How Modern Systems Depend on Things You Don’t Control

Ethan Cole
Ethan Cole I’m Ethan Cole, a digital journalist based in New York. I write about how technology shapes culture and everyday life — from AI and machine learning to cloud services, cybersecurity, hardware, mobile apps, software, and Web3. I’ve been working in tech media for over 7 years, covering everything from big industry news to indie app launches. I enjoy making complex topics easy to understand and showing how new tools actually matter in the real world. Outside of work, I’m a big fan of gaming, coffee, and sci-fi books. You’ll often find me testing a new mobile app, playing the latest indie game, or exploring AI tools for creativity.
3 min read 63 views
How Modern Systems Depend on Things You Don’t Control

Modern systems don’t run on what you build.

They run on what you depend on.

And most of those dependencies are outside your control.

Your System Is Not Your System

What you call “your system” is just a layer.

Underneath it:

  • cloud infrastructure
  • third-party APIs
  • identity providers
  • networking layers

And each of them has its own behavior, failures, and constraints.

This is the same reality behind invisible systems.

You don’t see them.

But they define everything.

Control Ends at the Boundary

You control your code.

You don’t control:

  • the network
  • the cloud
  • upstream services
  • shared infrastructure

This is exactly the limitation described in control as illusion.

Control doesn’t disappear.

It stops at the boundary.

Dependencies Define Behavior

Your system doesn’t behave independently.

It behaves as a function of its dependencies.

If a payment API slows down → your checkout slows down
If DNS fails → your system disappears
If authentication breaks → your system is unusable

This is why systems fail in chains — the same pattern behind global outages.

The More You Scale, the Less You Control

Small systems can be self-contained.

Large systems cannot.

Scaling introduces:

  • more dependencies
  • more integration points
  • more external control

Which means:

More scale → less control

Control Lives Somewhere Else

Modern infrastructure is not decentralized in control.

It is concentrated.

A few providers define:

  • infrastructure behavior
  • availability patterns
  • failure characteristics

This is the same dynamic behind control planes.

You don’t control the system.

You operate within someone else’s control layer.

Dependencies Fail Differently

The biggest problem is not that dependencies fail.

It’s that they fail differently.

  • they have different SLAs
  • different failure modes
  • different recovery times

Which means:

Your system inherits unpredictability.

You Can’t Test What You Don’t Own

Testing assumes control.

But dependencies break that assumption.

You can’t fully simulate:

  • cloud outages
  • third-party API behavior
  • network-level issues

This is why chaos engineering exists — to expose real behavior under stress, as shown in chaos engineering.

Failure Comes From the Outside

Most teams look inward for problems.

But many failures originate externally.

From:

  • upstream dependencies
  • shared infrastructure
  • invisible layers

Exactly the layers described in systems nobody understands.

Your Reliability Is Outsourced

This is the uncomfortable truth.

Your system’s reliability depends on systems you don’t own.

Which means:

  • your uptime is shared
  • your failures are external
  • your risk is distributed

Isolation Is the Only Real Control

You can’t eliminate dependencies.

But you can isolate them.

  • timeouts
  • retries (carefully)
  • fallbacks
  • circuit breakers

Not to control them.

But to limit their impact.

The Illusion of Independence

Modern systems look independent.

They aren’t.

They are tightly coupled through dependencies.

And that coupling is often invisible.

Until something breaks.

The Real System Is the Network of Dependencies

What actually exists is not a single system.

It’s a network of systems.

Each one:

  • partially controlled
  • partially understood
  • partially reliable

And your system is just one node.

The Final Principle

You don’t build systems that control everything.

You build systems that survive what they don’t control.

Share this article: