[Info-vax] implementing IPv6 on the internet

[David Froble]

Jan-Erik Soderholm wrote:
> Den 2016-09-24 kl. 00:05, skrev Dirk Munk:
>> Jan-Erik Soderholm wrote:
>>>
>>> The return package arrives at the WAN address (there is only one,
>>> don't really know what you mean with "WAN4").
>>
>> The IPv4 address of the WAN port
>>
> 
> Ah, OK then... :-) I see what you ment now... :-)
> 
> I thought it was meaningfull to have the NAT thing in the
> description since David seemed to think that it was based on
> having the internal address beeing sent to microsoft (in this
> example) and back again and then used for the routing of the
> reply package. It is not, it is managed by the NAT function
> and the internal address is never seen on the outside.

See, I learn something new every day.

First, I'm not arguing against IPv6.  I think it would be great.  But as stated 
by others, there is inertia for IPv4.  Expect that to affect some decisions.

That said, what else I am is a problem solver.  Show me a problem, and I'll look 
at it from all angles, come up with one or more potential solutions, then 
evaluate them to try to pick the best solution.

So now I'm being told there is a problem, and no solution.  Sorry, I have to (it 
just happens) think of potential solutions.

Jan-Erik mentions tables in the NAT router.  Memory is cheap.  A NAT router 
could keep a table of data from DNS look-ups, and be able to translate, using 
the table, an IPv4 address to the corresponding IPv6 address in the table. 
Could old entries fall off the list?  Sure.  Nothing is guaranteed.  But unless 
there is a really large amount of activity, it's my guess that most of the time 
the translation could happen.

That written, NAT can be a PITA ....