Re: [IPv6] RFC 6724 shouldn't prefer partial reachability over reachability

[Nick Buraglio <[e-mail]>]

On Sun, Nov 19, 2023 at 3:33 AM Philipp S. Tiesel <[e-mail]>
wrote:

> Hi,
>
> On 17. Nov 2023, at 15:25, Michael Richardson <[e-mail]>
> wrote:
>
>
> Ole Troan <[e-mail]> wrote:
>
> To work well, we need some kind of system-wide _give me a socket to "name"_
> daemon, which does all the getaddrinfo() work, knows the list of
> (per-provider-domain) network addresses, and can cache all of that info.
>
>
> I thought this idea had lots of promise:
>
>
> https://datatracker.ietf.org/doc/html/draft-ubillos-name-based-sockets-03
>
>
> That has some on-the-wire implications, as I read it.
> There is also this effort:..damn... it was presented at RIPE last year,
> but I
> can't find it.
>
>
> All these things are done in TAPS… with the difference that it is
> deployable one-sided and already has one widely deployed implementation
> from Apple:
> https://datatracker.ietf.org/doc/draft-ietf-taps-arch/
> https://datatracker.ietf.org/doc/draft-ietf-taps-interface/
>
> Yes, it needs a log of nifty details to get right including Rule 5.5,
> using src-dst-routing and keeping DNS separate per PD/Interface but may
> also solve the multi-homing without NAT.
>

The more I think about it, the more I am thinking that it may be a
deployable solution for multihoming, with all of the details in appropriate
stages of hand waving.


>
> It not that it is so hard to do, it's that it takes some perserverance to
> get
> it upstreamed into all sorts of desktop OSs, embedded OSs, and to patch the
> API into the dozens of tools that need it.
>
> I think it's a 7-10 year duration effort, requiring 1/3 FTE.
>
>
> Right… but it takes a lot of people to agree that it should be done to
> begin with,
>
> AVE!
>    Philipp
> --
> Philipp S. Tiesel
> https://philipp.tiesel.net/
>
>
>
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [e-mail]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
>