Re: draft-ietf-v6ops-nat64-pb-statement-req-0: native connectivity withdual stack hosts

[Brian E Carpenter <brian.e.carpenter@gmail.com>]

Marcelo,

On 2008-07-26 03:47, marcelo bagnulo braun wrote:
> Hi
> 
> This is follow up on the previous message about dual stack being
> susceptible to modification or not.
> As i mentioned earlier, Dave and Jari have expressed that nodes must be
> able to use native connectivity, in order to avoid apps that are not nat
> compatible to crash.
> However, in our current set of requirements, we have
>   R3: Interaction with dual-stack hosts
> 
>   Translation mechanism MUST allow using native connectivity when it is
>   available.  This means that if a v6-only nodes wants to communicate
>   with a dual stack, it must use native v6 connectivity and if a v4-
>   only nodes wants to communicate with a dual stack, it must use native
>   v4 connectivity.(In this case, dual stack means a host with both IPv6
>   and IPv4 stacks, wich are both active, i.e. they have v4 and v6
>   connectivity).
> 
> 
> whcih in conjuction with R1:
> 
>   R1: Changes in the hosts
> 
>   The translation mechanism MUST NOT require changes in the v4-only
>   nodes to support the Basic requirements described in this section,
>   unless explicitly stated in the particular requirement.  The
>   translation mechanism MAY require changes to v6-only nodes.
> 
> results that IPv6 only nodes, (which in our document must be interpreted
> as a node without a v4 address irrespectively whether it has a v4 stack or
> not) MAY require modifications to satisfy this requirement.

It seems to me that a dual stack node with no (valid) IPv4
address is going to have to do something today, even in the
absence of nat64, so we can ask what existing stacks already
do in that situation - and once again, it's a good start if
nothing gets worse when nat64 exists.

> So, this means that unmodified ipv6 only hosts may not be able to use
> prefer the native connectivity and hence the apps that are not
> nat compatible would not work in these unmodified v6 only hosts.

And if so, we have not made things worse. And isn't this a defect
that stack maintainers will need to fix in any case?

> Do people are ok with this, or people think that we should be stronger
> on this and try that unmodified v6 only hosts can prefer native
> connectivity?

Sorry, I don't really understand what you are suggesting.

> 
> One more comment: Imposing this requirement for unmodified v4
> only hosts may be impossible to reach... comments?

You mean that the above MUST NOT may be impossible? Well,
then we can all go home ;-)

    Brian