Re: DNS under o=Internet

[[e-mail]]

> There has been a document on the tabnle for (.... long time) that has had
> the DNS tree under the root.    This issue was pointed out explicitly as
> an issue by Alan Young. 

Right. Except that last time around, the issue was raised in a sufficiently
'quiet' way that sufficiently few people responded so you didn't feel
that consensus had been achieved.

> I retained the location as is, because I did not
> hear much support for this view and so retained the status quo.   

I have no quarrel with your doing this.

> First, this position needs to be concensus.   "Does anyone ohject if I..."
> is not the right approach.   I STRONGLY object to the decision approach.

I *am* seeking consensus. Given the lack of response the last time
around, I thought I'd phrase it more strongly. It got you to
respond, didn't it? :-)

> Let me note the possible positions, with arguments for them.
> 1) At the root.   This was proposed, as it has a certain elegance.   The
> scheme is global, and so hanging it off the root makes sense.   However there
> is no registration authority for this.

That's exactly my point: the DNS is not global. The scope of applicability
of the DNS is only the Internet plus some other networks that have
'borrowed' the technology. The Internet is not the world (although it
is certainly a significant part of it), and to be truly global, the
Directory needs to extend beyond its boundaries.

Your second point about the lack of a registration authority is well taken.
My solution is to avoid the problem altogether by viewing the DIT
as a listing hierarchy :-).

> 2) Move to a location with registration.    This means picking a country, and
> registering a suitable point within this country, and then using this as
> the starting point.  This has the big advantage of being acceptable to
> general registration authorities.   

I am opposed to doing this as the DNS extends beyond the boundaries of
any single country.

> 3) Picking a subtree off the root.    This seems to lack the elegance of 1),
> without gaining the authority of 2).   I think that it is getting the worst
> of both worlds.  

Elegance is in the eye of the beholder. It seems far more elegant to me to
recognize that the DNS is Internet-specific for all intents and purposes,
and group it with other Internet-specific things under a subtree containing
all Internet-specific things.

> If you do this, I think that a sinlge arc should be defined,
> which leads to only Domain Names (e.g., CN=Domains). 

You can only do this (add arbitrary things to the root) only if
you view our current pilots as working towards an Internet Directory,
not a global Directory.

This goes back to the first point of my long posting regarding
naming: what are we trying to build here, a Directory for the Internet
or part of the Directory for the world?

An argument can be made that the Internet is both international
and having a globally recognized existence, thus deserving of a
listing/registration point below the root. The same cannot be
said for the DNS.

> This could be
> a second level under O=Internet.   I'd argue against this, as a) two levels
> is too much and b) some bits of the 822 world are going to object to this
> being under O=Internet.   I'd suggest having a single arc of the root.

I agree that two levels is too much. But zero levels (putting things
directly below the root) is too little. In recognition of the fact that
the Internet is part of the world, not the whole world, we have to
put Internet-specific things in an Internet-specific place and not
claim that the Internet is the world.

> My recommendation.   Go for 1), and leave RFC 1279 and OSI-DS 12
> alone!   We should definitely do this until an agreement is reached.  
> Lets discuss this in San Diego

I have no objection to going with the status quo until some sort of
agreement is reached. However I lean towards the opposite of what
Steve recommends: (1) is a solution that will shut us (the Internet
community) off from the rest of the world. I think being isolationist
is a bad idea, and we should move away from it.


Wengyik