RE: [GLOBAL-V6]IPv6 Allocation Policy

["Michel Py" <michel@arneill-py.sacramento.ca.us>]

Gert,


>> Michel Py wrote:
>> /35s are not supposed to exist anyway. As far as I know,
>> all /35s have been extended to /32s without the need to
>> renumber; just extend the prefix. 

> Gert Doering wrote:
> Some clarification: a few LIRs have actually NOT extended
> their prefix (for reasons unknown), so there are still
> "official" /35s around.

I read this as they simply have not filled the paperwork, because if
they did the extension would have been automatically granted; is it what
you meant?


> What *is* the current state of affairs concerning
> multihoming solutions?

- Short-term the only thing early adopters can do is to obtain one or
multiple /32s as discussed earlier.

- For medium to large setups, MHAP is entering commercial development; I
cloaked the technicalities for a few months to give a competitive
tactical edge to the company that develops. No intent to create a
proprietary protocol though; when the code is ready to ship I will
publish the latest draft.

- For smaller setups Microsoft is working a multi-address solution along
the lines of what Christian Huitema proposed in Atlanta:
http://arneill-py.sacramento.ca.us/ipv6mh/Dual%20Homing%20Experiment.ppt
http://arneill-py.sacramento.ca.us/ipv6mh/HostCentricMulti6.ppt
I do not know what the current state of development is though.


> I have had quite some argument with a colleague yesterday about
> "whatever they are argueing about, the end customers will not
> care until it makes multihoming easy for 'large enterprise' type
> customers". While I don't like that - he's right, people that
> can have multihoming today are quite reluctant to give that up
> for some vague promises why IPv6 is better for them.

I have made that very point myself many times. For that reason and for
the sake of launching IPv6 I am willing to look the other way if we
assign /32s to end-sites (which is a blatant violation of today's
policies, but there is no other solution) as long as we keep the policy
tight and can enforce it later.


>> There are several proposals to provide PI-like addresses
>> that are moving forward though.

> Now that's another interesting sentence :-) - "end users"
> would *love* that (and it might turn out to be "the" incentive
> to go to IPv6).
> Do you have a pointer for me where I can read up on those
> proposals?

http://www.tndh.net/~tony/ietf/ipv6piaddressusage-04.txt
http://www.tndh.net/~tony/ietf/ipv6piaddressformat-04.txt
http://www.ietf.org/internet-drafts/draft-py-multi6-gapi-00.txt
http://arneill-py.sacramento.ca.us/ipv6mh/geov6.txt

It is clear to me from recent meetings and presentations that the issues
that Craig was raising here recently are not only Cisco's problem and
that there needs to be a PI-like solution as well.

Michel.