Re: Extending a /64

[[e-mail]]

>> Right. Take the following solution:
>> 
>> - there is no subnet prefix / on-link prefix.  - hosts are assigned
>> single addresses via DHCPv6
>> 
>> The whole concept of interface-id's then become quite abstract.
> 
> Assuming current deployments with shared access links, then in my experience,
> a router needs to know the subnet prefix. 

It's easier for the purpose of a undigressed thread of logic if you accept the premise. ;-)
It's uncommon that data-links are multi access anymore.
For the purpose of the argument let's assume routing on each port.
Ref: draft-troan-6man-p2p-ethernet-00
Even if you actually had a shared network, you obviously don't need a subnet prefix.
That's equivalent to L=0.

>> And you could argue that the concept of interface-id's only exist
>> within the constraints of subnet-prefixes and SLAAC.
> 
> So I'd argue no, on a shared access link you have, explictly or implictly a
> subnet prefix and therefor interface ID's

But you agree that interface ID's cease to exist if there is no subnet prefix?

>> Is it a requirement that solutions support SLAAC (or current SLAAC)?
>> I would argue no.
> 
> In the past some people have argued quite strongly against single addresses
> via DHCPv6. 
> 
> The lack of addresses problem associated with IA_NA could be solved with IA_PD.
> 
> However, if we say that we solve the problem with some combination of IA_NA
> and IA_PD, then we have come full circle and solve problem that comes from a
> lack of PD with more PD.
> 
> That may still work if the lack of PD is at an administrative boundary and
> then within the client network we solve the lack of address problem with
> IA_PD and IA_NA.

DHCP IA or PD is not a solution to general permissionless extensions of a link.
They might be building blocks.

E.g. you could use HNCP to divvy up a /64 into DHCP pools that HNCP assigns to the participating routers.
Where no links has a subnet prefix (apart from fe80::/10 obviously). And hosts gets addresses assigned via DHCP.
Nodes that further wants to extend the network does that by participating in HNCP.

You could imagine doing this with a DHCP "God" box (and relays or discovery) too. Regardless you would need some sort of routing, static or dynamic.

The building blocks are already all there and you can extend a /64 network, without changing or breaking the IPv6 addressing architecture.

Best regards,
Ole