Re: (ngtrans) Joint DNSEXT & NGTRANS summary

[Johan Ihren <johani@autonomica.se>]

"Tony Hain" <alh-ietf@tndh.net> writes:

> Nathan and Bernard have a good start to the summary, in 
> terms of boiling down the essential viewpoints. The things 
> I would add from the discussion include:
> 
> - It is clear that we know how to make A6 resolvers deal with 
>   AAAA clients.
> - It appears that with current tools A6 reduces the burden of 
>   renumbering on the zone administrator.
> - It is not clear that renumbering will occur faster than a 
>   AAAA zone file could be resigned.

> - It is clear that we could evolve from AAAA to A6 later if 
>   it became necessary.

I think that such evolution will quickly move from the realm of clear
possibilities to "things that simply will not happen" as the base of
resolvers that serve AAAA clients but doesn't do AAAA synthesis grows.

Today it would be easy, since the resolver base is small and mostly
based upon various versions of bind9 that will (all?) be upgraded to a
version with AAAA synthesis built in.

Tomorrow, with a large base of resolvers without that functionality,
anyone who attempts to move from AAAA to A6 will literaly saw off the
branch of the tree he is sitting on (from the perspective of AAAA
on-lookers on the ground).

On the other hand, I fully realize the issues with keeping a (mostly
unused) code-path of AAAA synthesis around in the resolvers with A6
not on the standards track. 

All I am saying is that I think any attempt att consensus through the
"we can always upgrade later" argument is bound to misfire.

Johan