Re: draft-savola-multi6-asn-pi-01.txt

[Brian E Carpenter <brc@zurich.ibm.com>]

marcelo bagnulo wrote:
> 
> > Isn't there an issue about needing to support N global locators (used
> > for external communication) plus at least one local locator?
> 
> I may be not following, but my understanding is that Brian is proposing to
> use a non routable address as identifier, so that it can be assigned
> independently from the topology.

Well, I wasn't really *proposing* anything. But indeed the identifier-address
would still need to be globally unique, but it wouldn't need to be
topologically assigned. But I think Erik would like to make it impossible
to forge, as well, and that is indeed a challenge.

I think we are drrifting well away from Pekka's draft here.

   Brian

> This location independent identifier would be have a global meaning, not
> local, as i understand it.
> The problem is that it cannot be used as a locator. IMHO you could store it
> in the DNS, as long as you make sure that it is not used as a locator.
> Perhaps a new RR could be defined to store this id.
> 
> regards, marcelo
> 
> > In the case of NOID, would it make sense to store the local
> > locator together
> > with the global locators in the (global) DNS?
> >
> > If you can't put them in the same lookup service, then you need some
> > mechanism to (securely) discover locators that are not in the lookup
> > service. (As an aside, if you have such a mechanism you could also use
> > it to share care-of addresses with your peer when being mobile.)
> >
> > But I haven't found a way to do this with acceptable security in a scheme
> > like NOID. I was hoping that the hash chains in WIMP could be
> > used for this
> > in NOID, but I think Jari convinced me that this has problems.
> >
> > It can be done securely when there is a new namespace as in HIP, SIM, etc.
> >
> >   Erik
> >
> >
> >

-- 
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Brian E Carpenter 
Distinguished Engineer, Internet Standards & Technology, IBM