home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.crypt
- Path: sparky!uunet!gatech!destroyer!caen!sdd.hp.com!mips!mips!sfcsun!weaver
- From: weaver@sfc.sony.com (Eric Weaver)
- Subject: RSA keys
- Message-ID: <1992Jul30.213709.26515@sfc.sony.com>
- Organization: Sony Advanced Video Technology Center
- Date: Thu, 30 Jul 1992 21:37:09 GMT
- Lines: 28
-
-
- Summarized for those who haven't a Knuth Vol. II:
-
- Pick two primes, P and Q, which are NOT one greater than a multiple of 3
- (e.g. 37 won't do). Probabilistic prime tests are good for this.
-
- Take a frame of bits F, no shorter than 1/3 the length of PQ (the
- public, encryption key), and less than PQ.
-
- Encrypt: F[enc] = F^3 mod PQ.
-
- Let D (private, decryption key) = [(P-1)(Q-1)/6]+1
-
- Decrypt: F = F[enc]^D mod PQ.
-
-
- P and Q should be of approx. 25% difference in length, and of at
- least 450 bits total. 1000 bits or more, if you can afford it, would
- be best.
-
- Decryption is pretty expensive, obviously, so this technique is
- commonly used only to exchange key and digest info.
-
- Question for cognoscenti: How hard is it REALLY to guess or search D
- from PQ?
- --
- Eric Weaver Sony AVTC 677 River Oaks Pkwy, MS 35 SJ CA 95134 408 944-4904
- & Eng. Asst., KFJC 89.7 - Foothill College, Los Altos Hills, CA 94022
-