[Ietf-krb-wg] PKINIT with smart-cards

Hello,
 I've been thinking of an attack relating to Public key Kerberos
with smart-cards and I'd appreciate if anyone could verify
that what is described is not due to my misunderstanding
of the protocol.

The legitimate owner is equipped with a smart-card containing
an RSA private key and a certificate that restricts it to
signing (meaning Diffie-Hellman PKINIT will be used).

(two scenarios illustrating the same impersonation attack)

1. The attacker steals the smart card and the PIN required to operate it.
Performs several precalculations for PK_AS_REQ with future timestamps
that span for next few years. Places the smart-card back to the owner.
The owner has no way of detecting that the card has been used and
the key although no compromised, it can be used to impersonate
him by the thief.

2. A malicious smart-card reader is installed in a way that legitimate
users will use it to connect using Kerberos. The reader operates
as expected, but during idle time that the card is inserted the reader
precalculates PK_AS_REQ messages with future timestamps. The
owner has no way of knowing that.

The attacker having the PK_AS_REQ signed message can initiate
connections impersonating the legitimate owner at any time
of his choice.

best regards,
Nikos
_______________________________________________
ietf-krb-wg mailing list
ietf-krb-wg@lists.anl.gov
https://lists.anl.gov/mailman/listinfo/ietf-krb-wg