home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.crypt
- Path: sparky!uunet!telebit!phr
- From: phr@telebit.com (Paul Rubin)
- Subject: DES generates A_(2^64)?
- In-Reply-To: keith@bnr.ca's message of Thu, 15 Oct 1992 21:13:00 GMT
- Message-ID: <PHR.92Oct15174530@napa.telebit.com>
- Sender: news@telebit.com
- Nntp-Posting-Host: napa.telebit.com
- Organization: Telebit Corporation; Sunnyvale, CA, USA
- References: <1992Oct13.174505.24230@b11.b11.ingr.com> <1992Oct15.125830.25539@bnr.ca>
- <unruh.719169829@unixg.ubc.ca> <1992Oct15.211300.27098@bnr.ca>
- Date: 15 Oct 92 17:45:30
- Lines: 17
-
- Michael Wiener and I submitted a paper to Crypto '92 entitled "DES is not
- a Group". The principal argument is as follows.
-
- Consider the closure of DES encryptions under composition. Our claim is
- that this group must include permutations which are not equivalent to DES
- encryption with any key. Each of the n_i collected above must divide
- the order of the group. Compute n = lcm( n_i, i=1..k ). If n > 2^56 then
- the group must be larger than the set of DES encryptions.
-
- The paper includes the lengths of several hundred cycles of DES.
- The lcm of those lengths exceeds 10^2499. It is clearly not practical
- to use E^n(x) to recover x for such large n.
-
- I heard that someone in Eastern Germany recently proved a stronger
- result, that DES generates the alternating group on 2^64 letters.
- This had been conjectured a while ago, though I don't know what the
- evidence for it was. Anyone know more about this?
-