[arin-ppml] ARIN Multiple Discrete Networks Policy

Richard A Steenbergen ras at e-gerbil.net
Sun Oct 2 17:02:24 EDT 2011

On Sun, Oct 02, 2011 at 12:07:43AM +0000, John Curran wrote:
> Correct.  I was pointing out that if the policy is changed to simply 
> require a well-defined "state" then no examples would be needed.  If 
> the policy is only changed to equate "discrete" networks to "networks 
> with unique routing policies", then there would still need to 
> judgement regarding whether there was a compelling need for the 
> networks to have unique routing policies.

Uhm... How could you possibly propose to change the definition of a 
discrete network to eliminate the need to justify using them? These two 
things have absolutely nothing to do with each other, this is a total 
non sequitur.

> The current policy says "multiple discrete networks" not "networks 
> with unique routing policies" You've asserted that they are the same 
> and that is not the case accordingly to the current policy.

You asked for an explicit definition and I gave you a perfectly sensible 
one. Are you now claiming that this policy can't be applied because 
nobody knows what a discrete network is? Why not, I guess it's the only 
possible argument left which hasn't already been defeated.

I suppose we could start with the context provided by the policy itself, 
which says:

> Some organizations have requirements for multiple discrete networks 
> that need individual address allocations. Discrete networks must often 
> have separate unique globally routable address space and will often 
> grow at different rates. In order for organizations with multiple 
> discrete networks to request additional address space under a single 
> maintainer ID, the organization must use the following criteria:

So far we know that whatever they are, they must often have separate 
unique global routable address space, and will often grow at different 
rates. Any why might an organization need separate unique globally 
routable address space which can grow at a different rate? The ONLY 
benefit to this would be if they are required to implement unique 
routing policies for each discrete network, and thus require unique 
globally routable address space to do so. If you have another 
explanation, I'm all ears. Otherwise, I'm going to say that this acts as 
the basis for defining what they are, multiple networks which must 
implement unique routing policies.

The policy then goes on to say that you need to have a good reason for 
doing so:

> The organization must have compelling criteria for creating discrete 
> networks. Examples:
> * regulatory restrictions for data transmission
> * geographic distance and diversity between networks
> * autonomous multi-homed discrete networks

These three specific examples are all reasons why one may need to 
implement unique routing policies. One assumes that if you situation can 
be shown to match one of the example "compelling reasons" for doing 
this, and if you have a need for separate unique globally routable 
address space which grows at different rates, you would infact be able 
to use the $%^&ing policy that says these exact words!

> Networks that cannot readily reallocate their existing allocations 
> (for compelling reasons such as those shown in the examples, e.g. 
> regulatory restrictions on interconnection, geographic restrictions, 
> being discontiguous or those that are autonomous by nature) have been 
> judged to be "discrete networks".  Organizations that acknowledge that 
> they can readily reallocate their existing allocations across 
> interconnected network infrastructure have been determined NOT to have 
> "multiple discrete networks", even if such reallocation would result 
> in a routing impact.

Again, you're adding an exclusion which DOES NOT EXIST IN THE POLICY. 
The policy doesn't say ANYTHING about this, but what it does do is list 
some problems, state some requirements, and then provide a solution. If 
your problem matches the example problems, and you have a good reason to 
do it which matches the requirements, then how does ARIN possibly 
justify adding random new exclusions which don't exist in the policy?

Or, let me try another approach. If this is really the position you want 
to take, I hearby call on ARIN to apply the (newly invented) rules 
consistently, and stop giving out allocations under the multiple 
discrete network policy to anyone who could potentially solve their need 
to implement unique routing policies with deaggregation. I believe this 
will exclude almost everyone, including the #1 example "intended user" 
that keeps getting cited, someone with two or more physically 
discontiguous networks. Unless they have some other random situation, 
like not being able to work swip because of a bizaare corporate 
structure as you've been trying to point out in the past, they should 
clearly be able to solve this with deaggregation and thus no longer 
qualify. Does that seem sensible to you?

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)

More information about the ARIN-PPML mailing list