[arin-ppml] Draft Policy ARIN-2013-4: RIR Principles

[William Herrin]

On Wed, May 29, 2013 at 12:17 AM, Jason Schiller <jschiller at google.com> wrote:
> Also your rewrite loses the
> importance of the numbers being unique, and properly documented.

Hi Jason,

How do you figure? My rewrite spoke only to 0.2, not 0.1 or 0.3. I
don't have conceptual problems with 0.3. Maybe some wordsmithing. As
for documentation, to the extent that it serves sustainability, it's
critical. To the extent that it doesn't, it's undesirable.

> Your point about aggregation is the cart before the horse is well taken.
> Andrew Dul, had a similar point, and suggested rewrite adding something
> about the state of technology.  I think his rewrite is is more universal in
> that it does not specifically call out BGP.

Andrew's rewrite continues to focus on hierarchy and aggregation which
IMO a statement of principles shouldn't. The principle is: support
scalability, however that's achieved.

I called out BGP because BGP has been the core routing technology for
the duration of ARIN's existence and there is no end in sight. What
value lies in abstracting it out of the principles statement? Doing so
can only lead to confusion.

Aggregation is the only known tool for achieving scalability in the
current technology (BGP). That's a central point that should not be
lost to abstraction when we consider policy based on these principles:
when we're talking addresses for announcement into BGP (and we usually
are), scalability requires aggregation. Period. Full stop.

When we're talking about addresses for use with another technology, we
have to consider scalability freshly without getting hung up on
aggregation.


> I hope you will find Andrew's rewrite sufficient.

You mean adding the phrase "as required by the current technical
limitations of global routing protocols?" No, that isn't sufficient at
all. The hierarchy statement is broken at its core and hanging a brief
conditional on it doesn't fix it up.


> WRT sustainability as opposed to conservation, could you please expand.  I
> don't think i quit grasped your point.

http://www.merriam-webster.com/dictionary/sustainable

2. a : of, relating to, or being a method of harvesting or using a
resource so that the resource is not depleted or permanently damaged


It's well enough defined that I'm reluctant to expand on it but I'll
give it a shot anyway.

Basically three phases to a protocol's existence with respect to its
number pool:

Phase 1: Early. Conservation is a distant concern. The priority has to
be: don't kill the baby. IPv6 is in the Early phase and the baby looks
underfed.

Phase 2: Mainline. Conservation comes into play here. It should
discourage unduly wasteful use of the number resources and studiously
avoid practices which permanently remove the numbers from circulation.

As use increases, so should the conservation efforts. Ideally, the
conservation efforts are sufficient for the protocol to reach
obsolescence before the number assignment process leaves the mainline
phase. Didn't happen with IPv4, but that's the goal.

Phase 3: Zero sum. If the number pool approaches depletion before the
protocol reaches obsolescence then the numbers have to move from less
efficient uses to more efficient uses. Exactly how is a subject of
policy rather than a statement of principles but this is where things
like markets, resource reviews and reclamation come in to play.


Point is, conservation is only front and center during the mainline
phase of the number management process. In the early phase it's a
hindrance and in the zero sum phase it's moot. The abstract principle
of which conservation is a part, the one that applies throughout the
process, is sustainability.


> I was hoping to have a very clear line from the current draft to RFC-2050,
> and tried to carry those principles forward and use much of the same
> language.

I think that would yield a net negative result. If RFC 2050 is the
primary resource then we should incorporate it by reference without
trying to repeat it in our own words. The repetition can only confuse
matters and set up needless conflicts. If we recognize that 2050 is an
antiquated document then let's read between the lines at what its
authors were targeting and avoid getting caught up in wording and
choices that technology has passed by.

Regards,
Bill Herrin



-- 
William D. Herrin ................ herrin at dirtside.com  bill at herrin.us
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004