On Fri, Aug 1, 2014 at 11:16 AM, Joe Touch <[email protected]> wrote:
> If the sole goal is anti-monitoring, we have what we need.

See the Opportunistic Security draft.

> I also fear that a TCP-based solution might present a misleading view that a
> session is E2E protected when a MITM acts as a TCPINC proxy; TLS already has
> well-established procedures by which users can examine a certificate and
> determine whether to trust it. Building that into an OS and integrating its
> management with the user application (esp. legacy apps) isn't going to be as
> intuitive.

I agree.  I see this effort as part of the OS (opportunistic security)
effort: you get protection from passive attackers that you'd not have
gotten otherwise, but no protection from MITMs without doing a bit of
extra work.

Now, the "extra work" to get protection from MITMs can be relatively
minor: in a DNSSEC/DANE world it's just a matter of discovering the
server's public key with DNSSEC and then using it to authenticate the
server's side of the key exchange.  API-wise there's several ways to
do this, and some would be really easy to use, but let's not go into
it.  The point is that we have a path to make things better all
around, and that's what OS is all about.

Nico
--

_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc

Reply via email to