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

[Michael Richardson <[e-mail]>]

Brian E Carpenter <[e-mail]> wrote:
    > One aspect of this is that there are two WGs involved, because we need
    > to consider the impact of draft-ietf-v6ops-dhcp-pd-per-device and
    > draft-ietf-6man-pio-pflag together.

Well, yes you are right.

If you feel strongly about something being experimental, then it should be
v6ops-dhcp-pd-per-device, not pio-pflag, which I think should just be
standards track.

    > halves the IPv6 address length and divides the available address space
    > by a factor of 18,​446,​744,​073,​709,​551,​616 (2**64)?

I have thought about this.
I'm not concerned for a few reasons.
a) we are actually connecting multiple hosts (even if they are virtual
   containers) to the network.  Plus tethering.  This is exactly why we made
   the v6 space so big.

b) It very much argues why ISPs need to be delegating more than a /64 to
   every network!

c) We learnt twenty years ago that a critical issue in v4 expansion wasn't
   how much address space was left, but how many routes the DFZ could handle.
   In v6, a critical resource is Neighbour Cache Entries, and pd-per-device
   reduces this significantly compared to proxy-ND.

    > I think it would be appropriate for 6man+v6ops to have a unified
    > discussion on all this, to avoid it coming up anyway during IETF Last
    > Call.

isn't that what we are doing now?
It's all the same/correct people.

I also wonder if we will need a DHCPv6-PD option that says, "I just need
internal enterprise connectivity, a ULA/64 would be fine"

--
Michael Richardson <[e-mail]>   . o O ( IPv6 IøT consulting )
           Sandelman Software Works Inc, Ottawa and Worldwide