Re: EE Times on PRZ

Hal wrote:
>This, from a sidebar, is really surprising:  "In contrast, public keys
>allow the overt publication of an encryption key, because decryption keys
>can only be derived through a mathematically difficult process, such as
>large prime-number factoring.  Contrary to popular belief, the NSA can
>decrypt public keys of most practical key sizes."  I wonder what this
>means?  If it is a claim that the NSA can factor 1024 bit moduli that
>would certainly come as a big surprise.  If they are saying that they can
>do 512 bits that would be more believable although of interest.  It is
>strange that the author would include a statement like this without
>attribution or evidence.

Another quote from the article posted elsewhere said that, "PGP, which is
based on the Diffie-Hellman public-key technology developed in the 1970s..."
This is technically true, since all public-key work (including RSA) is based
to some extent on DH.  It could be, however, that the author is confusing
public-key technology with Diffie-Hellman public-key in particular, which
(as I understand it) is not particularly secure.

THUS SPAKE jalicqui@prairienet.org (Jeff Licquia):
# Hal wrote:
# >large prime-number factoring.  Contrary to popular belief, the NSA can
# >decrypt public keys of most practical key sizes."  I wonder what this
# >means?  

Just as healthy paranoia, that's worth persuing.
But I bet the author didn't know what they was talking about. 
 
# Another quote from the article posted elsewhere said that, "PGP, which is
# based on the Diffie-Hellman public-key technology developed in the 1970s..."
# This is technically true, since all public-key work (including RSA) is based
# to some extent on DH.  It could be, however, that the author is confusing

DH uses "discrete log" as the hard problem, and very straightforward
mathematics.

RSA uses "factoring" as the hard problem, and a very clever back door.

How do you decide if one is based on the other?

# public-key technology with Diffie-Hellman public-key in particular, which
# (as I understand it) is not particularly secure.

It's still up in the air, isn't it, whether the discrete log or 
factoring is the harder to crack.   My intuition is they're the
same hard.

I know of no problem with DH that RSA doesn't have similar problems.

			strick