Networks and the Internet

A network is nodes sending data to each other. Build one, break it, zoom out from a home network to the whole Internet, follow a message hop by hop, and see how fast it all grew.

This page supports British University Vietnam students in the module Introduction to Computing Science (2024–2027 and onwards). It accompanies the Week 8 “Networks” lecture. The next lesson, Network protocols, shows how the data actually travels.

1 · What is a network?

Nodes and arcs: a network is a graph

Networks consist of nodes (N) and arcs (A). Writing a network as a graph lets a computer analyse it: if we start at N3, can we get to N2? The same maths describes rail networks, flights and social networks. Redundancy matters too: failures should not bring down the whole system.

Can I get there?

From to

Show single points of failure

Mathematical representation

2 · How big can a network be?

LAN → MAN → WAN → Internet

Lots of LANs form a MAN, lots of MANs form a WAN, and the Internet is a collection of networks. Move the slider to zoom out.

3 · How do we navigate it?

tracert: following a message hop by hop

The tracert (or traceroute) command shows the nodes between you and a site. It sends probes with a TTL (time to live) of 1, 2, 3, … Each router subtracts 1; the router where TTL reaches 0 discards the packet and reports back. The list of replies is the route. Addresses below are made-up examples.

What is happening

tracert output

4 · The biggest network

The Internet: how it grew, and who is on it

In the beginning

A US military communications network of the late 1960s and 70s, designed to be robust even against attack. ARPANET used packet switching rather than an operator connecting wires, with distributed control. Its first four nodes used different protocols; only when people moved jobs between universities did the machines really learn to talk.

The Internet is not “the web”

The World Wide Web is just one service that runs on the Internet. Email, file transfer, online games and video calls use the same network but different protocols.

Why so fast?

Accessibility: cheap devices, batteries, Wi-Fi to share one connection. Need: social media, business competition, taxes, gaming and media, shopping, education.

Hosts on the Internet (log scale)

Each step up is ten times more hosts. From 111 hosts in 1977 to a billion in 2014: growth of about 9,000,000×. Values from the lecture (2001 is quoted as 150–175 million).

The “world-wide” web? Who is connected

Internet users as a share of each region’s population (Sept 2020 estimates in the lecture slide). Access is far from equal.

Access inequality has more causes than infrastructure: cost, language, and rules such as national filtering (for example the “Great Firewall of China”) change what people can reach.

Things to try

Reachability. In the lecture network, ask whether N3 can reach N2. Break link a3 and ask again: is there still a route? How many routes are there?
Single points of failure. Tick Show single points of failure on the Star and then the Ring. Which is more fragile? Why do real networks add extra links?
Design a robust network. Start from Chain and add links until no node or link is critical. What is the fewest you need?
Peer-to-peer or client/server? In the LAN scale, switch the type and send a message. What happens if the server fails? What if one peer fails?
Zoom out. At which scale does a message first leave your own organisation’s network? Which scale is the Internet?
Trace it. Compare the near and far destinations. Why can some hops show “Request timed out”, yet the trace carries on?