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

[Brian E Carpenter <[e-mail]>]

On 25/09/2017 09:45, Lee Howard wrote:
> Personal opinion...
...
> Dual stack doesn't solve the problem IPv6 was created to solve, which is the lack of available IPv4 addresses. 

That made me sit up. In a sense it's true, but the reason for dual stack
as the original co-existence mechanism was so that *servers* could serve
both legacy (IPv4) and current (IPv6) clients throughout the transition
period. So dual stack, all the way from the server to to the client's
provider edge, was very much part of solving the lack of IPv4 addresses.

This became inadequate because too many ISPs and hosting providers were
too slow to support IPv6. That's why we're now in translation hell.

On 25/09/2017 23:27, JORDI PALET MARTINEZ wrote:

> Maybe you’re not there, but operators live in a world where they don’t have any more IPv4 addresses, so double-stack is over in the access networks. 

That isn't true everywhere. There are still plenty of places where consumer ISPs
have IPv4 addresses for new customers (one per customer, of course). Of course,
this won't be true for much longer. Therefore:

> Maybe you’re not there, but operators live in a world where they don’t have any more IPv4 addresses, so double-stack is over in the access networks. However, there are devices and apps (in both cellular/tethering and non-cellular), that are IPv4-only, so user networks/LANs, need some transition support to make IPv4 transparently available end-to-end.

Yes, sadly. And in that situation, stating that 464XLAT is BCP seems quite
reasonable. Sad but true.

    Brian