[arin-ppml] ARIN Multiple Discrete Networks Policy

Owen DeLong owen at delong.com
Mon Oct 3 10:44:26 EDT 2011

On Oct 3, 2011, at 5:56 AM, Richard A Steenbergen wrote:

> On Mon, Oct 03, 2011 at 12:10:06AM -0700, Owen DeLong wrote:
>> Admittedly, it uses the word criteria instead of reason, but, I think 
>> picking apart the synonyms really isn't necessary. The meaning seems 
>> relatively clear to me and it indicates that you need a compelling 
>> reason (criteria) for not giving the networks a common routing policy.
> Owen, please read what I said again, I think we're actually in complete 
> agreement here.
> So far what you've said is:
> a) You need a compelling reason to operate multiple discrete networks.
> b) The policy defines some examples compelling reasons for this.
> c) Not having a common routing policy equates to discrete networks.
> I couldn't agree more with any of the above.
> What John is saying is that, in addition to the compelling REASON 
> criteria of the policy, there is also a phantom compelling NEED 
> criteria. This shifts the burdon of proof from "show us that you have a 
> good reason for doing this" to "show us that you have an absolutely NEED 
> to do this", language which doesn't exist in the policy at all. He then 

Nope... I'm saying that criteria equates to NEED in the policy and in the
policy intent.

You're saying that it does not. That's where I agree with John and where
you and I appear to disagree.

> goes on to say that there is no definition equating having unique 
> routing policies with "discrete networks", and therefore he gets to make 
> up any definition for "discrete network" that he likes, and the 
> definition he uses just happens to include this non-existant compelling 
> NEED logic.

Correct. The examples in the policy are intended to equate the actual
NEED for distinct routing policies as criteria rather than the mere use
of distinct routing policies.

> You're absolutely right, the meaning of the policy is QUITE clear and we 
> both have the same interpretation of it. The problem is that ARIN does 
> not agree.

Apparently not since you just reinterpreted it differently than I did.

However, I will submit a proposal to correct this discrepancy. Thank
you for bringing the issue to my attention.


More information about the ARIN-PPML mailing list