Re: A common problem with SLAAC in "renumbering" scenarios

[Nick Hilliard <[e-mail]>]

Ole Troan wrote on 19/02/2019 13:08:
> It’s been the IPv6 addressing model for at least 20 years, so I
> think the other areas have had ample time to provide their input.

dynamic-assignment addressing protocols tend to assign dynamically.  As 
Barbara notes, the stability period may be long but there is no 
guarantee that things won't change, especially in the ipv6 world (the 
caida study refers almost exclusively to ipv4).  Addressing tends to be 
sticky, not static.

>>> There’s nothing a bit of regulation can’t fix. :-)
>> 
>> I'm sure you didn't mean to invite regulatory enforcement :-)
> 
> So far the ISP industry has managed to stay largely self-regulated. 
> Let’s hope it can continue that way.

Not inviting regulation would be a good start in this department.

> But that doesn’t mean the IETF should accommodate “operators cannot
> provide service with fixed addresses”.

We're not talking about that though - all (or nearly all) operators can 
provide static IP addressing and factor this into their design models 
and businesses to one degree or other.  What you seem to be talking 
about is whether operators should default to exclusively using static 
prefix assignments.  This is unrealistic at best.

Nick