Re: [Fwd: [wg-c] straw poll -- reminder]

["Ross Wm. Rader" <ross@ebarn.com>]

Personally, I don't question the technical stability of DNS with regards to
adding massive number of new TLDs to the root.

I do question the political, procedural and organization stability of DNS
with regards to adding a massive number of new TLDs to the root.

-RWR
----- Original Message -----
From: <pg@name-space.com>
To: Ann-Catherine Andersson <aca@telia.net>
Cc: <wg-c@dnso.org>
Sent: Wednesday, August 18, 1999 5:16 PM
Subject: Re: [Fwd: [wg-c] straw poll -- reminder]


>
> Many people have made statements similar to the following:
>
> "I would chose Option 1, as this seems to be an *operationally safer* way
to
> start."
>
> (emphasis added)
>
> Many, if not most of the people who make this statement are doing so
> because they don't know the facts first-hand, not because they have the
> day to day experience of operating a DNS service or otherwise have
> the experience and operational knowledge needed to make such
> determinations based on knowledge.
>
> I present this information to this working group in order to help them
> make a more informed choice with regards to adding new TLDs to the root:
>
> THERE IS NO TECHNICAL OR STABILITY ISSUE
> IN ADDING NEW TLDS TO THE ROOT.
>
> The number of domains, be they TLDs or SLDs added to a given zone
> in any period of time does not in itself effect "operational safety" or
> "stability" of the DNS.  Any perception of such is a MYTH.
>
> In a given day, there are somewhere between 3 to 8 MILLION second-
> level domains (SLDs) added to the TLD zone for "COM."
>
> Structurally, every zone file in the DNS is identical, from the ROOT.ZONE
> on down the tree.  DNS is HIGHLY SCALABLE, as evidenced by the
> daily growth of "COM."
>
> Adding ANY NUMBER of TLDs to the root WILL NOT adversely effect
> operations of the DNS any more than the addition of MILLIONS of SLDs
> in "COM.", a daily occurrance, effects operational stability of the DNS.
>
> If there are 1000 new TLDs added to the root at once, and not increased
> for a year's time, the impact would still be less than ONE DAY'S GROWTH
> IN "COM."
>
> The servers which currently operate the ROOT.ZONE, also carry the
> TLD zones for "COM."  "ORG." "NET." "EDU." "ARPA." "MIL." "GOV."
> as well as the "ROOT.ZONE" and the DNS functions reliably,
> although this is not a "technically pure" implementation of the
> DNS hierarchy (see below).
>
> Each day, the growth of the SLDs in the "COM." zone is in the MILLIONS
> with NO ADVERSE OPERATIONAL EFFECTS ON THE DNS.
>
> The load on the ROOT-SERVERS, even with ONE MILLION TLDS
> would be LESS than the load when those same servers run the
> "COM." TLD, if the TLD-SERVERS were *properly* delegated and
> separated from the ROOT-SERVERS. It is possible, however,
> that the current systems can likely operate a "MILLION-TLD-ROOT"
> in addition to running the "COM."  "ORG." "NET." "EDU." "ARPA." "MIL."
> and "GOV." TLD zones with todays resources.
>
> BOTTOM LINE:  Adding large numbers of TLDs to the ROOT will not
> adversely effect the operations or stability of the DNS.
>
> [Note:  In a telephone conversation last August with Dr. Paul Mockapetris,
>  the computer scientist who INVENTED DNS, I discussed the above scenario.
>  Dr. Mockapetris agreed that every zone in the DNS is EQUALLY SCALABLE,
>  FROM THE ROOT ON DOWN.
>
>  He also stated that the practice of running the TLD zones mentioned above
>  on the ROOT-SERVERS is not a technically pure implementation of the
>  hierarchical DNS structure.  The TLD zones should run on servers other
>  than the ROOT, and the ROOT-SERVERS should run ONLY THE ROOT.ZONE
>  file, with the TLD zones delegated to separate machines. This not only
>  reduces the load on the ROOT-SERVERS, but it also allows plenty of
>  resources for the TLD NAMESPACE to GROW.]
>
> Respectfully submitted,
>
> Paul Garrin
> Founder/CEO
> Name.Space, Inc.
> http://name-space.com
>
>
>
>
>
>
>
>