[IPv6]Re: Working Group Last Call for <draft-ietf-6man-pio-pflag>

[Brian E Carpenter <[e-mail]>]

On 03-Jun-24 21:33, Eliot Lear wrote:
> Hi Lorenzo (and all),
> 
> On 03.06.2024 06:37, Lorenzo Colitti wrote:
>> FWIW, if we do want to change the subnet boundary from /64 to /80, one of the ways to do so might be to do that a decade in advance and tie it to one of the other prefixes (say, 3000::/4). After 10 years have passed, hopefully most of the implementations will be ready. But again, it's not clear that we actually need to do this now.
> 
> Without taking a position on whether /64->/80 is a good or a bad idea, while I like the idea of taking time to make a change of this nature, I would not be a fan of tying yet more semantics to IP addresses.  For one thing, what do you do if an implementation *doesn't* make the change in that time table and still lands in this new block?

It seems likely that if we did change (say) the IPv6-over-Ethernet spec to allow a choice between /64 and /80, it would be a site deployment option, so I'm not sure that Lorenzo's model would work anyway. But that said, I agree that the less semantics built into address bits, the better. So far, we've really only done that in f000::/4, I think.

     Brian