[Ach] Improving Applied Crypto Hardening Draft

[ianG]

On 10/01/14 16:32 PM, Manuel Kraus wrote:

> Well, the naming scheme seems to be quite inconsistent anyways. DHE vs. EDH, for example, meaning the same.


I think (not 100% sure) that it is this:

DHE == Diffie-Hellman (key) Exchange
EDH == Ephemeral Diffie-Hellman (key exchange)

The term Ephemeral refers to the key being a public-key style of 
operation (DH), but not saved in any context.  The assumption of the PKI 
crowd is that public keys have to treated as religious artifacts to be 
preserved and handed from generation to generation in ceremonies 
involving sacrifices and cold and fire and brimstone.

Hence, when someone invented a new approach (use DH to generate a 
protocol-forward-security PFS), they needed a whole new term to indicate 
how evil this particular perversion was to everyone.  And evil as it 
was, typically the EDH modes were not promoted, along with that other 
abomination:

ADH == Anonymous Diffie-Hellman (key exchange)

So sad.  If they hadn't had such massive broomsticks up such tight 
places, we would have had opportunistic crypto across the entire 
webspace within a few years, and the upgrade path for cert SSL would 
have been about a tenth as hard.  Old rant.....


> But I'd like to recommend to strictly use the common crypto stack given acronyms, instead of putting new ones in. Since this draft ist targeted at system administrators, other naming variants are useless confusing (it has confused me already). For me it is the openssl package, which rules, so I would vote for "ECDHE" in this case. Well, unless it isn't named something else in other valid crypto stacks... or the guys around openssl change it. :-P


good luck !

iang