[arin-ppml] ARIN Multiple Discrete Networks Policy

[Richard A Steenbergen]

On Mon, Oct 03, 2011 at 10:29:06PM +0000, John Curran wrote:
>
> The policy text does not reference unique routing policies.

Even Owen, who is trying desperately to disagree with me (I'll pass on 
any particular speculations as to why :P), has stated that a discrete 
network is one which must implement unique routing policies.

Since I apparently need to cite examples from every previous e-mail 
every time now, allow me to point out his attempt to "clarify the 
current defintion":

> Insert new 4.5.3: Discrete networks are separate networks which have 
> cannot usefully share a common routing policy. Examples might include 
> networks with any of the following characteristics:

So, if a network has one of the (previously defined in the policy 
language) characteristics, and thus cannot usefully hsare a common 
routing policy, they have discrete networks AND good justification for 
running discrete networks. I'm more than happy to have this 
clarification passed as a policy change, since it's just restating how 
it "already is".

> Did you cut and paste the wrong section?  Please resend the section 
> which refers to "unique routing policies".  You quoted a section 
> regarding "Separate unique 'globally routable' address space' "

I believe I outlined how "separate unique globally routable address 
space" equates to "unique routing policies" in a previous e-mail, but 
let me know if you didn't understand it and I'll be happy to try and 
explain it again.

> > What does "I've already allocated address blocks to my POPs, which I 
> > cannot rearrange", have to do with "geographic distance and diversity 
> > requirements"? Where does this language come from? What is your basis in 
> > the policy to support this claim?
> 
> You need a 'compelling reason' to have your infrastructure considered
> _multiple discrete networks_

Uh huh... Funny, I would have thought that meeting one of the example 
"compelling reasons" listed in the policy would serve as proof that you 
have a compelling reason. Meanwhile, you seem to feel that you need to 
add new, undocumented conditions to the example "compelling reasons" in 
order to justify a compelling reason. I fail to see how this makes any 
kind of sense, or can be justified under the current policy.

> > When has this claim ever been applied 
> > to any other MDN applicant? And for the record, that part *IS* 
> > completely true in my specific situation, so it's not even a blocker, 
> > but it's also complete nonsense too, and needs to be called out as such.
> 
> Richard - If that part is completely true in your situation, then 
> please reapply asap. You indicated otherwise on your last request to 
> ARIN for space under the MDN policy.

No really I didn't, and this is the part you keep failing to understand! 
All I said was that I could potentially work around the need for this 
policy by doing something in a suboptimal way, but in the SAME WAY that 
everyone else could be too. If you think that everyone else can't do it 
too, you're confused, but fortunately the policy says NOTHING about "you 
must not be able to do this any other way". Maybe I should have just 
lied and saved myself the trouble, but I was trying to educate you on 
the practical realities of routing on the Internet. I guess no good deed 
goes unpunished, and all of that.

> If that's the case, then the most constructive path would be to 
> propose unequivocal text for an MDN policy change.  If you believe 
> that "discrete networks" are "networks that implement unique routing 
> policies" then that would be a fine change.

I believe Owen has already done this for us, and I'm happy to support 
his proposal.

> That would also result in the elimination of ARIN having to determine 
> if a "compelling reason" was present at all, since with such a 
> definition, since we would simply be able to ask: "Do you have two or 
> more networks with unique routing policies?", and if the applicant 
> said "yes", then they would meet the MDN policy by having _multiple 
> discrete networks_ per definition of _discrete network_.

*NO* it absolutely would not! The fact that you keep saying this 
indicates a clear and massive misunderstanding somewhere, which is what 
I keep trying to correct!

Having a network with a unique routing policy is a STATE OF BEING. It is 
EASY to do, all you have to do is GO CONFIGURE IT AS SUCH. This is in NO 
WAY justification for doing so, and it is certainly not justification 
for obtaining any kind of special treatment from ARIN. For that, you 
need to have a COMPELLING REASON to *BE* in that state, which the 
current policy defines (quite well too, if you would stop trying to make 
up random undocumented rules just because you think you should be :P).

-- 
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)