NODE 9f9ab965Re: +-=*^
George A. Gleason <gg@well.sf.ca.us>Sat, 10 Oct 92 01:59:50 PDT
I've seen a number of postings so far about "secure *and* authenticated"
requiring advance in-person distribution of key material. This would seem
to eliminate the main advantage of public key systems, i.e. open
distribution of public keys. So as long as we're handling key material in
person, how about one-time systems, eh? Absolutely secure, provably so,
obsolescence-proof, simple & straightforward. Not particularly exciting
from a theoretical point of view, but one-time systems are practical and on
the bottom line, they work. What do y'all think...?
-gg@well.sf.ca.us
NODE 6ef03568+-=*^
Eric Hughes <hughes@soda.berkeley.edu>Sat, 10 Oct 92 07:29:03 PDT
George recommends one-time pads.
The key distribution problem for one-time pads is *much* worse than
for public key systems, or even conventional secret key ciphers for
that matter. You still have to exchange keys without transmission
(i.e. face to face meetings again, or mail, etc.). Anything that is
secure for exchanging a one-time pad is also secure for exchanging
public keys. Then you have to do this again when your pad runs out.
The bandwidth required for one-time keys is much higher than for
conventional keys to boot.
But the biggest advantage of public key systems is that I can sign
someone else's key, and if you know my key, then you know his.
To put it more humorously, you will have exchanged cryptographic
fluids with everyone I have as well. This is a good thing.
Eric