Re: Random seeds

Eric Young (eay@cryptsoft.com)
Fri, 26 Jun 1998 09:53:15 +1000 (EST)

• Next message: Wei Dai: "Re: Microsoft patent"
• Previous message: Jeff Foy: "Re: Blowfish II name"
• In reply to: Gary Howland: "Blowfish II name"

On Thu, 25 Jun 1998, Tom Weinstein wrote:
> > At 08:37 PM 6/24/98 -0700, Tom Weinstein wrote:
> > >available, but the best thing is a good hardware RNG.
> > The only good RNG will be a hardware RNG. And if you
> > are really picky, then the only good RNG is a NSA
> > certified hardware RNG, like on a Fortezza card.
> How about: The only good RNG is a hardware RNG that I can analyze the
> design for (unlike on a FORTEZZA card).

What always worries me is the potential for the smart-card RNG to be a good
PRNG (how can you tell the difference if you cannot analyze the card design?)
in which a retrievable serial number is critical to the algorithm. If the
right people know the algorithm and serial number (initial seed), the random
stream can be generated etc etc.... I still like the idea of external input,
which is known to be used in the mixing process.

eric

• Next message: Wei Dai: "Re: Microsoft patent"
• Previous message: Jeff Foy: "Re: Blowfish II name"
• In reply to: Gary Howland: "Blowfish II name"