The layers should prevent some attacks, reveal others, limit their movement or impact, and support recovery.
Effective depth comes from complementary and reasonably independent safeguards. Repeating the same control in several places can create the appearance of protection while preserving a common failure mode — for example, several tools that all depend on the same identity system, telemetry source, or mistaken policy.
Real depth also spans functions, not just vendors: prevention buys time, detection notices the failure, response limits the spread, and recovery restores service. Exercises that deliberately bypass one layer — a missed alert, a phished credential — reveal whether the remaining layers genuinely compensate or quietly share the same assumptions.
Key points
Possible layersSecure design, identity controls, hardened endpoints, segmentation, monitoring, response, protected backups, and exercised recovery.
Design principleAssume individual safeguards can be bypassed, misconfigured, unavailable, or compromised.
Validation needTest whether layers have sufficiently independent failure modes and whether detection and recovery still work after a preventive control fails.
Important limitationExcessive layering can increase complexity, obscure ownership, and create new failure paths; every layer needs a defined purpose.