-----BEGIN PGP SIGNED MESSAGE----- On Fri, 15 Nov 1996, Peter Hendrickson wrote:
(My apologies to those who are seeing this twice. I sent it yesterday but it does not seem to have made it through.)
A number of us have been concerned about how PGP generates entropy. Striking the keyboard beats using time as a source of random numbers, but the degree of entropy is not well understood. Are there machines where - for some reason - the keyboard strikes fall into some sort of pattern?
And that's just when you are generating your public/private key pair. What happens when you are just generating 128-bit keys for individual messages? Where is the entropy coming from? I don't understand completely, but somehow PGP collects entropy from the system and then runs it into IDEA and then uses the numbers from the output. When the program is not running, a pool of this data is kept in randseed.bin.
PGP uses the hash of the message to preprocess the randseed.bin file and postprocesses it with the session key. It uses the randseed.bin file as a PRNG to generate the session key.
We already know it's a problem because it is hard to understand what it is even doing, much less determine if it is consistent with sound engineering practice.
PGP uses a variation of ANSI X9.17 for key generation. This is considered to be a secure method. Mark - -- finger -l for PGP key PGP encrypted mail prefered. -----BEGIN PGP SIGNATURE----- Version: 2.6.3 Charset: noconv iQEVAwUBMozI/CzIPc7jvyFpAQG+NAgAvPWsnTb7S/ohtHcyQG8DCaSlKtHt+FMF 4qtLNtBzi8Y27WRtl9HHojdgYGVxuT45e4W9r3WOtLjaMkrYPdIKUcxSABkhr3Zd 03pfVePg/ws3p+ynmpInMj8vr3lAYPlcFp5cPJdKl8WhZTSxXFCQ92q7xckzoW5S J9iUiIgnYf8n7qcfNSQI9rVw2d3Dv6rccqdYtfNA+UUe6jlwIbooITZ89EhVHWzw 2BTZF+xOVenK058uQQFIzU99Bkaz35Hl3CC41TX/Ka2CZXBhAzjD++xtcapu1PE7 moEzKde5cgYZ+R1d7TZwvudkAtdlx7xo4GDfxI5O+KR+FOyDNqMdoQ== =JTNn -----END PGP SIGNATURE-----