Re: [Endymail] Restart Endymail to discuss the Mathematical Mesh?

[Phillip Hallam-Baker <[e-mail]>]

On Mon, Apr 8, 2019 at 2:32 PM Steven M. Bellovin <[e-mail]>
wrote:

> Slightly off-topic, but folks on this list may be interested in
>
> https://www.cs.columbia.edu/~smb/papers/eurosys-2019-submission408-e3-final-1.pdf
> .
> We plan to release the code as open source.
>

Looks interesting. I can see some clear convergence in mechanism. Though it
is not clear to me how you are dealing with the per device keys.

My approach is to split the decryption key. Which means I can add and
remove devices etc. in no time at all. When I provision a new device, the
administration device knows the private key for each of the applications I
might provision to it. To add S/MIME capabilities, it splits the private
key in two, sends one to the service and encrypts the other under the
public key of the device. So the service only has a random number and the
device can't decrypt without the device helping.

The crypto is a scheme Matt Blaze and Torben Pederssen discovered.
Depending on how you look at it, you can consider this to be Proxy
Re-encryption or Distributed Key Generation. It is actually a simple case
that sits at the intersection of both.