Re: IPv6 addresses for Exchange Points

[David Kessens <]

Mirjam,

Just for the record, I think that a few statements in your mail need a
correction.

On Mon, May 07, 2001 at 05:35:11PM +0200, Mirjam Kuehne wrote:
> 
> [Apologies for multiple postings. I suggest to keep the discussion on
> lir-wg@localhost as it is related to address allocation policies].
> 
> Dear colleagues,
> 
> At RIPE 39 last week in Bologna the issue of IPv6 address assignments
> to Internet Exchange Points was discussed (see also my mail from
> 24 April 2001 to these lists).
> 
> There was consensus to assign a /64 to an isolated Exchange Point.  It
> was further suggested to assign the agreed standard assignment size to
> a site (currently a /48) to a group of inter-connected Exchange
> Points.
> 
> An Internet Exchange Point was defined as follows:
> 
> 3 or more ASes and 3 or more separate entities attached to a LAN (the
> same infrastructure) for the purpose of peering and more are welcome
> to join.

I don't think both paragraphs are a good summary of our discussion.
There was no consensus on /64 assignments to isolated Exchange Points,
neither was there a consensus on which definition to use for an
(isolated) exchange point.

We had consensus that more discussion was needed, that you and Randy
would post a proposal for further discusion on the topic, and that
other proposals were certainly welcome too. I think that the
discussion on the lir-wg list clearly proofs that there was no
consensus yet. Your mail certainly is a nice starting point for this
discussion.

> The RIPE NCC proposes to proceed with assignments for Exchange Points
> under the above policy. 

While I would like to see a quick resolution to this problem, I don't
think the current proposal is detailed enough for people to make a
decision for or against the proposal.

We first need to answer a couple of more fundamental questions (in
order of importance):

1) are we going to do special purpose ipv6 allocations in the first place ?!?
2) what is the intended use of the address space for exchange points ?!?
3) how do we define an exchange point
4) what is the size of the actual allocation 
   [ this should actually be the easiest question to answer since we 
     are not as constrained as with ipv4 address space :-) ]

I have explicitly not answered those questions myself (though I
obviously have my own opinion :-)) - did I miss any other questions ?!?

And, yes we can do the further discussion on the lir-wg list.

David K.
---