[noise] Thoughts on semi-deterministic encryption

[Jonathan Moore]

On Tue, Aug 26, 2014 at 2:16 PM, Stephen Touset <stephen at squareup.com>
wrote:

>
> For the purposes of Noise, if you don’t have a good RNG you aren’t going
> to be able to come up with strong ephemeral keys anyway, right?
>

This is a further issues to be addressed. I think some progress is possible
here as well.  One possibility is to use a  construction for ephemeral keys
that mixes in the long term keys so that pore entry means one only looses
forward security but not confidentiality, another possibility might be to
make sure that the session key mixes entropy from both sides of the
connection so that only if both hosts have pore entropy do you loose.  I am
not claiming that either of these will definitely work out but they are
things to think about.

Also I don't specifically bring these questions up with respect to noise
today. Trevor suggested to me that this list is open to generally crypto
discussion that might be relevant to noise like systems.

-Jonathan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://moderncrypto.org/mail-archive/noise/attachments/20140826/6ec4ab81/attachment.html>