Post-compromise security
Post-Compromise Security
Property that a connection recovers its confidentiality after an attacker has obtained its secrets, because later traffic uses fresh keys independent of the compromised ones.
In our briefings
The claim each of the newest briefings on Post-compromise security would have you leave with.
Revision -09 reverses -08 on passport results. Revision -08 let an attester cache a verifier's signed results and use them "for any number of subsequent TLS handshakes" while its freshness policy allowed. Revision -09 says results "from a previous TLS handshake MUST NOT be presented in a subsequent handshake, even if their validity period has not expired." The attester must fetch them after ServerHello and before its Certificate message, the verifier must sign the attested TLS identity key and this handshake's binder into them, and the relying party must check both. Revision -08 had said a passport topology was the equivalent of OCSP stapling (where a server presents a status response it fetched earlier); -09 says the passport model now differs from stapling, because the verifier is contacted on every handshake. The change log in Appendix A.1 describes this as binding results to the current handshake and does not say that -08's caching allowance is withdrawn.
The attested-TLS draft drops cached verifier results, so the verifier now sits inside each handshakeBI-FIND-2026-2184
RFC 8446 requires an application profile before a protocol uses 0-RTT, the early data a client can send in its first flight. The application must explicitly request its use, define which interactions are safe to replay and handle rejection; early data has no forward secrecy. Servers should prevent replay across the deployment, but the mandatory minimum is accepting a given 0-RTT handshake at most once per server instance. Application retries need application defenses, and clients must send only early data they consider safe to replay.
TLS 1.3 requires applications to choose which early data is safe to replayBI-FIND-2018-0012
Cited in 2 briefings
Every briefing that discusses this term, newest first.
- The attested-TLS draft drops cached verifier results, so the verifier now sits inside each handshake
Draft · · IETF Datatracker · Thomas Fossati et al. · Oct 14, 2026
- TLS 1.3 requires applications to choose which early data is safe to replay
Article · · RFC Editor · Eric Rescorla, Mozilla · Aug 10, 2018