Replay Attacks
A valid request can still be harmful when captured and repeated; authenticity does not automatically imply freshness or exactly-once effect.
Frame the problem
Security starts with a concrete asset, attacker capability and trust crossing.
Why the system fails
The receiver verifies identity or signature but does not bind the message to a time window, nonce or processed event id.
The important question is not “what is Replay Attacks?” but “which assumption let untrusted data or an over-scoped identity cross previously valid message → current state change?” Trace the decision at the boundary, then constrain what can happen after the first control fails.
Design the control in layers
Start with the control closest to the interpretation or privilege boundary: Bind signatures to timestamp and body Then add a control that reduces blast radius and telemetry that proves the decision was enforced.
The resulting design is not labelled secure. Record the identified controls, the known failure paths, the remaining exposure, and the evidence you would need during an incident.
| Prevent | Detect | Recover |
|---|---|---|
| Bind signatures to timestamp and body · Reject old or reused nonces and event IDs · Make operations idempotent where possible | Duplicate event IDs and signature timestamps outside the window | Contain the affected identity or component, scope impact from audit evidence, and preserve a regression test. |
Key points
- Asset: Integrity of payments, mutations and signed messages.
- Boundary: Previously valid message → current state change
- Primary control: Bind signatures to timestamp and body
- Detection signal: Duplicate event IDs and signature timestamps outside the window
- Always ask what limits damage when the primary control fails.
Boundary control exercise
This lesson uses the shared boundary-control exercise.
Follow the attack
Safe conceptual simulation: capability → missing control → crossed boundary → asset impact.
- 1Attacker starts with: Someone who obtains a previously valid request or event.
- 2The receiver verifies identity or signature but does not bind the message to a time window, nonce or processed event id.
- 3The weak or missing boundary control is crossed: Previously valid message → current state change
- 4Impact: Duplicate charges, repeated mutations or reused authorization.
- Duplicate charges, repeated mutations or reused authorization.
Defend, detect, recover
One prevention is a single point of security failure. Layer it and make failure observable.
- • Bind signatures to timestamp and body
- • Reject old or reused nonces and event IDs
- • Make operations idempotent where possible
- • Duplicate event IDs and signature timestamps outside the window
- • Contain the affected identity or component.
- • Scope access from audit evidence.
- • Fix the boundary and add a regression test.
- • Misconfiguration and new access paths can bypass the intended control.
- • A privileged insider or compromised control plane may still reach the asset.