Re: [Captive-portals] A final check on draft-ietf-capport-architecture-09

[Erik Kline <[e-mail]>]

On Tue, Sep 1, 2020 at 9:49 AM David Dolson <[e-mail]> wrote:

> How do such devices obtain IP addresses?
>
>
In IPv4, this would be NAT.  The first host to connect would probably pass
the captive portal for all other devices, but only because of the
HTTP-intercept technique.  No clients would see the API URL, and they would
never be able to learn the venue URL

In IPv6 it's more complicated, and largely not yet addressed.  With Proxy
ND, the downstream clients would see the RA/DHCPv6 option(s).  With
64share, the same is possible but would, I suspect, be
implementation-dependent.

Arguably the domain of Captive Portal solution is limited to the case when
> the agent assigning IP addresses is controlling access to the network as
> well.
>
> -Dave
>

As we've currently scoped things, I agree.

The more I think about it, the use case I had in mind (capport
implementation in an ISP's modem/CPE) would probably require things we've
not yet completed.