Chapter 1 · Foundations· Beginner· ~10 min

The CIA Triad

Three letters that quietly sit behind almost every security decision: Confidentiality, Integrity, Availability.

You will learn

  • What each pillar of the CIA triad means in plain language
  • A real-world example of each pillar failing
  • Why the three pillars often pull against each other
  • How to use the triad as a quick checklist when reviewing any system

Why three pillars?

Ask ten engineers what "secure" means and you will get ten answers. The CIA triad is the oldest and most widely used model for cutting that question down to size. It says: a system is secure when, against the threats you care about, it preserves three properties at once, confidentiality, integrity, and availability.

If any one of them breaks, you have an incident, even if the other two are perfect. A bank whose customer data is leaked has a security failure. So does one whose data is silently altered. So does one whose website is unreachable during a payday.

Confidentiality

Confidentiality is the property that information is only revealed to people (or systems) authorized to see it. Encryption is the most famous tool for this, but access control, careful logging, and boring things like locked server rooms count too.

Example: your email password is stored on a service and never shown to anyone, not even the engineers who run the service. When that promise breaks, confidentiality has failed.

Integrity

Integrity is the property that information is not modified in ways the system does not allow. It covers both deliberate tampering and accidental corruption.

Cryptographic hashes, digital signatures, version control, and checksums all serve integrity. Even something as simple as a database constraint ("a balance cannot go negative") is an integrity defense.

Example: you transfer $50 and the receiving bank records $5,000. Confidentiality is fine, no one saw your password. Availability is fine, the transfer went through. But integrity failed.

Availability

Availability is the property that the system is usable when it is supposed to be. A site that is unreachable, a database that has been corrupted into uselessness, a key you have lost, all of those are availability failures.

Distributed denial-of-service attacks, ransomware, hardware failures, and even bad deploys all attack availability. So do well-meant security controls that lock everyone out.

Example: a hospital cannot read patient records during an emergency because its systems are encrypted by ransomware. The data is still confidential and still intact, and still useless.

The three pillars pull against each other

Real systems trade these properties off constantly. A few examples:

  • Strong confidentiality (lots of access checks) can hurt availability when the auth service goes down.
  • Aggressive integrity rules (read-only backups) make it harder to recover quickly after a bug.
  • Maximum availability (cache everything everywhere) often means data lives in more places than is ideal for confidentiality.

Good security is rarely "more of everything". It is choosing where on the triangle each decision should sit, given what you are protecting and against whom.

A small trick

When reviewing any feature, ask three short questions: "Who is allowed to see this?", "Who is allowed to change it?", "How available does it have to be?", and notice when the answers conflict.

Beyond the three letters

Many people add a fourth or fifth property to the triad: authenticity (proof that a message came from who it claims to be from) and non-repudiation (the sender cannot later deny sending it). They are useful additions, but every modern framework still begins with CIA.

Take the quiz