Re: [v6ops] The need for local-ipv4 socket transition solutions -- NAT64/DNS64 remains insufficient

[James Woodyatt <[e-mail]>]

On Fri, Apr 3, 2015 at 7:29 PM, Lorenzo Colitti <[e-mail]> wrote:

> If so - then how do you see us getting to IPv6-only? Do you see
> application developers moving to support IPv6 by themselves, including on
> legacy apps that haven't been updated for years, or that can't be updated
> because they were sold on CDs? Or do you see network operators removing
> IPv4 regardless of whether that breaks their users' applications?
>

I don't see us getting to IPv6-only unless and until network operators are
compelled by some external force to start breaking IPv4-only applications
for their users. Yes, I understand that no such external force exists
(thanks in no small part to regulatory capture, but we can leave that aside
for now) at present. They will only do so when the costs of supporting
IPv4-only exceeds the cost of breaking IPv4-only. And that's probably just
as it should be, unless you're the sort who thinks that Internet service
operators ought to be treated like public utilities instead of private
enterprises.

I would support V6OPS publishing a BCP that recommends, to those network
operators which cannot abide breaking IPv4-only applications for their
users, that subscribers be provided with native dual-stack IPv4/IPv6
service, instead of them trying to make a hard transition to IPv6-only,
which relies on them somehow forcing customers to bring their own hosts
that can support 464XLAT to the network.

I would probably support 6MAN publishing a matching pair of Informational
specifications that update RFC 3493 and RFC 3542 with considerations for
hosts with a CLAT function integrated.


-- 
james woodyatt <[e-mail]>
Nest Labs, Communications Engineering