The four variables
How long must this data stay confidential
The first variable: how many years a specific piece of data must remain secret, and why that number, not the algorithm, drives urgency.
5 min read
What you'll be able to do
- Define confidentiality lifetime and explain why it is the variable that starts the reasoning chain.
- Give examples of data with short, medium, and long confidentiality lifetimes.
- Explain why a short-lived-data assumption should be justified, not assumed.
The reasoning chain in this course starts with data, not with cryptography. Before asking what protects a piece of data, ask how long that data actually needs to stay confidential. This is its confidentiality lifetime, and it varies by orders of magnitude depending on what the data is.
| Data class | Typical confidentiality lifetime | Why |
|---|---|---|
| Session analytics, ephemeral tokens | Hours to weeks | Value to an adversary decays almost immediately after the session ends. |
| Financial transaction records | Several years | Regulatory retention and dispute windows, not secrecy forever. |
| Health records, genomic data | Decades, sometimes the patient's lifetime | The sensitivity of the underlying fact does not expire on a schedule. |
| State, diplomatic, or long-range trade-secret data | Decades or longer | Value can persist well past the lifetime of the people or products involved. |
The number that matters is not how sensitive data feels today — it is how many years from today it must remain confidential. A record created five years ago with a twenty-year lifetime has fifteen years left, not twenty.
Confidentiality lifetime alone does not create risk. Data that must stay secret for thirty years but can never leave a single air-gapped machine is not HNDL-exposed no matter how long its lifetime is. The next lesson adds the second variable: whether the data can actually be captured.
Marking a lesson complete only updates this browser.