How the Internet actually works (and what is an ASN?)

The Internet is not a single system owned or run by any one organisation. Instead, it is made up of tens of thousands of separate networks, each operated independently.

These networks are known as Autonomous Systems, or ASes. Each of these networks is identified by an Autonomous System Number (ASN), which plays a critical role in how Internet traffic is routed worldwide.

An Autonomous System might be:

  • A broadband or telecoms provider
  • A mobile network operator
  • A cloud platform such as AWS, Google or Microsoft
  • A content delivery network (CDN)
  • A large enterprise, university, or government body

Each of these networks connects to others to exchange traffic, forming the global Internet.

What is an ASN?

Every Autonomous System is assigned a unique identifier called an Autonomous System Number (ASN).

An ASN tells the rest of the Internet:

  • Who a network is
  • Which IP addresses it is responsible for
  • How traffic should be routed to and from it

Routers across the world use ASNs to make routing decisions thousands of times per second. Without them, the Internet simply wouldn’t work.

While you may recognise the names of some networks such as Google, Vodafone or Microsoft, many of the largest and most important networks operate almost entirely behind the scenes and are rarely known outside the industry, such as Arelion, GTT, Cogent & NTT.

Who can run an Autonomous System?

In theory, anyone can run an Autonomous System.

In practice, the technical, financial and operational requirements are extremely high. Running an ASN means:

  • Operating your own routing infrastructure
  • Managing global routing tables
  • Maintaining secure, resilient connectivity
  • Negotiating interconnection agreements
  • Monitoring and defending your network 24×7

For this reason, most organisations, including the vast majority of businesses, rely on an established ISP (like Enlink Networks) that already operates an Autonomous System on their behalf.

The so-called "Internet backbone"

Some of the largest Autonomous Systems operate enormous fibre networks spanning countries, continents and oceans, including vast undersea cable systems.

Together, these networks form what people often refer to as the Internet backbone.

They are sometimes described as Tier 1 networks.

What does "Tier 1" actually mean?

Tier 1 networks are defined by one thing only:

They do not buy Internet connectivity (IP transit) from anyone else.

It’s that simple.

Instead, they connect to every other Tier 1 network through arrangements known as Settlement-Free Interconnection, or peering. Each agrees to exchange traffic without payment so that their customers can reach the entire Internet.

This definition is informal rather than official, but it has been widely used within the industry for many years.

Tier 2 networks

Tier 2 networks also peer directly with other networks — often including:

  • Other Tier 2s
  • Content delivery networks
  • Cloud providers
  • Internet exchanges

However, unlike Tier 1s, Tier 2 networks also buy IP transit from other carriers to ensure they can reach the entire Internet.

In practice, most Tier 2 networks purchase transit from multiple upstream providers, often blending Tier 1 and Tier 2 carriers to improve resilience, routing choice and performance.

Tier 3 networks

Tier 3 networks sit at the edge of the Internet.

They:

  • Do not peer directly with other networks
  • Rely entirely on purchased IP transit
  • May still use multiple providers for redundancy

Many access networks, smaller ISPs and enterprises fall into this category.

Why tier labels can be misleading

The tier system is a broad concept that was more useful in the past than it is today. In some ways, it now creates more confusion than clarity, as it implies a hierarchy that no longer reflects reality.

Being connected to a single Tier 1 network is not inherently safer or better.

Large networks don’t always agree. When two Tier 1 carriers fall out, whether over traffic imbalances, commercial disputes or policy decisions, they may de-peer, cutting off direct connectivity between their networks.

When this happens, customers of either carrier, including downstream ISPs and end users, can suddenly find large parts of the Internet unreachable unless alternative routing paths are already in place.

These events are rare, but they do happen, and when they do, the impact can be widespread.

Why interconnection matters more than "tier"

A well-connected Tier 2 network, with:

  • Extensive peering
  • Multiple diverse upstreams
  • Intelligent routing policies

can often outperform a poorly connected Tier 1 in terms of:

  • Latency
  • Resilience
  • Path diversity

What matters most is how a network is interconnected, not the label attached to it.

At Enlink, we design our network around interconnection..

We use a carefully selected blend of:

  • The most highly connected global carriers
  • Extensive local peering
  • Multiple independent routes

This gives our customers:

  • Lower and more consistent latency
  • Greater resilience to upstream issues
  • Protection from single-network failures or disputes

Our goal is to build a resilient, fast, business-only network that performs reliably in the real world.

The bottom line

The Internet is a constantly evolving web of independent networks, commercial relationships and technical agreements. While tier labels can be helpful shorthand, they don’t tell the full story. In today’s Internet, interconnection strategy matters far more than hierarchy.

For more information, visit our Internet Access page or contact us with your questions.