[ipv6hackers] opportunistic encryption in IPv6

[Eugen Leitl]

On Mon, Jun 10, 2013 at 03:12:59PM -0700, Owen DeLong wrote:
> Because hardly anyone has the PKI that would be necessary to make it actually work?

A lightweight, low-security setup would exchange public keys (short ones, ECC)
during session setup. You will need need active (MITM) attacks to
disrupt this, and it is detectable in principle. 

Keys could be ephemeral (one for each session) or cached, and fingerprints
checked. Changed fingerprints could be silently logged, or session setup
denied.

Keys could be stored in a DHT, according with a trust metric which e.g.
uses social graph information for validity.