[arin-ppml] IPv6 in the Economist

[Lee Dilkie]

Dean,

I'm unsure why you think IPv6 is broken simply because the DNS 
infrastructure isn't completely rolled out yet or is designed in a 
fashion not to your liking.

Dean Anderson wrote:
> 1.  You could claim that there are actually IPv6 root servers. 
> [that would be something objectively determined.]
>   
This will take some time, but is hardly holding up deployment.
> 2.  You could claim that IPv4 and IPv6 DNS records aren't mixed. [that
> could also be objectively determined.]
>   
I see no reason why there is a problem mixing IPv4 and IPv6 addresses in 
DNS. DNS is just a database, it contains a lot of other non ip address 
records like SPF and DKIM, TXT.. So putting both IPv4 and IPv6 addresses 
into the same nameserver is quite reasonable. How you access the 
nameserver, either via IPv4, IPv6 or X.25 even, isn't the question.
> 3.  You could claim that one must remove IPv4 NS records for IPv6 NS
> records. [that could also be objectively determined]
>   
huh? I don't understand your point.

The fact is. IPv4 isn't going away for a very long time. The *only* 
reasonable rollout involves dual stack existing for a long time. 
Separating transport from content (be it web, dns records, anything) and 
making it accessible from both IPv4 and v6 is going to be the norm, for 
a long time.

Owing to the nature of dual stack, I don't think anyone is in a hurry to 
roll out end-to-end IPv6-only deployments. Our first goal should be to 
get IPv6 networks and routing rolled out to our IPv4 customers so 
dual-stack will even have a chance of working.

-lee
>
>