[dns-operations] renesys blog: Identity Theft Hits the Root Name Servers

[David Conrad]

Kurtis,

On May 22, 2008, at 7:50 AM, Kurt Erik Lindqvist wrote:
> On 22 maj 2008, at 16.29, David Conrad wrote:
>> I'm suggesting that since it is so hard to change root server  
>> addresses, we remove the need to.
>
> What you are suggesting is that we do a global renumbering event  
> instead of the current trickle that is converging on a stable set  
> from what I understand. And your arguments with all the problems  
> associated with renumbering would probably be even worse in the  
> aftermath of a global renumbering.

The problem with continuing the slow trickle is that _today_, we can  
be reasonably certain that the folks who run the root operators are  
willing (out of the goodness of their hearts) to do the right thing.

In the future, I for one have far less certainty that this will be the  
case, particularly with the increased 'monetization' and  
politicization of the DNS.  This is only going to get worse.

>> Unfortunately, some root server operators see the idea of making it  
>> easier to disassociate the address with the organization providing  
>> root service and re-associate it with a different organization as a  
>> threat.  Perhaps this is understandable since it is much nicer to  
>> not have to be formally accountable to anyone, less of a burden to  
>> operate in non-transparent and non-open ways, more fun to have  
>> secret meetings, etc.
>
> I feel pretty targeted by the above remarks, but I am 'somewhat'  
> surprised.

I am, of course, not targeting you with those comments, but rather the  
existing root server system.  In a world in which national economies  
depend on the Internet which, in turn, depends on the DNS which, in  
turn, depends on the root servers and their operators, the lack of  
openness, transparency, and accountability, is simply insane.  As Ed  
said, it might have made sense back when Jon first anointed the  
original root servers, but that was a different Internet.

> I explicitly said that a change in accountability model might or  
> might not be desirable,

Laudable words.  However, I will note that for more than 10 _years_  
have passed since the need for some sort of formalization of  
relationships was raised and, to date, only ISC has deigned to enter  
into any sort of agreement (as toothless as it might be).  And note,  
this isn't to minimize what ISC has done -- I personally believe they  
should be congratulated for taking the fist step.

Of course, the lack of movement isn't particularly surprising -- there  
is no incentive for the root server operators to do otherwise.

> But I think it's important to point out that there are parts of the  
> world who would believe that the current distributed model is better  
> than a model where a US corporation operating under a DoC contract  
> could make unilateral decisions of who is to serve the root-zone,  
> less desirable as well.


Yes, yes.  Evil US government and all that.

There are other models of accountability.  While it might get me  
fired, I personally wouldn't have a problem if there were publicly  
binding SLAs with penalties for non-performance entered into with  
entities other than ICANN (in fact, this has certain advantages).

And this says nothing about openness or transparency.

Regards,
-drc

P.S. And for those who have accused me privately of driving an ICANN  
agenda, these are personal opinions: I argued this back when I was  
part of the root server cabal, after I turned in my secret decoder  
ring, and will continue to argue it after I get fired.