Who is in charge of the 2002::/16 reverse DNS ?

[Jeroen Massar]

Michel Py [mailto:michel@arneill-py.sacramento.ca.us] wrote:

> > Jeroen Massar wrote:
> > Also 6to4 is a _transistion_ method... so I wonder why
> > somebody wants to get their company/organisation relying
> > on something that will go away in the long run, which could
> > take it's time ofcourse.
> 
> The time frame we are looking at here is ten years or more, 
> so it is perfectly legitimate to want reverse lookup for 6to4 
> addresses for the time being.
"which could take it's time ofcourse." and which unfortunaly will be in
the range 10 year period you mention :(
Ofcourse it's perfectly 'legit' to have reverse lookup, it's even quite
overseeable most of the time as IPv4 ranges
can be whoised and are known where the end up at.

> > You'd still be better off getting some real IPv6 space from
> > your upstreams.
> 
> Yes, but sometimes, it is not available. Most of the time, 
> these days. It is sad to say, bit in terms of performance 
> 6to4 is better than regular tunnels and will be used because of this.
That's why it's one of the available transition methods.
But IMHO *any* transition method should be avoided as much as possible.
If it ain't possible too bad, use your favourite and possible transition
method, but if you can get a native uplink... why not ?

> > And renumbering a network with IPv6 is a piece of cake
> 
> This is simply not true. In the 6bone meeting in Minneapolis, 
> the renumbering of /24 and /28 pTLAs was discussed, and 
> several people contributed that renumbering anything bigger 
> that a dial-up connection is not a piece of cake.

Ofcourse if one sizes down renumbering isn't easy...
But if you move from one /48 to another /48 it should be a piece of
cake.
The AMS-IX did it perfectly well last weeks.
First the routers on where in 3FFE:3000::/64 (out of the /48) now they
have moved to 2001:07F8:1::/64 (out of the /48).

eg: 3ffe:3000::a500:8954:1 became 2001:7f8:1::a500:8954:1

And one can do that with almost a flip of the switch.
It ofcourse comes down to a good numberplan and if one needs to scale
down (/24 to /28)
the numberplan shouldn't be completely filled up ofcourse, though
sometimes that is quite
unavoidable and indeed one will have a hard time sizing down.
At least in theory it all should be much easier as IPv4 though.

Don't know if there is something like this but maybe it's a good idea to
create a document setting
up some of the currently known scenario's which explain which problems
can occur when renumbering.
And ofcourse include a solution which explains how to avoid these
problems.

I only wanted to state the "if possible use native IPv6, not one of the
transistional methods" ;)

Greets,
 Jeroen