> From: "Joel M. Halpern" <[email protected]>
> O[n]e reason I like the registry/registrar architectural model fo[r]
> EID allocation is that it avoids end-site lockin with someone who might
> raise their annual maintenance fees.
This is another one of those issues that exist whether or not that EID-only
IPv6 block is allocated.
The real issue here is the way DDT (or, probably, any similar system) works.
If you're allocated (whether from the special EID-only block, or just regular
PI space, whatever) a block of EIDs, then for the mapping system, you're kind
of stuck using whoever it is who provides the mappings for that branch of the
delegation hierarchy.
This is similar to DNS, where you can buy .com names from a large number of
_registrars_, but if you don't like the way the _registry_ is handling
lookups, you're stuck.
You and I had discussed (offline) a way of allowing competing _registries_
in DDT (basically, all providers of node N have to share their mapping
databases), so it it _possible_ to have competing 'registries' in LISP,
but this has not yet been explored in detail.
I wrote 'registries' since, in general, the EID namespaces in LISP are
provided/controlled by _other_ entities - e.g., for IPv4, IANA and the RIRs.
So it's not a 'registry' service in the terms used in DNS - their 'registry'
service includes _both_ i) allocation, and ii) lookup, whereas DDT (and
whatever entites provide DDT service) only provides the latter.
Noel
_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp
