "Dynamic Web Solutions Pty Ltd"... a business name.

In Australia domain names are registered by licensed registrars of the Australian Domain Name Authority, the AUDA. Large registrars include: Enetica, Netregistry and MelbourneIT. These companies are authorised by the AUDA to license domain names. They also operate reseller programs to allow smaller companies to register domain names on behalf of their customers. In Australia you need an ABN, ACN or RBN to register a domain name.

Once a domain name is registered, it is "delegated" to a server. Delegation means pointing the domain name at the server. That server then answers file transfer protocol requests, receives email and serves web pages for that domain.

To delegate a domain name you need to know the name servers of the company that will maintain your web site and email. This company is referred to as the "host".

The host maintains a DNS server, or "domain name server". This is generally a software server running continuously answering calls for your domain (and others) that are delegated to that server and directing the requests and inbound data to different facilities of the host's server. You can think of a DNS server as a table that helps the server route traffic to different IP addresses.

So why do domain names have to have an IP anyway? An IP address looks like 201.517.635.124... I think you will agreed, that is not very memorable. That IP address has a letter equivalent... one that humans can easily recall. Therefore the domain name system is designed to say:

1. which server will answer requests for which domain name i.e. which server is "authoritative" for which domain name

2. which IP addresses "resolve" to which domain names".

Return to top

The Hardware of the Internet

Broadly speaking the Internet is comprised of servers, routers and clients.

1. Servers are powerful computers that run 24 hours a day, 7 days a week, without stopping and hopefully without failing. Servers are specially constructed with redundant hard drives (this is called a RAID array), extra cooling and are generally "rack optimised", which means they are very flat and deep and are intended to slip into a rack of 32 similar machines.

Most servers are housed in "data centres". The datacentre offers physical security, redundant power supply (with back up diesel generators), redundant air-conditioning, special fire retarding systems, anti static mats... and the list goes on. In other words, short of an earthquake or nuclear strike, datacentres are designed to handle many contingencies. The objective of a datacentre is to keep those servers cool and running continuously. During hurricane Katrina only one New Orleans datacentre was able to do this.

2. Routers pass data onwards to its destination IP. They aren't all that smart, but they do their job quickly and efficiently.

3. Client machines, that is the end user, the person requesting a web page or downloading email. The client machine uses a browser and an email client to perform these functions. The server waits for a request from a client, then grabs the requested data and serves it back.

4. Cables and conduits. The physical connections between servers, routers and clients vary enormously. Between you and your telephone exchange, it is likely you are using twisted pair copper wire. Between your exchange and your ISPs network, it is likely you are using fibre optic cables. Between ISP and ISP it is likely you are using T1 carriers, the main trunk routes of the Internet.

Today, there are good arguments to include VoIP phones, some mobile phones and Blueberry-styled PDAs as "Internet hardware" as all of these devices are connecting to the Internet for various functions.

How Stuff Works FAQs on internet infrastructure.

Internet Maps

The Internet is vast and to try to better understand it people have depicted it with maps or radial graphs. Links below are to these sites. I add them here for your interest:

Was this tutorial helpful? If so, please support it with a donation >>

Return to top

[Web Site Design] [Search Engine Submission] [Web Site Hosting] [About Dynamic Web Solutions] [Internet Tutorials] [Sitemap] [Links] [IRC] [Homepage]