riseup.net OpenPGP Best Practices article

[Robert J. Hansen]

> While in principle I agree that 2048 bit key is strong enough for most
> uses, comparing 3DES keys space (or any other symmetric encryption
> algorithm) and RSA (or some other public key system) key space is a
> bit like comparing apples and oranges. If you crack the 3DES
> encryption of a message you have cracked that particular message. If
> you crack the RSA key, you have cracked all messages. So the effective
> key space of your public key should be larger then the key space of
> the session key(s).

This is, IMHO, a complete nonissue.

If your adversary has the ability to brute-force a 112-bit keyspace,
then you are now living in a world where crypto cannot protect you.