Re: [Arcing] A bit more on the problem statement

[Edward Lewis <[e-mail]>]

On 2/20/16, 13:24, "Arcing on behalf of Douglas Otis"
<[e-mail] on behalf of [e-mail]> wrote:

>This is likely my last comment.

Apparently something said touched a nerve.  If it's frustration over
repeatedly making a point, I'll use this old reply of mine: It might be
the 10th time you've said "it" but this is the first time I've heard "it."
 If there's something else, I'll just state that I haven't been closely
following the homenet wg in the past few years, nor the dnssd wg for that
matter, so perhaps there's a boatload of context buried in mail archives.

>Don't insist adoption of draft-cheshire-homenet-dot-home-02
>must follow some grander scheme that might make use of an
>alternate root or URL modification. Any such approach
>doubtlessly precludes essential compatibilities needing
>consideration for Homenet deployment.

Due to what I read as exasperation, I looked at previous messages on the
mail list and see you've mentioned this draft in each of your posts.  So,
I looked it up and gave a quick scan.

Before I'd comment on it, I would need to read it more thoroughly and I'm
not sure this mail list is the appropriate place to make comments.
(Perhaps it is, it's not a draft-ietf-homenet-...)  You mention "Don't
insist adoption of" - what does that mean?  E.g., Adoption as a WG item in
homenet, adoption as an IETF standard, adoption by the ARCING mail list,
or something else?

Still, the use case raised seems rather significant, obviously akin to the
IPv4 address ranges that by convention are not advertised to the open
Internet (inter-enterprise is the term used in "Address Allocation for
Private Internets").  The trouble is that unlike routers, DNS elements
don't have an innate notion of "internal" and "external" interfaces.
(There are de jure definitions of "views", multiple implementations of the
concept but no protocol standards-track document has ever been produced.)
So, we can't straight up look to how IPv4 addressing and routing for a
solution to the use case covered in the draft cited.

Beyond that I do see that this use case is interesting to bring into any
discussion on the nature of Domain Names or a general namespace.  The
string "home" could qualify as a resolution switch.  Like "onion" the
process for resolving is different from traditional means.  Unlike onion,
the process for "home" uses the protocol already spoken over port 53
(i.e., the DNS protocol).  I can see this as fertile ground for work.