Re: [v6ops] I-D Action: draft-ietf-v6ops-unique-ipv6-prefix-per-host-07.txt

[Simon Hobson <[e-mail]>]

Lorenzo Colitti <[e-mail]> wrote:


> /64 gives you something that no longer prefix does: the ability to run SLAAC and connect unlimited devices behind the host. 

From previous discussions, I am led to believe that SLAAC will work with other (longer or shorter) prefixes. It is only the deprecated EUI64 method that *needs* 64 bits.


> No. Someone wrote code that gets a prefix via DHCPv6 PD, and blindly announces it in an RA.

Are you suggesting that someone who realises that not everything is a /64 would do that ? You've not really done anything to counter this being a case of "everything is /64" thinking and thus ignoring the potential for anything else. It may be a simple case of blindly re-announcing the prefix you got delegated - but it still comes down to not understanding that the prefix may be other than /64.


> That doesn't work because SLAAC only works with 64-bit IIDs, and for good reason.

As above, I've seen comments here stating that this isn't the case.


> Already? 64 everywhere has been the standard for 20 almost years now. It predates pretty much every IPv6 network and every IPv6 implementation out there today.

And that hardcodes in a restriction which IN THE REAL WORLD is known to cause issues. For example, you (incorrectly) only get a single /64 from upstream, but want to something other than a single host or very simple network. I agree that it shouldn't happen, but in the real world it does - you either stick with a simple basic network, or you run prefixes smaller than a /64.