defense in depth
termsecuritygovernance & risk
Layering multiple independent controls so that if one fails, others still stand.
Defense in depth assumes any single control can be bypassed, so it stacks them, network, host, application, identity, forcing an attacker to defeat several in a row. No layer is perfect; the strength is in the overlap.
Defense in depth assumes your controls will fail, and asks what happens next. Any single mechanism, a firewall rule, an authentication check, an input validator, will eventually be bypassed, misconfigured, or found to have a flaw disclosed on a Tuesday. Layering means the failure of one is survivable rather than terminal.
The idea is borrowed from fortification, and the analogy is worth taking seriously because it includes the part people skip: layers must be different in kind, not merely repeated. Three firewalls from the same vendor sharing a common vulnerability are one layer wearing three hats. Genuine depth mixes prevention with detection, network controls with identity controls, and automated enforcement with human review, so a single class of mistake cannot defeat all of them at once.
The cost is honest and worth stating. Every layer adds latency, expense, and operational surface, and layers that nobody monitors are decoration. The discipline is deciding which failures you are actually defending against, then building the smallest set of genuinely independent controls that survive those failures, rather than accumulating products because each one seemed prudent in isolation.
Also known as: defense-in-depth, defence in depth, layered security