Re: [v6ops] I-D Action: draft-colitti-v6ops-host-addr-availability-01.txt

[Philip Homburg <[e-mail]>]

In your letter dated Tue, 28 Jul 2015 14:50:04 +0200 you wrote:
>On Tue, Jul 28, 2015 at 02:48:07PM +0200, Philip Homburg wrote:
>> If you have more a 100 devices requesting DHCPv6-PD, then maybe it is time to
>> get a /48?
>
>For each single L2 network in a large enterprise?  Is that covered by RIR
>policies?
>
>Please do me the honour to actually *read* what I write.  This is not
>about "home", because home does not have scaling issues of the sort Andrew
>was talking about.

I don't see the problem. You can get a /48 no questions asked. This is 
equivalent to a /16 for IPv4. You can easily reach 10000 devices without nat on
IPv4 out of a /16. Why would it be different for /64s out of a /48?

Of course, the more devices you want in a /48 the more careful you addressing
plan has to be. But certainly for large flat structures it should be easy to
achieve high density.

I don't know if RIR policies allow for one /64 per device from PI. If not, then
maybe somebody should submit a policy proposal.