symmetric vs. asymmetric in group use

peter.segment at wronghead.com peter.segment at wronghead.com
Tue Aug 28 08:37:43 CEST 2012


On 28/08/12 01:54, No such Client - nosuchclient at gmail.com wrote:
>> If you are restricting heavily the people you share your public key with,
>> why not simply use a symmetric algorithm, forgetting public key
>> cryptography completely?

 > Uhh. because the benefit of pubkey encryption is still there, minus the
 > risk of having pubkeys there forever permanently.

As a note of some possible interest, members of (obviously
hypothetical) C-Z/SUV (cf. the "GPG simplified" thread that this
one appears to be an offshoot of) have considered using symmetric
crypto. The most important argument was that by the same 
"out-of-channel" method used to verify correspondent's public key, a
two-correspondents specific symmetric key could be exchanged,
and that the public key system implementations are much more
complex and therefore fragile, and cryptographically, public has
three critical crypto algorithms (or components) that must not
break: RNG, asymmetric and symmetric cipher, while the symmetric
has only one: symmetric cipher.

However (since for obvious reasons a single, group-wide key is
out of the question) there would be only ~2*n~ keys to manage
for public, and (n**2 - n) for symmetric.

Peter M.





More information about the Gnupg-users mailing list