On 07/26/2010 18:33 GMT+02:00, William Herrin wrote:
On Mon, Jul 26, 2010 at 3:24 AM, RJ Atkinson<[email protected]> wrote:
I agree with Tony Li and others who believe that there is a
fundamental tussle between anonymity and reachability.
Then the smart move is to represent the tussel in the protocol so that
it can be decided at run time by the user rather than by us in the
protocol design.
I missed Ran's message, but found it in the archives ...
Yes, the end user must be empowered. IMHO this isn't a tussle at all in
the real world. Government policy and laws, if nothing else, will
require that people not be required to reveal even their general
location in order to communicate at all, particularly if others can find
out their location and track them without their knowledge. If nothing
else, knowing someone was at a medical facility, or knowing someone is
on vacation so you can rob them. Or: "I know where you were last night,
and unless you pay me I'm going to tell your wife/husband", where the
victim can't even determine if the claim is true! Regardless of what we
might think technically, not being able to cloak your location at all
will not work in the real world. Using a data plane rendezvous point
(HIP proxy, MIP HA, etc.) allows the ultimate user or representative to
decide how much to reveal and how much not to. This has to be the
engineered default.
This could be used with ILNP. ILNP doesn't require the optimizations
that have been proposed.
BTW this isn't really about IP routing and forwarding, which are to get
IP packets between endpoints specified at the IP layer. This is about
how those IP endpoints are determined, and one of the endpoints might
not be an endpoint for the higher layer, just for IP. As an example see
the SMTP server mesh.
And looking at Ran's mail ... I get packages delivered to me all the
time without revealing even the domain I'm in. They get sent to my
department, where they are "encapsulated" (new label stuck on the
outside) and shipped to me. We "pay for shipping" twice, but for
correspondents that I rarely interact with, and only for short times,
this is less effort than going through the overhead of optimizing the path.
Scott
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
