Re: Confirmation to advance: draft-ietf-6man-ipv6only-flag-05

[Sander Steffann <[e-mail]>]

Hi,

> Actually their isn’t a alternative solution.  There are several things trying to be achieved.

This reads as a list of goals invented to justify the solution...

> 1) remove ALL IPv4 traffic from the network.

Why?

> 2) stopping hosts attempting to send IPv4 traffic onto the network.

Considering the noise they cause I agree that this is a worthwhile goal.

> 3) being able to turn off IPv4 on the router interface.

Why?

> 4) needs to work on a BYOD network.

Agree.

> 5) eventually being able to turn off ALL IPv4 packet processing in the router.

Definitely.

> If you have a solution that achieves all these goals without signalling via IPv6 I’d like to see it.  Filtering can’t achieve 2.  RFC2563 can’t achieve 1 and 5.

A "hard" ipv6only flag would achieve that but doesn't have consensus. A "soft" flag used as a heuristic could get consensus but doesn't achieve all those goals either.

While many devices still have IPv4 turned on by default RFC2563 can keep the noise down (2). When IPv4 becomes less and less common (3) and (5) are fine. Optionally with filtering on ethertype as well if you really want (1).

There have been many protocols on Ethernet that have disappeared all by themselves, possibly enforced by filtering. I'm still not convinced we need anything more than what we already have.

Cheers,
Sander