Re: [Softwires] Host based translation: v4-v6

[Zhen Cao <[e-mail]>]

On Thu, Dec 3, 2009 at 10:25 PM, Simon Perreault
<[e-mail]> wrote:
> Zhen Cao wrote, on 2009-12-03 09:10:
>>> New "IPv6 apps" are usually not IPv6-only. They are version-independent. See
>>> e.g. RFC4038. So your future app will try IPv4 if it cannot get IPv6
>>> connectivity. Which, it seems to me, would make case 4 fold into case 2.
>>
>> Hi Simon, I checked 4038 and found it is informational. And I do not
>> think applications developers will indeed follow this.
>
> The fact that it is information is irrelevant. This is how applications are
> being developed *right now*. We have experience with this, having ported many
> applications to IPv6. This method is taught in seminars, books, etc. There is no
> speculation here.

Good. You are saying that this happens for ported applications, but
that's only part of the story. We will have many new IPv6 applications
in the future.

>
>> For example,
>> RFC4294 defines IPsec as a MUST for IPv6 node, but from our equipment
>> test, we found this is a myth.
>
> How is IPsec relevant to an application?

I referred to this to prove that applications developers do not go
ahead as documented.

>
> As far as I know, the IPsec API is still in early stages:
> http://tools.ietf.org/html/draft-ietf-btns-c-api-04
> http://tools.ietf.org/html/draft-mglt-btns-ipsec-api-requirements-00
>
> Simon
> --
> DNS64 open-source   --> http://ecdysis.viagenie.ca
> STUN/TURN server    --> http://numb.viagenie.ca
> vCard 4.0           --> http://www.vcarddav.org
>