Re: [Rift] IPv4 vs IPv6

good, let's remove the sentence but we have to resolve the following
scenario now.

v6 LIEs are exchanged, with that we have v6 gateways

now, if v4 addresses are not present, no v4 LIEs can be exchanged but a
node may be able to forward v4 over v6 gateways (in fact highly desirable).
How do we know whether v4 can or cannot be forwarded over the link?

A possible solution is to send on the adjacency which address families can
be forwarded like this

/* Link capabilities */

struct LinkCapabilities {
    /* indicates that the link's `local ID` can be used as its BFD
       discriminator and the link is supporting BFD */
    1: optional bool                           bfd =
            common.bfd_default;

*    2: optional bool                           v4_forwarding_capable = true;*}

which will allow a node to decide whether it installs the v4 route v6
nexthop with v6 gateways on this link if it does not have a v4 address. Or
we just say "v4 can be forwarded over RIFT links even if only v6 LIEs have
been exchanged?" ...

-- tony

On Fri, Jul 19, 2019 at 12:30 AM Bruno Rijsman <brunorijsman@gmail.com>
wrote:

> That would be my vote.
>
> AFAIK the protocol supports v4-only, v6-only, and v4-and-v6.
>
> I don’t see a need to put a requirement in the spec that a router must
> support a specific subset of those options.  That is up to the implementor
> / operator.
>
> — Bruno
>
> On Jul 19, 2019, at 4:45 AM, Tony Przygienda <tonysietf@gmail.com> wrote:
>
> If the consensus is that we should be dropping the sentence completely in
> the next versions and not specify any RIFT router requirements, I think
> that's perfectly reasonable as well.
>
> --- tony
>
> On Thu, Jul 18, 2019 at 4:39 PM Alvaro Retana <aretana.ietf@gmail.com>
> wrote:
>
>> On July 18, 2019 at 3:50:34 PM, Tony Przygienda (tonysietf@gmail.com)
>> wrote:
>>
>> [Speaking as a WG member.]
>>
>> Aha ok, what you say is reasonable. I was steered here originally by the
>> fact the even cheapest stuff today
>> <http://airmail.calendar/2019-07-18%2012:00:00%20GMT-3> does v4 and if
>> we have v6 support on the fabric v4 forwarding over v6 nexthops is possible
>> at basically zero config cost and you have a great v4 to v6 migration
>> story. but with the angle you propose I see that writing "SHOULD support
>> v4" would be better assuming silicon will become v6 only over years and v4
>> will vanish.
>>
>> We can't write of course "MUST support v6" since that makes the stuff
>> undeployable in foreseable future albeit we all wish the world would have
>> moved on for last 20 years :-}
>>
>> Other opinions?
>>
>> I would go all the way and not make specific support Normative.  Even a
>> “SHOULD support IPv4” could raise a lot of questions.
>>
>> My 2c.
>>
>> Alvaro.
>>
>> _______________________________________________
> RIFT mailing list
> RIFT@ietf.org
> https://www.ietf.org/mailman/listinfo/rift
>
>
>