ICANN/DNSO
DNSO Mailling lists archives

[ga-full]


<<< Chronological Index >>>    <<< Thread Index >>>

Re: [ga] Alternative roots




On 23 Apr 2001, at 17:23, John Charles Broomfield wrote:
<snip>

> Yes we do, no we don't... Generalities are such fun aren't they?
> "We don't want collisions" sounds quite nice as a battle cry and something
> that everyone would want to stand behind, but what EXACTLY is a collision?
> A purist might say that a collision is when ANYWHERE in the world two people
> typing the same thing could potentially get different answers. This Is a
> BAD THING (tm) some would say...

Ah, if it were just that simple, I'd say let's see who does what.  
Unfortunately, the potential for real problems is a much larger issue 
than some private remote network offering up a collider.  This is the 
USG root which has a huge market share even though it is being eroded 
a bit.  It involves email, ftp, hostnames...  And who is to say that the 
"alternatives" may not be so alternative after a while?  I don't see it as 
trivial, but that's just me.
> 
> ICANN (or anyone operating ANY bunch of root-servers -legacy or not-) has to
> decide what exactly it is going to consider legitimate enough to warrant
> respect. No, this is NOT "higher than thou", it is just common sense. *ALL*
> of the alternative roots out there have their own rules as to what they
> consider legitimate or not (to an extent). Just blindly saying that ANY
> previously set up TLD *ANYWHERE* in the world is to be respected is just not
> good enough.

I don't think that is what is being said, but again, it's my view of the 
issue.  As for respect, yes, any existing entity should be respected.  If 
a TLD is accessible to the public (not an intranet will not routables) and 
there are live domains out there, it should be respected.  Your business 
is just as important as mine and if you have a business product it 
should not be usurped by a government agent simply because they 
want it.  In addition, if it will cause a technical mess, the agent (ICANN) 
is mandated to avoid it.  

What does it take to earn respect?  Must there be a certain amount of 
money or assets? $50,000?  $150,000? $1,000,000?  Customers?  How 
many is enough?  100?  1000?  100,000?  If 5,000 individuals and 
businesses will be disrupted, is that enough to warrant notice?  You 
said that it is not a higher than thou attitude, but is common sense.  Is 
it common sense to deliberately duplicate a TLD knowing that it will 
disrupt those businesses, email, hostnames...  And what kind of 

 What stops everyone in the world in that case just setting up
> half a dozen TLDs on their own home machine? (no, I won't go down the line
> of argument of "why shouldn't they?", we're talking about expansion of what
> exists, not creation of a totally different beast, go elswhere if you want
> that argument).

> 
> ICANN, *according to its own rules* is not condoning, promulgating or creating
> any type of collision whatsoever... It starts of from a point where it takes
> care of only its own namespace, and as such will not have two entities
> running the same TLD within its own servers. The rules it setup to decide
> who is legitimate to run ADDITIONS to its server, you know all about ($50K
> registration etc...).
> 

And that's the fallacy.  "Their" name space.  A collision occurs when 
there is a duplication in any level of the tree, regardless of which set of 
servers you're talking about.  If the public is accessing those servers, 
there is a problem.  It's one thing if they can or cannot reach an 
address, but quite another if they reach an erroneous address.  Oh 
well... We'll see, huh?

> The alternative root servers have each and every one of them their own rules
> to define collisions and to define what is or not worthy of respect. Just
> saying as a blanket statement that the alternative root servers do not have
> collisions is a falacy. 

No one has said that. It has now been greatly exacerbated because the 
message is out that it's okay to introduce them.   However there has 
been a concerted effort to resolve those collisions.  

Just in case, I remind people that to prove a
> point, about 2-3 years ago -maybe less, not too sure-, I publicly claimed on
> a few lists [dns-policy comes to mind amongst others] a stake to ALL
> up-to-then non-used anywhere in the world alphanumeric combinations of 3-10
> (maybe 20 can't remember) characters. So, ANY additions since then obviously
> collide with my personal namespace (nameservers running on my own network,
> but for obvious reasons the addresses won't be divulged).
> 

If they are not public, what is the problem?  If they are public, it could 
be a problem.  

> Of course, they will all claim that my nameservers and my namespace don't
> count as far as their own rules to how they manage their servers. Of course,
> they are right, but at the same time they fail to see that ICANN can decide
> what is worthy of respect and what isn't, what merits a claim to
> non-collision and what doesn't. In other words, as far as ICANN is
> concerned, there is NO collision. As far as OTHER peoples namespaces go,
> that is their own problem...

So much for generalities.  So, according to your argument, if the so-
called alternative roots must work within their name space and ICANN 
must work within its name space and ICANN is okay with duplicating 
any TLD they choose in their name space, it is by the same token okay 
for any alternative root to do the same, right?  So ORSC, if it chose to 
do so, could then have their version of .com and not worry about the 
collision, right?  After all, it's their rules and they can do as they see fit 
in their name space.  I don't think so.  

So while we are trying to get everyone to the table to have reasonable 
discussions, it's really okay to say, "no, we don't have to" and be 
perfectly correct in saying so.  However, what will be accomplished?  
The DNS will be a mess and everyone will go his own way.  ICANN 
rules the USG root and that is all that matters in the world.  Hmmm...

> If we were talking about IP address space instead of DNS namespace, the
> problm would be akin to a bunch of people unilaterally deciding to use in a
> private manner a bunch of IP addresses that nobody up to then had used (but
> not using the addresses set aside as private address space), and then
> suddenly yelling like crazy once they find out that the addresses they had
> been using for a while (in an unsanctioned manner) were assigned to some
> other entity. Nothing stops them from continuing to use those addresses as
> they wish, but its not something that I would recommend. (I see this type of
> scenario every day by the way, speaks a lot about who sets up their
> network).
> 
> Yours, John Broomfield.
> --
> This message was passed to you via the ga-full@dnso.org list.
> Send mail to majordomo@dnso.org to unsubscribe
> ("unsubscribe ga-full" in the body of the message).
> Archives at http://www.dnso.org/archives.html
> 


--
This message was passed to you via the ga-full@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga-full" in the body of the message).
Archives at http://www.dnso.org/archives.html



<<< Chronological Index >>>    <<< Thread Index >>>