Re: [dtn] Marking RFC5050 as Obsolete?

Hi,

>From an implementation point of view it would make sense to obsolete
RFC5050 as it would make it clear that new implementation shall be done
following the BPv7 specification. It would send a clear signal that the
working group has decided to no longer support this protocol and will
instead focus its effort into developing BPv7.

Regarding the analogy with IPv4/IPv6, when IPv6 came out there were already
hundreds of millions of IPv4 users interconnected to the same network.
Today most of the existing BPv6 networks are not really "public" nor
interconnected to one another so obsoleting the RFC doesn't prevent the
administrators to simply ignore this status update and continue operating
their legacy network as usual.

Regards,
Lucien

On Thu, Sep 26, 2019 at 3:39 PM Carsten Bormann <cabo@tzi.org> wrote:

> On Sep 25, 2019, at 21:52, Templin (US), Fred L <Fred.L.Templin@boeing.com>
> wrote:
> >
> > It is the same with BPv6 and BPv7 - there is a non-negligible deployment
> of BPv6
> > that will still continue after BPv7 is published whether we say
> "obsoletes" or not.
> > There is operational experience with BPv6 that will continue onwards the
> same
> > as happened with IPv4, and that is not a bad thing.
>
> There will always be protocols in real world use that have been replaced
> by newer ones.
>
> The question here is one of expressing intent.  Is BPv7 intended to
> supersede BPv6 or not?
>
> I’m not talking about “deployment realities” here (heck, I still have some
> Python v2 on my system), I’m talking about intent going forward.  Either
> the intent is to sustain both versions indefinitely (with bug fixes and
> extensions still going into BPv6), or the intent is to move to BPv7.  Like
> with Python v3, which ultimately needed a strong statement (and even a
> deadline) that it is now time to stop using Python v2 (and even then, the
> Python v2 is not going to vanish from my systems magically, and there will
> likely be some people hacking v2 and keeping it alive even beyond
> 2020-01-01).
> (We don’t need a deadline here, but we need to be clear about the intent.)
>
> I don’t think IPv4 vs. IPv6 is a good analogy here, but if you have a
> massive infrastructure processing BPv6 bundles, it may seem to be that way
> to you.
>
> Grüße, Carsten
>
> _______________________________________________
> dtn mailing list
> dtn@ietf.org
> https://www.ietf.org/mailman/listinfo/dtn
>