Re: [v6ops] [Lsr] New draft: carrying/renumbering assigned prefixes in IS-IS/OSPFv3 (lsr-v6ops-pd-aargh-00)

[Acee Lindem <[e-mail]>]

Hi David, 

> On Nov 5, 2023, at 8:55 AM, David 'equinox' Lamparter <[e-mail]> wrote:
> 
> On Sat, Nov 04, 2023 at 09:33:57PM -0400, Acee Lindem wrote:
>> I thought that DHCPv6 Prefix Delegation (PD) could have multiple
>> levels of delegation and that mechanism alone could be used for
>> assignment within a domain?  I must admit that it has been a very long
>> time since the home net WG discussions.
> 
> It can, but using it that way essentially forces you into picking an
> "upstream" interface on each router and building a spanning tree that
> the assignment travels down on.  It's just running the same protocol
> up- & downstream as server and client, not some actual "levels" feature
> in the protocol.
> 
> Unless I'm misremembering (yeah, long time...) there was consensus that
> hierarchical DHCPv6-PD is not desirable for homenet;  if a network
> warrants using OSPF or IS-IS, hierarchical PD is going to be even less
> desirable.

I agree that the partitioning of the PD prefix is not going to be correct in anything but a trivial network. If we were to distribute it in OSPFv3 we would need logic outside the protocol to partition the PD prefix as well. As we discussed offline, this could be configuration that statically assigns prefixes independent of the first 64 bits of the prefix received from the upstream providers. 

In any case, your next step to test the waters to see if users actually want this. 

Thanks,
Acee



> 
> 
> Cheers,
> -equi/David