PTR records for IPv6

Bjørn Mork bjorn at mork.no
Tue Sep 3 13:06:56 CEST 2013


Florian Lohoff <f at zz.de> writes:
> On Mon, Sep 02, 2013 at 09:30:22AM +1200, Brian E Carpenter wrote:
>> So, is there any real operational value in this, or is it just
>> a case of "we did it for v4 so it must be right for v6"?
>
> Its a nice to have IMHO. 
>
> Whats missing is an idea how to get forward and reverse DNS out to
> residential customers e.g. delegate a /56 prefix to its pre:fix::53
> address so a residential gateway/CPE/iad can generate forward
> and reverse DNS for its addresses handed out.
>
> I'd like to see this - We might have the opportunity to get rid of NAT
> for residential customers so we should enable to them to participate in
> the internet as a full member not some "client beeing able to use the
> clickable part of the internet"

I have played with the idea of implementing some sort of selv service
for the end users, allowing them to either have their prefix delegated
or to add a few simple *static* PTR records for specific addresses. I do
not see any sort of dynamic DNS solution as part of this.  The only real
use case AFAICS, is where the end user want to run a server with proper
reverse mapping.  That means that the end user should select a static
address first.  If they choose delegation, then they would of course be
free to implement their own dynamic DNS.

Simple to do, but realistically there isn't going to be much demand for
it. So it will probably just remain a thought.

I don't think blind delegation to pre:fix::53 is wise. You will end up
mith most of these being lame.  I am pretty sure I will end up with lots
of lame delegations too, if I let users choose to delegate without
verifying that the server they delegate to answers.

But you could problably probe a fixed address, like pre:fix::53, and
delegate iff there is a name server there, answering authoritatively for
the reverse of that prefix.  That would also be pretty simple to do.


Bjørn


More information about the ipv6-ops mailing list