ietf
[Top] [All Lists]

Re: [saag]: Review of: Opportunistic Security -03 preview for comment

2014-08-16 16:52:45
On Fri, Aug 15, 2014 at 05:06:45PM -0400, Paul Wouters wrote:
On Fri, 15 Aug 2014, Viktor Dukhovni wrote:
Opportunistic DANE TLS for SMTP is security

Some disagree about the use of the term opportunistic in this case.
If an SMTP client supports DANE, and is contacting an SMTP server
supporting DANE, there is nothing opportunistic about it. It MUST use
encryption and MUST NOT fall back to cleartext.

I suppoe it's all about definitions.  Viktor's point is that the client
is willing to use TLS or not depending on whether it knows (local
config) or could know (DANE) that it must, with the latter being
opportunistic.  The opportunity in the DANE case is really the server's:
if it publishes TLSA RRsets...

It is security against passive attacks,
that is, for a different threat model.

I don't disagree. But it is still only encryption.

I would have objected regardless.  Opportunistic security is a
better match than OE for the content of the draft.  I would not
have objected to Opportunistic Cryptosecurity, but it is not a
compelling improvement.

While not compelling, it is an improvement :P

Cryptosecurity reals like snake oil to me.  Or, at least, that's how I
fear it will read to non-initiates.  I _feel_ that cryptosecurity is
worse than not an improvement.  It's a feeling, difficult to back up
with facts and readon.  But bikeshedding is like that (please do not
take that as an insult.  Bikeshedding is a very common kind of argument,
and I've done it myself.  It happens, we call it, and then generally it
ends.).