[address-policy-wg] 2013-03 New Draft Document and Impact Analysis Published (No Need - Post-Depletion Reality Adjustment and Clean up)

* Malcolm Hutty

> Firstly, thank you for being so willing to engage with a view to 
> meeting objections, rather than just defending your proposal. I hope 
> my answer meets you in the same spirit.

Yep, and thank you for being constructive as well!

> That amendment would *definitely* make the proposal more appealing to
> me. To my mind, this moves us from "2013-03 looks bad" to "2013-03 
> seems broadly OK but needs thought to see whether a little further 
> wordsmithing is needed". Which is, I think you'll agree, major 
> progress.

Absolutely!

> How long do we reasonably have to think about that wording? I realise
> we're at a late stage, but I would welcome introducing more eyeballs
> to this problem.

Since we're still discussing, the review period will be extended with
another couple of weeks, so there is some time. Hopefully we've have a
text we're happy with by then, which will then as I understand it be
sent to the NCC which will have to re-do the Impact Analysis. The first
IA took 11 weeks to complete, so I *really* hope that we only have to
ask them to re-do it once.

(As you probably can imagine my first preference would be not to do it
at all, as I consider myself a pragmatist and I think these changes have
minimal *practical* impact, but oh well...)

> In particular, I'm not sure I like limiting fairness to End Users, as
> you have done.

Actually, those words aren't mine, they are there in the current policy
document (ripe-592) as well. This was 100% intentional, as I would
strongly prefer for the proposal to not change previously established
consensus on any topic that is not "core" to the proposal itself, or to
introduce any new lofty principles/goals/ambitions.

Nick's cautionary "a hemina of wine" example is highly relevant here. If
we try to define *new* principles, we'll probably need further
wordsmithing, which may in turn lead to bikeshedding and thus grinding
the progress of the proposal to a standstill. Another risk is that even
though we find agreement on a new text, the NCC may say in its IA that
they interpret this to mean X, while our intention was Y - and we'll
have to start over again.

So while I have no desire whatsoever to stifle any discussion on any new
goals and principles and any tools or mechanisms that would enforce
those, I would strongly prefer if this discussion could be done in
another policy proposal independent of 2013-03.

With that in mind - can you live with simply retaining the "fairness
goal" as stated in our current policy document and nothing more? I'm not
necessarily asking for your *support* of the proposal here (although
that would of course be very welcome too!); "I no longer have any
objections" or "I abstain" is sufficient.

(BTW: I don't think it is a problem limiting fairness to the End Users.
The whole point of the Internet Registries is to get the resources into
the hands of the End Users. Internet Registries aren't using the
resources themselves, so they have no need for "fairness". Of course,
often a single organisation will be both an Internet Registry and an End
User, but in that case it will be covered by the fairness goal anyway.)

> Referring back to your other comments on fairness, to the effect
> that "fairness" has already been abandoned in the last /8 for
> NCC->LIR, I don't necessarily agree or think it should be presented
> like that.

That's fortunately just my personal and pragmatic opinion, and not
something I have any intention of putting into the policy. :-)

> I suspect if we adopted the text above without alteration the effect
>  would be to still require a documented need for at least a single IP
>  address. We could further refine it to remove even that - but it has
>  already been mentioned that this is a negligible requirement; how 
> would you feel about leaving it in, given that it's so little? Taken 
> together with the bit in the second set of square brackets, this 
> policy would exclude those that apply simply in order to re-sell,
> but nobody else. I would think that a simple checkbox "Do you intend
> to use at least one of these IP addresses on a network you operate?"
>  should satisfy the documentation requirement.

(Filiz: I understood that this is essentially the same issue you
described in your last post to the list. Therefore I'm CC-ing you in
here rather than replying to your message directly, hope that's OK.)

I have no strong feelings against this, no, the amount of red tape is
minimal and in any case something an LIR will have to deal with at most
*once*. The main reason I have resisted this is that in practical terms
it is nearly meaningless as a barrier against those who would abuse it,
and asking the NCC to re-do the IA just to get it in there would IMHO be
a waste of time and effort. However, if we need to change the proposal
and re-do the IA anyway, might as well do this one too.

So here's my suggested text. In the proposed section 5.1
(http://www.ripe.net/ripe/docs/other-documents/draft-v2-ipv4-address-allocation-and-assignment-policies-for-the-ripe-ncc-service-region-new-policy-text/#allocations-made-by-the-ripe-ncc-to-lirs)
we add a new point:

«4. In order to qualify for receiving the allocation, the LIR must state
an intent to use it for making assignments to one or more End Users.»

I think this describes the current status quo pretty well, and should
result in a check-box like the one you described in the request form
(something I'll try to confirm with Andrea *before* the whole updated
proposal is sent back to the NCC). Would this be sufficient to remove
your concerns?

Best regards,
Tore Anderson