[arin-ppml] Policy Proposal: Predicable IPv4 Run Out by Prefix Size

[David Farmer]

On 10 Jun 2009 Martin Hannigan wrote:

> Based on the recent discussions here I think that the language should
> be modified to include reference to inventory that ARIN may hold. For
> example, if someone does return a /8 and there is a need, the policy
> should possibly not be in effect unless there is only "one" /8
> available for allocation. 

In the rationale I do talk about returns, etc...  But lets dig in a little, if by some 
miracle more /8s are returned or recovered I think the policy still works fairly 
reasonably as stated.  For the sake of this discussion lets say ARIN got back 
several /8s and kept them, I'm not say that will or should happen, I'm just 
playing what-if with this policy.

ARIN's inventory of /8s is I, maximum allocation size would be M, then; 
8 > I >= 4; M = /8;
4 > I >= 2; M = /9;
2 > I >= 1; M = /10;

Even if very large inventories of IPv4 addresses were returned, the policy still 
produces what seem like reasonable limits to me.  

Even if we were to assume something completely silly like IANA were to give 
all the currently available IPv4 addresses to ARIN and implement 10.4.2.2 
today then, ARIN would have an inventory of 28 /8s or so;

32 > I >= 16; M = 4 /8s
16 > I >=  8; M = 2 /8s 

So even with crazy large inventories it seems to work ok.

If those limits really created a problem for anyone it wouldn't be that hard to 
remove the policy.

> I have some issues with justification time
> frames as well, but interested to hear the response on the inventory
> question.

What do you me by "justification time frames"?  Do you mean the other 
proposal "Predictable IPv4 Run Out by Allocation Window"?  Or, do you 
mean the that 10.4.2.2 is the trigger for this policy?

> Best,
> 
> Martin

Anyway it is getting late and crazy thoughts are running through me head like 
the Maya calendar ending on Dec 21, 2012, so IPv4 should almost last to the 
end of the world anyway.  Or even crazier, the Maya knew when we need to 
go to 128 bit addresses to start the new IPv6 world.  

(switch register overflow -  PANIC!) 

bye



===============================================
David Farmer                                      Email:farmer at umn.edu
Office of Information Technology
Networking & Telecomunication Services
University of Minnesota		       Phone: 612-626-0815
2218 University Ave SE		       Cell: 612-812-9952
Minneapolis, MN 55414-3029	       FAX: 612-626-1818
===============================================