Re: [v6ops] reclassify 464XLAT as standard instead of info - double stack coexistence

[Alexandre Petrescu <[e-mail]>]

Le 25/09/2017 à 12:12, Mikael Abrahamsson a écrit :
[...]
> It doesn't matter if it's GTP over IPv4 or IPv6. The end device
> doesn't care.

Well it should care.

It should care about not breaking the apps that are IPv4, not breaking
DHCPv6-PD.  And dont try to be smart with funny translation methods;
because others tried too.

Dumb network.

> The important thing is what type of PDP context that is supported.

For IPv4 apps, the PDP context type should include IPv4 (in addition to
IPv6).  That's about it.

> If it's only IPv6 type PDP supoorted (or used), then this is "IPv6
> only access".

Well, that's "IPv6-only PDP type", it's not "IPv6-only access".

The Access Network is a huge concept that I am afraid no-one has seen an
IPv6-only version of.

> That's all the device sees. What the ISP uses to produce this service
> is irrelevant in this discussion.
> 
>> The ideal is "double stack" - no encap, no translation, no IN A in
>>  UDPv6 packets, no IN AAAA in UDPv4 packets, and PDP type v4v6.
> 
> You're pretty much alone in that definition.

Hmm... and how about the many networks that provide simultaneous IPv4
and IPv6 access without any translation, encapsulation nor DNS-v4-v6?

If you sum them up I think you end with more than a single mobile
operator's subscribers.

Look at those stats of IPv6 google http hits - how many come from a
single mobile operator's subscribers?  All the rest are from "double
stack" networks.

Alex

>