[arin-ppml] ARIN Multiple Discrete Networks Policy

[Richard A Steenbergen]

On Sat, Oct 01, 2011 at 04:13:41PM +0000, John Curran wrote:
> > 
> > Ok so, your current position is that matching an example from the 
> > specifically defined list of example compelling reasons is not a 
> > compelling reason, because of an exclusionary rule which doesn't exist 
> > anywhere in the policy (namely that you can't possibly solve the problem 
> > with deaggregation) that you're making up on the spot? 
> 
> Incorrect.  If an organization can demonstrate a compelling need 
> for multiple discrete networks (along the lines of the examples),
> it can make use of the MDN policy. This is as the policy requires.

That would be fine, but how does one go about doing that? I would assume 
that one would demonstrate how their situation matches one of the 
examples in the policy, which I've already shown, but so far that hasn't 
been good enough for you. After that is done, you then come back to add 
a rule which doesn't exist in the policy that says "if you can possibly 
implement this with deaggregation instead, you don't have a compelling 
need". This language simply doesn't exist, and for good reason, because 
it would exclude almost everyone from using the MDN policy.

I have yet to see you argue anything about how a specific situation 
doesn't match one of the example compelling needs. If you did, I could 
easily explain how it does (as I tried to do in a previous e-mail, 
talking about the "geographic distance and network diversity" example), 
but that HASN'T been your objection so far, all you've had is this made 
up rule about "aggregation".

Can you please address this point? Where in the policy does it say that 
if you can solve the problem with deaggregation you don't have a 
compelling reason? If it doesn't, what is your basis for claiming that 
someone who matches on of the example compelling reasons doesn't have a 
compelling reason? Or is the arguement about the matching itself? Please 
pick one so I can figure out where your objection is and explain the 
confusion.

> Refining the examples is also an option for improving the clarity
> of the policy, but that still leaves ARIN staff having to make a 
> judgement call regarding what constitutes "compelling criteria."
> I believe that the more certain outcome would be to change the 
> policy to specifically include a clear technical definition of 
> what constitutes "multiple discrete networks" in the policy text.
> 
> Richard - Can you propose an appropriate definition of 
> "multiple discrete networks" that could be added to the 
> policy?

I would say that a discrete network is one which must implement a unique 
routing policy which is distinct from and other network, and multiple 
means there is more than one of them. There are a variety of technical 
and political reasons why someone might have a compelling reason to do 
this, such as the 3 examples specifically listed in the policy already 
(political/legal issues, geographic distance or network diversity 
needs, and discontiguous networks).

But how does this help you with ANY of the issues above?

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)