[Info-vax] implementing IPv6 on the internet

[Dirk Munk]

David Froble wrote:
> Dirk Munk wrote:
>> Chris wrote:
>>> On 09/23/16 21:22, Dirk Munk wrote:
>>>
>>>>
>>>> And now we get a problem. How is the router suppose to know it has to
>>>> build a new IPv6 packet with the same payload, but destination address
>>>> X6, and sending address WAN6?
>>>
>>> The router will already have it's own V6 address, supplied by the isp.
>>> It will also know about the V4 addresses on it's own private subnet
>>> and how to work NAT for those addresses.
>>
>> Great, and how does the router know the IPv6 address of Microsoft, or
>> even better, how does the router know it has to send the packet to the
>> IPv6 address of Microsoft?
>>
>> The PC did a nslookup remember, the router didn't. The router doesn't
>> know it has to exchange the IPv4 address of microsoft by the IPv6
>> address, how could it? If the PC sends something to an IPv4 address,
>> how can the router know if this destination also has an IPv6 address?
>
> In your own words:
>
> 2. it will get the IPv4 address X4 (and the IPv6 address X6) of
> microsoft.com
>
> So, the PC, or whatever, knows both addresses.  If the PC sent out the
> packet with both addresses, then the router could decide which it wants
> to use.  WOuld be a change in the PC software, but, that happens every day.
>

Sure, that is thinkable. However that is not "a change in the PC 
software", it is a change in the IP stack. You just proposed a whole new 
set up for IP packets. Given the fact that it took some odd 20 years or 
so to get where we are with IPv6 today, implementing such a proposal 
could take at least 10 years.

And why should we want it? What could be the reason to keep an old 
fashioned protocol like IPv4 running on our LAN?

And what to do with internet servers that are IPv6 only (yes, these do 
exist already)? There's now way you can reach them.