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

["Dale W. Carder" <[e-mail]>]

Thus spake Bjoern A. Zeeb ([e-mail]) on Tue, May 07, 2019 at 06:33:50PM +0000:
> On 7 May 2019, at 18:21, Philip Homburg wrote:
> 
> > But nobody is doing any work on what IPv4 should look like in the
> > future.
> 
> You mean inside the IETF given sunset4 was concluded without conclusion?
> 
> 
> What does SNA, Frame Relay, FDDI, Token Ring, … ok bad examples … what does
> IPX/SPX, Appletalk, VINES IP, XNS, NetBIOS, .. look like today?  Should we
> assume IPv4 will have a different fate?

I had worked on migrations of IPX, Appletalk, NetBIOS, and DECNET to IPv4
in a large academic network.  We can debate if you want to call that
a "managed" environment! ;-)

The lesson to me was that they were all migrated in the same manner, by 
removing the services that were making the advertisements causing the
clients to use it.  Slowly as the fileservers, printers, and so on were
turned off the clients became irrelevant.  I recall the SAP broadcast
load at its peak as decidedly non-trivial.

If you want to draw an analogy to IPv4, this would be when the 'A' records 
disappear from DNS and nothing is advertised via mDNS.  There was not a 
"don't use IPX flag" in IPv4 for example, though I don't recall if
netware, apple, or windows did anything special at the application layer.
By the time we got all this turned off, the OS vendors certainly had us 
beat as these protocol stacks were not on by default.

I don't see how IPv4 would be much different.  Yes, we could provide a
hint flags, but removing 'A' records from DNS and whatever the equivalent 
is in mDNS would be more robust.  That leaves with just the IPv4 clients
chirping, which is either mitigated by ethertype filtering in managed
environments, or otherwise negligible in volume in unmanaged networks.

Dale