North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

RE: Statements against new.net?

  • From: Mathew Butler
  • Date: Tue Mar 13 22:42:42 2001

Title: RE: Statements against new.net?

(My comments interspersed)

-----Original Message-----
From: Roeland Meyer [mailto:[email protected]]

>    To remain a global network, the Internet requires the
> existence of a
>    globally unique public name space.  The DNS name space is a
>    hierarchical name space derived from a single, globally
> unique root.
>    This is a technical constraint inherent in the design of the DNS.

False. It is not a constraint, it results from a restricted implementation.

--My Comment--
And how would you propose unrestricting this implementation?  The DNS design is ratified as an Internet Standard.
--End My Comment--

>That one root must be
> supported by a set
>    of coordinated root servers administered by a unique naming
>    authority.

This statement is a pure non-technical value-judgement, supporting the
legacy implementation, and is entirely self-serving. Of course, this
statement, itself, is also a value-judgement. However, note that the defeat
of the false premise, and it's direct improper conclusion, make this
statement a pure political phrase.

--My Comment--
The DNS SOA record specifies a single point of contact for all failures related to the zone, as well as the name of the true master of the zone.  Yes, it's possible for there to be an arbitration authority who is ultimately responsible for the root zone, but THERE IS A SINGLE AUTHORITY WHO IS RESPONSIBLE FOR THE ZONE.  Read that statement again.  Be it a corporation, a partnership, or an individual person (such as Jon Postel, may he rest in peace), there is ONE single authority.  There -must- be, as a purely operational issue -- just as there must be a single POC for a NOC for a backbone transit provider.  Or else nobody will use it, due to the issues involved in trying to coordinate between two or more authorities.  Miscommunication == Downtime.  Miscommunication Between Parts of the Authority == More Downtime.  Downtime is Unacceptable in the Root.

--End My Comment--

>    Put simply, deploying multiple public DNS roots would raise a very
>    strong possibility that users of different ISPs who click
> on the same
>    link on a web page could end up at different destinations, against
>    the will of the web page designers.

This entire paragraph is a NOP. Webpage designers would never link to
external resources under such condition as stated here. Those that do,
deserve to get appendages whacked.

--My Comment--
How would you propose to resolve this problem?  Without a SINGLE point of authority for the Root, there is no one person/organization/etc who is ultimately responsible for the shape or design.  Which means that there is Confusion.  From Confusion comes Chaos and Politics... in the old days, when it was primarily a technical decision, the Techs could deal with it.  Now, with Users and Politicians and Managers (collectively Chaosmongers), the Order (such as it is) of the Techs will not be heard.

--End My Comment--

You really should calm down. I use external programs to build db.root.zone
files, from external sources. All of the ORSC agrees to the core TLDs, which
includes the legacy roots, and there are dispute proceedures for handeling
collisions.

--My Comment--
That's fine, but ORSC doesn't attempt to hijack any TLDs that were already claimed and invested into by other entities.  You choose to use your own root -- this means that your "private" TLDs will not be visible to people outside of the ORSC's sphere of authority -- that's the risk and choice that you have made.  But as soon as your "private" TLDs are taken over by another corporation/entity that can make deals to hijack them in the public eye, you're probably going to be raising bloody hell.

Also, how many webpages actually refer to the ORSC's TLDs?
--End My Comment--

> RFC2826 SAYS YOU HAVE TO CHOOSE FOR YOURSELF.  Which is more important
> to *YOU*?  100% consistency with the rest of the world, or access to
> your private name space?  *YOU* evaluate, *YOU* choose, and RFC2826 is
> nice enough to point out the problems you'll encounter.

It is not at all an either/or situation. There *is* the path of
"accommodation".

--My Comment--
Indeed... but will other people accommodate -you-?
--End My Comment