Skip to main content

How the Internet Works

A. The Internet Is a Collaborative System​

The internet is one of mankind's greatest achievements, but it was not created by a single person, company, or organization.

It evolved through the cooperation of many people, organizations, and technologies.

The internet is built on cooperation and rules.

These shared rules allow completely different systems and technologies to communicate with each other.

B. Why "How Does the Internet Work?" Is a Good Interview Question​

"How does the internet work?" is a broad, open-ended interview question.

It allows candidates to explain any part of the technology stack they understand well, such as:

  • Operating Systems
  • Web Servers
  • Application Servers
  • Networking

The answer can demonstrate not only technical knowledge, but also curiosity and understanding of the broader engineering ecosystem.

C. Cooperation Through Standards and Protocols​

Different systems can communicate because they agree on common standards and protocols.

For example, Ubuntu and macOS can both use SSH because they implement the same protocol.

The basic idea is:

Agree on a common standard → implement it → different systems can communicate.

The internet is essentially built from many of these shared decisions and standards.

D. One Possible Internet Request Path​

A request does not follow one fixed sequence of devices or providers. From a user's point of view, the simplest home setup looks like this:

The box people call a Wi-Fi router often performs several jobs:

  • Wi-Fi access point: Lets wireless devices join the home network.
  • Router: Moves traffic between the home network and the internet provider.
  • NAT: In a typical IPv4 network, lets devices with private IP addresses share one public IP address. It remembers which connection belongs to which device so replies return to the correct place.
  • Firewall: Blocks network traffic that is not allowed.

There may also be another box between the Wi-Fi router and the internet provider:

  • Fibre internet usually uses an ONT, which translates the signal from the fibre cable into a connection the router can use.
  • Cable or DSL internet usually uses a modem for the same general purpose.

The internet provider may combine the Wi-Fi router and ONT or modem into one device, so users do not always see them as separate boxes.

Beyond the home, traffic may cross several transit or peering networks before reaching the destination. The destination may use a load balancer and many servers, but those components are not required for every website.

E. Transit, Peering, and the Internet Backbone​

Your ISP (Internet Service Provider) exchanges traffic with other networks through transit and peering relationships. Some routes use large global backbone providers, but there is no single backbone or required Tier 1 path for every request.

These backbone networks:

  • Handle extremely large amounts of traffic
  • Connect different networks together
  • Help move traffic across regions and countries

From there, traffic may reach a Data Center operated by providers such as AWS or DigitalOcean.

F. Data Centers, Load Balancers, and Servers​

Inside a Data Center, a large application may run on multiple servers rather than a single server.

A Load Balancer distributes incoming traffic across those servers so that one server does not receive all the requests.

This is the final part of the simplified journey from the user's computer to the application server.

Key Idea​

The internet is a network of networks built on cooperation, shared standards, and interconnected infrastructure.

Even loading a single web page involves many different technologies working together behind the scenes.

Network Tools Exercise​

G. Ping — Check Whether a Host Is Reachable​

ping is a simple network tool used to check whether a host is reachable over the network.

It uses ICMP (Internet Control Message Protocol), a network-layer protocol used by devices to send diagnostic and error messages. ping sends an ICMP Echo Request to the destination and waits for an ICMP Echo Reply.

ICMP does not use a TCP or UDP port. A successful reply shows that the destination responds to ICMP, but it does not prove that a website, API, or another application service is working correctly.

For example: ping google.com

It is useful when you want to quickly determine whether a network problem exists.

For example, if a website is loading slowly, you might want to know:

  • Is my network down?
  • Is the remote host reachable?
  • Is the problem with my browser or application?

ping is lightweight and can quickly tell you whether a host responds.

ping is a quick way to check whether a host is reachable.

Some servers may intentionally disable responses to ping, so a failed ping does not always mean the server itself is down.

This means that the server or firewall intentionally does not respond to ping requests. Common reasons include:

  • Security: To reduce network scanning and discovery.
  • Firewall rules: The firewall blocks ICMP Echo Requests.
  • DDoS protection: To prevent large amounts of ping traffic from overwhelming the server.
  • Network policy: Administrators may simply configure the network not to respond to ping requests.

H. Traceroute — See the Path Across the Internet​

traceroute shows the path that traffic takes from your computer to a destination host.

For example: traceroute google.com

Instead of showing only whether the destination is reachable, it displays each intermediate hop.

A simplified route might look like:

Each hop represents another network device or network that the traffic passes through.

This makes the complexity of the internet more visible.

Traceroute and Latency​

traceroute also shows how long each hop takes to respond.

This can help you understand:

  • How far away a network is
  • How quickly it responds
  • Where network latency increases
  • Where a connection may be failing

For example, a request to google.com might travel through many different networks before reaching Google's data center.

The instructor's example required around 20 hops to reach Google, but hop count should not be treated as a direct measure of geographical distance or connection quality. Routes can change, some routers do not reply, and load balancing may produce different results between runs. Many traceroute implementations also stop after a default maximum number of hops unless configured otherwise.

The internet does not connect your computer directly to a remote server. Traffic passes through many intermediate networks.

Using Traceroute for Troubleshooting​

traceroute can help identify where a network problem occurs.

For example:

If latency suddenly increases or requests begin timing out at a particular hop, the problem may be with:

  • Your local network
  • Your ISP
  • An intermediate network
  • The destination's data center

This makes traceroute useful for determining which part of the network path is causing a problem.

I. Netstat — Inspect Network Activity​

netstat provides much more detailed information about network activity on your computer.

It can show information such as:

  • Open ports
  • Connected IP addresses
  • Active network connections
  • Processes communicating through the network
  • Local network activity

Compared with ping and traceroute, its output can be much more verbose.

ping checks reachability, traceroute shows the route, and netstat shows network activity on your machine.

Key Idea​

Network tools help expose the normally hidden infrastructure behind internet communication.

These tools are especially useful for Network Administrators and DevOps Engineers, but they also help developers understand and troubleshoot how applications communicate over the internet.