Starting a discussion thread here as suggested by Lee Howard in the the
session yesterday.

My sense of the comments in the room (while I don't remember a hum) was
that while it was still too early to advance "IPv4 to Historic", there
would be value in sending a clear message that we are on a path to move
IPv4 to Historic at some point in the future.  This could either be through
sunset4 or preferably by an IETF governance statement.

Some topics for discussion seemed to be:

* Should there be clear guidance to IETF WGs and other standards bodies
that it is acceptable to develop IPv6-only protocols without consideration
for IPv4?   (With the caveat that practical deployments may still have a
need to inter-operate with IPv4, if only as providing IPv4-as-a-service
over IPv6.)  Are there particular efforts (eg, 5G) where we should be
encouraging an "IPv6-first" mindset as part of the design, and is this
something we can do?

* Do we set a time-table for moving to Historic or base this off some
adoption metric?  (80% global IPv6 adoption?  IPv6-only deployments
becoming commonplace in multiple scenarios outside of a few large
companies?)

* Are there other things we can do to reduce the time-window where everyone
has to deal with full dual-stack complexity?  (For example, fixing issues
that remain with some of the transition technologies that allow for
IPv6-only connectivity with IPv4-as-a-service?)

* What would be a good thing to call this: directionally historic,
end-of-engineering, "It's Complicated", to-be-deprecated, ...     (In
particular, we need to make sure the statement made sets expectations but
is is inline with present reality.)
_______________________________________________
sunset4 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sunset4

Reply via email to