Mahesh Jethanandani has entered the following ballot position for
draft-ietf-ipsecme-ikev2-downgrade-prevention-07: No Objection

When responding, please keep the subject line intact and reply to all
email addresses included in the To and CC lines. (Feel free to cut this
introductory paragraph, however.)


Please refer to 
https://www.ietf.org/about/groups/iesg/statements/handling-ballot-positions/ 
for more information about how to handle DISCUSS and COMMENT positions.


The document, along with other ballot positions, can be found here:
https://datatracker.ietf.org/doc/draft-ietf-ipsecme-ikev2-downgrade-prevention/



----------------------------------------------------------------------
COMMENT:
----------------------------------------------------------------------

I agree with both of Boucadair's DISCUSS points.  RFC 9242 is a normative
dependency because Section 7.1 contains a MUST that depends on the IntAuth
term defined in RFC 9242, and the entire IKE_INTERMEDIATE interaction is
meaningless without reading that RFC.  RFC 5723 is a normative dependency
because Section 7.2 contains two MUSTs ("MUST be stored in the ticket" and
"MUST act the same way when doing resumption") that depend on the ticket
mechanism defined in RFC 5723.  Both should be listed as normative references.

Section 9, first paragraph — incomplete statement:

433 >    The IKEv2 extension defined in this document protects against
434 >    downgrade attacks on IKEv2 described in Section 4.  It only provides
435 >    this protection when both peers implement the extension.

In my view, the second sentence is an incomplete statement.
Section 5 correctly identifies two conditions: both peers must support the
extension AND at least one non-compromised authentication key must be used in
the protocol run.  The opening of Section 9 omits the second condition, which
is the substantive one that the rest of Section 9 spends several paragraphs
analyzing (PSK, EAP, etc.). I would suggest aligning it with the Section 5 
language.

---

Section 5 note — important limitation not cross-referenced in Section 9:

318 >    |  This mechanism only detects tampering with the IKE_SA_INIT
319 >    |  transcript.  It does not change the transform-selection rules
320 >    |  for the responder.  In particular, it does not by itself prove
321 >    |  that the strongest mutually supported cryptographic algorithms
322 >    |  were selected.  That property depends on responder selection
323 >    |  policy, which remains governed by [RFC7296] and local
324 >    |  configuration.

This limitation — that the extension authenticates the final
IKE_SA_INIT transcript, but cannot prevent an attacker from first using a
forged INVALID_KE_PAYLOAD to force a retry on a weaker key exchange, thereby
influencing what that final transcript contains — is security-relevant and
belongs in Section 9.  Readers who consult Security Considerations directly to
assess the residual attack surface will miss this caveat entirely.  I would
suggest adding a paragraph in Section 9 that summarizes this limitation and
cross-references Section 5.

---

Section 8 — attack indistinguishability not noted:

427 >    Deployments that rely on this mechanism for downgrade resistance
428 >    should provide a local policy to require successful negotiation of
429 >    this extension and to abort the connection otherwise.

This section is the entirety of the operational guidance.  In my view, it
would benefit from noting — consistent with what Section 5 says — that when
the mechanism detects a downgrade attempt, the resulting authentication
failure is indistinguishable from any other authentication failure (incorrect
credential, configuration mismatch, and so on).  This has direct operational
consequence: operators cannot distinguish a downgrade attack from a
misconfiguration, so logging and alerting on these failures carries limited
diagnostic value.



_______________________________________________
IPsec mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to