On Wed, Feb 20, 2013 at 6:14 PM, sofiane Imadali <sofiane.imadali.i...@gmail.com> wrote: > Hi, > Thanks for your interest Roger, your question is very pertinent and I > wanted to share it with the group (hope you don't mind bringing this > offlist mail to the list).
no I don't mind, I'll comment on your answer inline <snip> > - Today in a closed operator domain, the vehicle (Mobile Router) > can generate a ULA prefix for site-scoped communications. It is > subject to collisions if used on a large number of vehicles (10^4 or > more - RFC4193). In the vehicular context you're facing this > limitation. > For applications such as remote diagnosis, or other services > (involving access to in-vehicle Machines), the solution we propose > allows to create a predictable (deterministic manner) unique > (collision-free) prefix, that in a closed loop (operator domain, > manufacturer domain) allows you to determine which vehicle to access > with which prefix: No DNS, No Prefix Delegation, No DHCPv6, nor any > other additional artefact involved. > Again, with this proposal, if you are an operator/manufacturer with a > list of your vehicles fleet, you can determine a list of ULA prefixes > to access each one of them: No additional protocol involved, no > collisions. I don't think it's clear enough in the draft that we're talking about closed circuit system, not communicated initiated from within the car to outside, any outside device including other cars. The case of accessing the car's network from a service point of view are still inside the car's domain. Am I understanding this correct? Another issue is the use of the term site, you could explain how you are using that term. I personal would consider any inner car communication to be within a site, including when I connect my phone to the on-board media-system to stream music or movie. Any communication toward outside the car are out of the site. <snip> > To sum up; in order not to trigger (too soon) the privacy concerns and > be blocked by that, the primal use case (domain, site) would be a > closed operator network, or car manufacturer in the manufacture > (closed loop). This method allows us to set internal routing to access > the in-vehicle devices, by a deterministic manner: VIN --> ULA prefix. > No additional protocol is needed to do that. Any service provided by > this devices (monitoring vehicle state for example) is a use case that > applies to this addressing approach. as said above, I understand those two draft to talk about inner car communication, not communication toward _any_ device outside the car. That include communication to other nearby cars, parking automate and more. Or to sum it up as I see it with regard to the privacy issue - the issue is when the car, or someone in the car use an IP toward _any_ outside device that can be traced back to the VIN of any car (vehicles). > Thanks for reading and feedback; no problem, good to get such a good explanation back :) -- Roger Jorgensen | ROJO9-RIPE rog...@gmail.com | - IPv6 is The Key! http://www.jorgensen.no | ro...@jorgensen.no -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
