Re: [address-policy-wg] IPv6 assignment for the RIPE meetingnetwork

[Nick Hilliard nick@localhost]

Andrei Robachevsky wrote:
> I think the RIPE meeting network meets the requirement for multihoming,
> since it is multihomed, both topologically and in time.
> 
> But meeting the "Contractual requirements" is more difficult, since in a
> way that will require the RIPE NCC to have a contract with ourselves and
> to evaluate our own request. Perhaps a more elegant solution here would
> be the one proposed by Remco back in November (to establish a policy
> that lets the NCC file a request in the ordinary way).

Andrei,

Maybe you could explain what sort of problems you're having with address
space requests?  Is it that:

1.  the ripe ncc seems to have no official means of assigning address space
to itself,

1a. due to political neutrality requirements, the ripe ncc cannot engage
the services of a third party LIR,

2.  engaging in a contract with itself under the terms of 2007-01 will
cause a bureaucratic singularity to occur, and therefore it should not be
attempted,

3.  the registration department at ripe applies the same rules to the rest
of the ncc that they apply to everyone else, and because 2006-01 isn't
policy yet, the meeting organisation department cannot get ipv6 pi, or

4.  something else.

From your email of yesterday, it seems like #2 is definitely a candidate
problem.  I'd suspect #1a too, although not so much #1.

#3 will be solved with 2006-01. #1 and #2 can only be solved with a policy
change.  #1a could be solved by a policy change to create a pseudo LIR
(eu.ripencc), if there is a requirement to push the ripe ncc's numbering
requirements through a LIR.

Please explain more.  There's a lot of discussion going on about this
topic, but it's just not clear what problem you're trying to solve, other
than the immediate technical requirement to get ipv6 PI space for ripe
meetings.

Nick