home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.crypt
- Path: sparky!uunet!infonode!ingr!b11!craig!craig
- From: craig@jido.b11.ingr.com (Craig Presson)
- Subject: Re: DES Encryption/ Encrypting more than once.
- In-Reply-To: ferret@works.uucp's message of 12 Oct 92 21: 45:43 GMT
- Message-ID: <1992Oct13.174505.24230@b11.b11.ingr.com>
- Sender: usenet@b11.b11.ingr.com (Usenet Network)
- Reply-To: craig@jido.b11.ingr.com
- Organization: Intergraph Corporation, Huntsville, Alabama
- References: <wa6JsB7w165w@works.uucp>
- Date: Tue, 13 Oct 1992 17:45:05 GMT
- Lines: 47
-
- In article <wa6JsB7w165w@works.uucp> ferret@works.uucp (Dave Ferret) writes:
- Just a sidenote to 'Hackers' words...
-
- There are also encryption algorithms that when used to encrypt the
- plaintext over and over and over, will yield the un-encrypted text. (Ie:
- Its a circular encryption -- Sorry, I don't know the correct term here)
-
- Cyclic, as in subgroup. The question is often phrased, "Is DES a
- group?" (under composition of encryption). Last I read (Crypto '90?),
- it was strongly suspected but not conclusively proven to contain no
- cycles* (recall that a finite group must contain at least one cycle).
- A common suggestion for using DES as a superencipherment for itself is
- E_k1(D_k2(E_k1(P))), where E_k1 = encrypt with key k1, etc. If
- encrypting twice with the same key were a lot stronger than encrypting
- once, then no doubt the algorithm would just do that (bearing
- performance limits in mind, of course). If you get FIPS pub 46 (and
- 81) or an intrology to cryptoduction and read through the algorithm,
- you will see that it already contains 16 rounds with key
- transformations and that things get pretty messy in there :-)
-
- Also, the reasons for tinkerig with DES could possibly be for one reason,
- it was claimed 'unreversable' (Right?) and also that its the basis for
- Unix' CRYPT().
-
- Eh? It is used in Unix password encryption to _create_ a one way
- function -- a random 12-bit salt value is added in to the password,
- then that is used as a key to encrypt a 64-bit constant 25 times. The
- multiple encryption is not there to strengthen the algorithm as much
- as to slow it down, to make dictionary attacks less feasible.
-
- The account on pp. 30-31 of Garfinkel and Spafford, _Practical Unix
- Security_ (O'Reilly & Assoc., 1991) is brief and complete.
-
- I'm sure this will be in the promised FAQ, either in all its
- gory detail or by reference to G&S or equivalent.
-
- -- Craig Presson
- * - here's a reference I haven't read yet, that should be definitive:
- @inproceedings{campbell,
- author = "Campbell, K.W. and Wiener, M.J.",
- year = 1993,
- title = "Proof that {DES} is not a group",
- booktitle = "Advances in Cryptology --- Crypto '92",
- publisher = "Springer-Verlag",
- address = "New York",
- note = "To appear"}
- Note to me: call & see if this is out yet.
-