Independent research · 2015–2026

Not a gap in speed.
A gap in aim.

Cyberattacks and infrastructure failures are growing. So is the machinery meant to contain them. This study measures both sides — and finds that defense isn't just slow. It's pointed at the wrong target.

4 independent measures EuRepoC v1.3.2 · 3,416 incidents 6 incident-to-response pairs 12×8 cascade matrix

The question

Over the past few years, one failure after another — a data-center fire that took a nation's chat app offline, a telecom breach exposing 23 million people, a ransomware attack that locked students out of their coursework mid-finals — cut off services millions rely on. The obvious question is whether attacks are rising. The harder one, and the one this study asks: has our ability to defend kept pace? Answering it means measuring both sides — the pressure, and the capacity built to absorb it.

SK Telecom breach
Kakao data-center fire
Canvas outage

Four measures · three of pressure, one of capacity

The through-line

The two widest cascades in the study — MOVEit and CrowdStrike, hitting eight and seven of eight sectors — produced no new binding rule. The narrowest did. Regulation doesn't track how far a failure spreads. It tracks whether personal data leaked inside a sector a regulator already polices.

CrowdStrike Progress MOVEit

The two widest cascades in the study ran through single vendors — CrowdStrike's update pipeline and Progress's MOVEit platform — reaching banking, healthcare, aviation, government, and education at once.

The framework

The whole study reads through one idea, the efficiency–fragility tradeoff: the same design choices that make digital systems cheaper and faster also make them more fragile. Moving everyone's services onto a few cloud providers cuts costs — but it means a single break can take down far more than before. Fragility isn't the price you pay to get efficiency. It's part of what efficiency is.

That idea predicts exactly what the capacity analysis finds. Concentration is not a flaw any single company can be cited for, so a regulatory system built to police data breaches simply can't see it. The pressure grows along one axis — structure. The defense is lined up along another — confidentiality. And the two never meet.

Perrow, Normal Accidents (1984)
Interaction × coupling diagram

Charles Perrow, Normal Accidents: Living with High-Risk Technologies (Basic Books, 1984) — the source of the tight-coupling framework this study builds on.