[2272] in Kerberos
Re: Clarifying preauthentication
daemon@ATHENA.MIT.EDU (Theodore Ts'o)
Thu Oct 8 15:19:12 1992
Date: Thu, 8 Oct 92 14:46:09 -0400
From: tytso@Athena.MIT.EDU (Theodore Ts'o)
To: John Linn. LKG1-2/A19. {-486 or DTN 226}-7384. 08-Oct-1992 1147
Cc: kerberos@Athena.MIT.EDU, linn@albeit.enet.dec.com
In-Reply-To: John Linn. LKG1-2/A19. {(508)-486 or DTN 226}-7384. 08-Oct-1992 1147's message of Thu, 8 Oct 92 11:45:10 EDT,
Date: Thu, 8 Oct 92 11:45:10 EDT
From: John Linn. <linn@albeit.enet.dec.com>
As I understand the scheme, it doesn't directly prevent off-line password
attacks. Rather, its real value (and it is real value) is indirect,
making it harder for an attacker to actively acquire the necessary
basis against which to attempt such attacks.
John,
Oops! You're absolutely right; I wasn't sufficiently careful
when I was typing up the text of my announcement. I was being far too
loose there --- thank you for calling me on it.
The correct statement is that it makes off-line password attacks
much more difficult, since the attacker must eavesdrop on the network
for a ticket exchange of the user who is under attack, instead of merely
asking the KDC for a ticket with which to mount the off-line password
attack.
One way to absolutely prevent off-line attacks would be to
create a new preauthentication type which requests that the encryption
portion of the KRB_AS_REP also be encrypted in a service key found in
/etc/srvtab, as well as in the user's secret key. But of course, that
only works if the machine on which you are logging in has an /etc/srvtab
file to which the login or kinit program has access.
The other method, which you mentioned, is that with certain
types of hand-held authenticators, it might be possible to use the value
sent back from the authenticator as part of the key encrypting the
KRB_AS_REP message. This assumes that you do not send that value back
to the KDC (so it is not exposed to the network), and that the value
displayed by the hand-held authenticator is long enough to be used
either exclusively or supplementarily as keying information.
Again, my apologies for my mis-statement in my announcement; I
really should have been more careful.
- Ted
