[Cryptography] Toxic Combination

[Viktor Dukhovni]

On Mon, Dec 01, 2014 at 09:58:25AM +1100, Alfie John wrote:

> And this is taken advantage of every day by phishing attacks. However
> although your solution of setting up DNSSEC and DANE is the _correct_
> solution, it's just too complex and hard to get right for a lot of
> system admins so it's not going to get uptake - just look at how PGP is
> also the _correct solution_ for encrypting messages and yet has not had
> the uptake since 1991!

[ Note, I am not saying that I expect imminent DANE adoption in the
  HTTP stack. ]

The nature of the difficulty is much greater for PGP than for
DNSSEC.  I am not going to deep dive into the details of that.
Email content encryption imposes usability barriers on both parties,
while DNSSEC is only a burden on the zone maintainer side and
tooling to address this "local" problem is improving.

> I think a better solution would be something like implementing Digest
> Authentication (RFC 2069, but replacing MD5 with something like AES-256
> and allow it to be upgradable) in the browser. The password field value
> would then be replaced with the value from the DA call and no secrets
> would be leaked. This solution would get way faster adoption.

This also faces adoption barriers.

-- 
	Viktor.