Re: New Version Notification for draft-koodli-ipv6-in-mobile-networks-02

[Rémi Després <remi.despres@free.fr>]

Le 19 avr. 2010 à 19:47, Fred Baker a écrit :

> 
> On Apr 19, 2010, at 10:30 AM, Cameron Byrne wrote:
> 
>>> - If DNS64-NAT64 synthesizes IPv6 addresses that hosts don't distinguish today from real IPv6 addresses, this destroys the IPv6 guarantee of e2e transparency.
> 
> Only for communications that leave the IPv6 domain and are translated to IPv4, which already has this problem.

This problem is indeed inherent to NAT64.
But fortunately it is not inherent to the more general problem of IPv4 connectivity across IPv6-only access network:
- If hosts are dual stack, and have means to encapsulate IPv4 in IPv6 across provider access networks, no IPv6-IPv4 translation is necessary.
- This is possible in particular with DS-lite, where operator NATs are NAT44s.
- This is also possible with the reversed 6rd I plan to document and which is essentially as simple as 6rd (encapsulations with stateless address mappings). With it, customer sites can derive from their IPv6 addresses their IPv4 addresses (typically shared, with disjoint port sets). Where applicable, no operator NAT at all is needed.


> If the communication stays within the IPv6 domain, whether the two hosts know it or not, they are communicating transparently.

Right, but, unless it is impossible, e2e transparency in IPv6 should also apply across the Internet in general. 
(And AFAIKT this is NOT impossible).

Regards,
RD