[arin-ppml] 2014-1 Out of Region Use

Milton L Mueller mueller at syr.edu
Wed Feb 25 16:43:18 EST 2015


Jon
Yes, it's clear that you support the intent of 2014-1, which is to reconcile actual staff practice with an approved policy. But it also sounds like you think our threshold requirement can  be gamed?  The "speaking Romanian" issue (or more likely, speaking Chinese) has been the subject of fierce debates. Do you have specific suggestions for how we could fix this? 

--MM

> -----Original Message-----
> From: Jon Lewis [mailto:jlewis at lewis.org]
> Sent: Wednesday, February 25, 2015 1:10 PM
> To: Martin Hannigan
> Cc: Milton L Mueller; arin-ppml at arin.net
> Subject: Re: [arin-ppml] 2014-1 Out of Region Use
> 
> On Wed, 22 Oct 2014, Martin Hannigan wrote:
> 
> >>> resources stating that this is a no op as well:  already using
> >>> numbers in other regions and even ARIN (Curran) chimed in and said
> >>> that it wasn't a problem.
> >>
> >> I think you're interpretation of the situation is WAY out of line
> >> with the reality. Staff wants to STOP out of region use, and is doing
> >> so de facto because of the ambiguities in the policy. Yes, lots of
> >> people are already using numbers in other regions but if you want to
> >> continue to do that with new requests we need to solidify the policy
> >> and make it clear that this is ok.
> >>
> >
> > I don't think I took the discussion out of context at all. Both Matt
> > Petach and I stood up and said that we both use numbers out of region.
> > What's the big deal? John commented that it wasn't. If that's not
> > accurate, I'd be interested in your interpretation.
> 
> Sorry about digging up this old thread...
> 
> The problem (my problem anyway) is that I was just told by ARIN staff (Leslie
> Nobile at the event yesterday and confirmed just now via the
> helpdesk) that NRPM 4.5 can't be used by an ARIN member to request
> resources to be used out of region and, perhaps more troubling, that v4
> space used out of region does not count towards calculating space
> utilization when an organization requests additional space from ARIN unless
> a less specific route is advertised in-region.
> 
> I wasn't prepared to argue about either of these "policies" yesterday, but
> after searching the NRPM, I can't find any basis for either of them.  So, I
> called the helpdesk to double-check / ask where in the NRPM I could find
> these "policies".  I was told they're based on ARIN's interpretation of 2.2,
> specifically "The primary role of RIRs is to manage and distribute public
> Internet address space within their respective regions."  Those are some very
> specific "policies" derived from a very vague sentence, which I think could
> just as reasonably be interpreted as meaning the RIRs exist to serve
> organizations headquartered in their respective service regions.
> You're a US corporation, ARIN is your RIR.  You're a German corporation,
> RIPE is your RIR.
> 
> >> "The requirement to have a minimal level of resources deployed in the
> >> region (/44 for IPv6, /22 for IPv4, 1 ASN) is an attempt to respond
> >> to law enforcement and some community concerns. An absolute
> threshold
> >> ensures that those applying for ARIN resources are actually operating
> >> in the region and not simply a shell company, but it avoids the known
> >> pitfalls of trying to use percentages of the organization's overall
> >> holdings to do that."
> 
> I think I understand the intent of that requirement, but it's kind of pointless,
> especially as currently written:
> 
>   ARIN registered resources may be used outside the ARIN service region.
>   Out of region use of IPv4, IPv6, or ASNs are valid justification for
>   additional number resources if the applicant is currently using at least
>   the equivalent of a /22 of IPv4, /44 of IPv6, or 1 ASN within the ARIN
>   service region, respectively.
> 
> If I were a European org wanting to get space from ARIN instead of RIPE, all I
> have to do is buy a VM on some ARIN-region cloud provider, get them to do
> BGP with my VM, and I've satisfied the requirement?  If you think that's far
> fetched, see the recent thread on NANOG about cloud providers willing to
> do BGP with VMs.  All that does is slightly increase the cost of violating the
> intent of the policy.
> 
> I think 2014-1 shouldn't be necessary, but since current ARIN "policy" is
> based on an interpretation I disagree with of vague language in section 2, I
> suppose I'm all for passage of 2014-1.
> 
> I just hope it doesn't have the sort of unintended consequences I think it
> could easily have.  Does the helpdesk have many staffers who speak
> Romanian?
> 
> ----------------------------------------------------------------------
>   Jon Lewis, MCP :)           |  I route
>                               |  therefore you are _________
> http://www.lewis.org/~jlewis/pgp for PGP public key_________



More information about the ARIN-PPML mailing list