[2187] in Kerberos
Re: New User Accounts
daemon@ATHENA.MIT.EDU (Li Gong)
Thu Sep 10 21:12:45 1992
Date: 10 Sep 92 03:26:12 GMT
From: li@oracorp.com (Li Gong)
To: kerberos@shelby.Stanford.EDU
In article <9209031516.AA09658@Athena.MIT.EDU> smb@ulysses.att.com writes:
> Dictionary attacks can eb stopped using the protocls developed
> by li Gong et al, and by Bellovin & Merrit. Note that the
> latter necessitates the use of public-key, and both have
> (probably acceptable) an overhead.
>
>Both protocols require public key cryptosystems. One significant difference
>is that the Lomas/Gong/Needham/Saltzer protocol requires a known public
>key for the server, and the Bellovin/Merritt one does not. Both require
>high-quality random numbers. Bellovin/Merritt imposes some possibly-tricky
>constraints on what a public key can look like.
Just a bit of up-to-date information on this. In our new paper
L. Gong, T.M.A. Lomas, R.M. Needham, and J.H. Saltzer, ``Protecting
Weak Secrets from Guessing Attacks'', to appear in {\em IEEE Journal
on Selected Areas in Communications''}, first quarter, 1993.
we have consolidated and greatly expanded our previous efforts. In
particular, we described several more practical derivations of our
example protocol (published in ACM 12thSOSP, 1989) to deal with
concerns such as minimization of number of messages, using nonces
instead of synchronized clocks (since they may not be available),
one-way identification, and so on. In particular, we showed a simple
derivation that does away with the need to let the server's key known
to other parties in advance, and another derivation for direct
authentication (i.e., not via a server S). This last protocol's
functionality is similar to that of the Bellovin/Merritt protocol and
has similar constraint on the public system. However, due to our use
of confounders, even when a session key is compromised (say by
cryptanalysis), one cannot masquarade as another or mount a dictionary
attack on the password.
The final version should be ready in 2-3 weeks time. Anyone who
cannot wait till first quarter 1993 could drop me a line for a ps copy
or maybe a hardcopy.
Li
--
Li GONG, PhD | Email : li@oracorp.com
ORA Corporation | Fax : 607-277-3206
301A Dates Drive | Switch : 607-277-2020
Ithaca, New York 14850, USA | Direct : 607-272-0736 (dial 217 on tone)
