Re: [dane] [saag] Need better opportunistic terminology

[Joe Touch <[e-mail]>]

On 3/12/2014 10:06 AM, Nico Williams wrote:
> On Wed, Mar 12, 2014 at 11:49 AM, Joe Touch <[e-mail]> wrote:
...
>> BTNS uses unsigned key exchanged, and there's nothing "opportunistic" about
>> it. Unsigned authentication is the goal from the start.
>
> For me the goal was to use channel binding at the application layer.

Having unsigned key exchange has two ultimate uses:

	- raising the bar for attacks
		above 'send a RST' but below MITM

	- providing a network-level mechanism that can be linked
	to security at higher layers

App-layer channel binding could be useful for BTNS, or for other 
approaches too (e.g., where a single key protects a set of ports).

> But we never got there: no one seems to care much about end-to-end
> IPsec, sadly.  (Well, it's not that no one cares, but that it's too
> late now; TLS is king.)

TLS still doesn't protect the transport or network layers.

>> OE as defined in RFC 4322 isn't about using unsigned key exchange; the
>> "opportunistic" sense is derived from using keys retrieved from DNS without
>> prior agreement. That's not what happens in BTNS.
>
> Stephen has it right: OE in the RFC4322 sense is about applying
> protection even when SPDs don't agree on this, but it still requires a
> keying infrastructure (i.e., trust paths).

Right, but then "O" isn't quite the right term for security that avoids 
the need for keying infrastructure altogether.

Joe