[dns-operations] A new (bad) idea for DNS: Distributed DNS Implementation in IpV6

Michael Sinatra michael at rancid.berkeley.edu
Fri Dec 7 22:23:57 UTC 2007

Rodney Joffe wrote:
> hey Niall,
> On Dec 7, 2007, at 8:57 AM, Niall O'Reilly wrote:
>> On 7 Dec 2007, at 14:12, Rodney Joffe wrote:
>>> I'm sorry, I think I may be misunderstanding. How does the size of  
>>> the
>>> IPv6 address space affect the Root DNS ?
>> 	Checking the archives of dnsop at ietf.org may help avoid duplication
>> 	here on dns-operations ...
> I appreciate the suggestion. But the subject was introduced on this  
> list. And much as I try, I cannot understand why adding IPv6 will make  
> much difference to the existing 3,000 records in the root zone. And  
> given the fact that we're talking about address space and not name  
> space, delegation may mean there is increased load down the tree. But  
> not much at the top, which is where Lutz is pointing (the "Root").

I agree with the sentiment on the IETF dnsop list that:

o this is not a dnsop WG issue
o the i-d author is improperly confusing DNS operations, address space, 
demand for DNS names, and naming semantics
o it is another case of the proponents of DHTs and similar techniques 
attempting to apply the DHT hammer to a screw
o the proposed solution does not follow logically from the alleged "problem"
o there's very little evidence that the alleged "problem" exists

Given the "logical" premise of the i-d--that the large address space of 
IPv6 will create a huge demand for names, I am actually more concerned 
that we will run out of words to use as DNS names.  I'd suggest that the 
author write an internet-draft proposing that new letters or characters 
be created in each of the major alphabets and syllabaries, so that we 
have more permutations to create more words.  We will then easily be 
able to stave off word exhaustion without having to resort to hacks such 
as k3wlsp33k.


More information about the dns-operations mailing list