Re: [homenet] Homenet Naming Architecture

[Douglas Otis <[e-mail]>]

On 1/19/16 5:05 PM, Suzanne Woolf wrote:
> Hi,
> 
> On Jan 18, 2016, at 7:14 PM, Douglas Otis 
> <[e-mail]> wrote:
> 
>> RECOMMENDATION 1: The TLDs .corp, .home, and .mail be 
>> referred to the Internet Engineering Task Force (IETF)
>>  for potential RFC 1918-like protection/treatment.
> 
> There was an individual draft proposing these names be 
> reserved in the special use names registry that was 
> extensively discussed in DNSOP last year. The WG did not
>  decide to advance it.
>> 
>> Moving forward, the role for .home and .corp TLDs with 
>> respect at establishing local naming conventions needs
>> to be clarified before meaningful headway can be made.
>> Can anyone offer meaningful guidance on this point?
> 
> I'm not sure I understand the claim you're making here--
>  maybe I missed a great deal of progress on the homenet 
> naming architecture, but it seems to me that there's 
> plenty of work to be done in deciding what behavior we 
> want from names before we start worrying about which 
> specific strings to use. (One of the key questions to me,
> if in fact we end up choosing domain names for this 
> purpose, is what value is actually added by using a 
> human-friendly string at all in the homenet context; 
> YMMV.)
> 
> In regards to the specific names you mention, it's my 
> personal opinion that we wouldn't be doing anyone a 
> service by trying to use them in homenet, both because 
> they're contended within the ICANN policy space (as I 
> think Andrew already pointed out) and because if we take
>  the risk of collision with a global delegation of those
>  names seriously, we should also take seriously the 
> possibility that they're being used in ways that could 
> lead to collisions in homenets.

Dear Suzanne,

Names rather than addresses, especially for IPv6, is well
ensconced as a means to establish user friendly network
configuration access. As such, special names greatly aid
network development, such as 'localhost' and '.local'.

Homenet is attempting to instantiate multiple router
configurations that function independent of external
services.  While names like 'localhost' and '.local' have
greatly aided network development, these names are limited
to immediate hosts or multicast. For this new environment, a
name able to establish user friendly configuration access is
needed.  DNSSD WG has proposed and supports '.home' for this
purpose.

Lacking standards body leadership or guidance, equipment
providers tend to use arbitrary names or hard-coded
addresses to bootstrap even single router configurations.
Homenet needs to standardize how to manage the configuration
of multiple routers.

As in the past, use of specialized names offered the least
onerous solution. As '.home' has been adopted in various
products, a special use names registry that includes at
least this name is both prudent and vital at aiding local
multiple router development.

Methods to further instantiate names within this specialized
environment is then better able to proceed and avoid
interfering with existing services. Names included within
this service might be derived from existing protocol hints
and less likely though explicit user interaction. In any
case, configuration must not depend on external services.

Regards,
Douglas Otis