Re: [TLS] PSK in 1.3?

[Yoav Nir <[e-mail]>]

> On Oct 21, 2014, at 11:05 AM, Peter Gutmann <[e-mail]> wrote:
> 
> Manuel Pégourié-Gonnard <[e-mail]> writes:
> 
>> In addition to what Watson said, it's generally a bad idea to roll one's own
>> crypto protocol, it's too easy to get things wrong. Even with PSK, TLS brings
>> you important things: a record protocol, a way to derive session keys from the
>> long-term secret key, etc.
> 
> I'm just reviewing a design right now for which there's been an argument put
> forward that it's too complicated and you could roll your own protocol that
> would be much simpler.  The implementers are using EAP-PSK, which is a wee bit
> complex, but then it has exactly the benefits you cite, everything's well-
> defined (although the lack of test vectors is annoying), you just grab the RFC
> and code it up and you're done.  Sure, you could homebrew your own
> application-specific protocol that's simpler, but then you'd end up with a
> proprietary, non-vetted, nonstandard one-off custom job that nothing else on
> earth can talk to.  That may be fine if you're building an Arduino-controlled
> electronic cat-door for your home, but not if it's commercial SCADA gear
> that's going to be deployed all over the world.
> 
> We'll see how the discussion over this pans out...
> 
> Peter.

Is that related to http://tools.ietf.org/html/draft-nir-tls-eap-13 ?

Yoav