Re: DNS heirarchy, multiple roots, etc [was Re: Split the IANA functions?]

[Phillip Hallam-Baker <[e-mail]>]

On Tue, Jan 7, 2014 at 5:59 PM, John C Klensin <[e-mail]> wrote:

>
>
> --On Tuesday, January 07, 2014 20:20 +0100 Patrik Fältström
> <[e-mail]> wrote:
>
> > On 7 jan 2014, at 19:48, Thomas Narten <[e-mail]>
> > wrote:
> >
> >> So any talk about having a different/better naming scheme is
> >> really just wishful thinking and a mostly a waste of
> >> everybody's time. If there was a better system, some set of
> >> smart folk would surely have already clued the rest of us in
> >> on what that was. Anyone remember Tim Bass?
> >
> > On top of that (if you look at some arguments that "change is
> > needed") I would like to know what the problem really is. And
> > yes, I am happy(?) to, and have tried to, understand what the
> > problem is.
> >
> > As others have explained, often "the problem" has to do with
> > misunderstanding of how Internet and DNS works. In other cases
> > it has to do with real issues, like the actual decision making
> > process for what strings can exist.
> >
> > When knowing what the problem is, there is often not much
> > disagreement on the fact improvements can be made. But more
> > disagreements in what the best path forward is. But THAT
> > discussion is much more fruitful and effective than just
> > saying "things must change".
>
> While agreeing with this and with most of Thomas's and Andrew's
> comments, I think there are a few substantive problems that can
> be easily identified and that keep coming back (I do not
> consider arguments about who controls the root of a strict
> hierarchy to be substantive, no matter how entertaining they
> become).   Most of them are associated with expectations of the
> DNS that it, as a strictly hierarchical system with one name per
> node, one-way links, and fairly weak aliases, cannot satisfy.
>

That is a much better argument.

What I was trying to object to is the use of 'mathematical possibility' as
a slapdown as if the design of the DNS were so perfect that anyone
proposing an alternative approach is a complete fool.

That sort of argument can work inside IETF but it looks really bad when it
is made in an external forum where the audience does not start from the
same assumptions as to what is immutable fact.


The choices are constrained by the legacy technical infrastructure and the
requirements.



> For example, we see repeated requests (or "requirements",
> demands, or fantasies) about treating two (or more) names as so
> identical that retrievals or actions that affect one of them
> affect the others.  It is certainly within the state of the art
> to do that in a strictly hierarchical  system.  Hierarchical
> database and file systems with support for those relationships
> have been around for close to 50 years if not longer.  But it
> isn't a modification that can be patched into the DNS -- it just
> isn't going to happen without replacing the DNS.  Some of the
> people who make those requests can be educated; others prefer to
> believe that, if only they repeat their demands often enough and
> loudly enough, they will get their way.
>

This is a better statement but the statement suggests that the problem is
only in the DNS infrastructure when the problem is also with the clients
that try to make lookups. For names to be useful, they have to work
predictably which means that at minimum they need to work the same with
every app on a given machine. Changes to the semantics of names will only
take effect after all the clients that perform recursive DNS resolution are
updated. And there are many of those now.

Yes, making two names look the same is also an example of the general cache
coherency problem, but Soviet science may believe it has an answer to that.



-- 
Website: http://hallambaker.com/