[arin-ppml] Against 2013-4

Steven Ryerse SRyerse at eclipse-networks.com
Tue Jun 4 17:30:41 EDT 2013

David we have disagreed quite a lot in the past but I pretty much agree with the spirit of your comments below.  

I have stated several times in the past that allocations should always be made - but they need to be made with some sort of rightsizing.  Obviously the difficulty is defining a policy that does that but I think it is possible to create one that just takes into account the size of the organization and/or the size of the existing network and the requested size of the block.  So when an allocation request comes into ARIN, at least the community defined minimum size block would always be allocated, and larger requests would be matched against the size of the organization and/or the size of their existing allocations or maybe networks.  There could be a reasonable appeal process if a larger request was only filled with a smaller block. The remaining amount of ARIN's IPv4 allocations should NOT be factored in.   In this role ARIN is the registrar and is basically performing the role Jon Postel did of getting right sized allocations out there.  (ARIN and this community should also perform the technical role the same as they do now.)  Jon gave me 5 Class C's when I asked for them for a customer but I'm guessing he wouldn't have given me a  Class A without making sure there was a large organization behind the request.  Don't know if he would have given me a Class B in 1994 (I wish I had asked).  I'm guessing he would have if I could have convinced him of a possible need.  

This approach worked pretty well in the past and it could work well again.  I also believe that this is the approach that perfectly aligns allocation policy with ARIN's Mission Statement.  It seems to me that the Mission Statement goes out the window on a lot of discussions in this community and it should be the driving force behind all discussions.

Finally I think a free market will bring out plenty of IPv4 addresses that are currently unavailable and this will keep hoarding and prices down.  I would point out that most of the IPv4 addresses are already being hoarded (just look at the /8 list) and only easy access to reasonable sized IP allocations will reduce hoarding.  Tight restrictive policies like ARIN uses now encourage hoarding and we are seeing that happen.  Organizations hoard when they fear not being able to get needed resources and are less likely to hoard when resources are easy to get.  I suspect that a free market in ipv4 addresses would make the equivalent of many /8's available to the open market.  Certainly there is enough unused IPV4 resources to get us thru until IPv6 is more prevalent.  The real world I have to live in has one of my upstream (BGP) fiber providers supporting IPV6 and the other not supporting it at this time.  I have to have the right amount of IPv4 to get me thru and so does everyone else.

-----Original Message-----
From: David Farmer [mailto:farmer at umn.edu] 
Sent: Tuesday, June 04, 2013 1:23 PM
To: Steven Ryerse
Cc: Blake Dunlap; Kevin Kargel; arin-ppml at arin.net; David Farmer
Subject: Re: [arin-ppml] Against 2013-4

On 6/4/13 10:52 , Steven Ryerse wrote:

> Who out there knows what all of these stakeholder think?  I don’t 
> think anybody knows.  So to make comments in this forum that 
> presupposes that this entire community feels one way or another is 
> inaccurate because nobody really knows.

I for one don't presuppose that the whole community feels one way or the other.  However, every time we've touched this set of issues from one side or the other in the last couple years we have not gained consensus one way or the other.

> Of course we could ask everyone in the ARIN community to comment.

We are, that is the point of ARIN-2013-4, everyone should take it as asking the question of the whole ARIN community.


> Free markets with reasonable governance always work.

I agree, however what constitutes "reasonable governance" is the very question being debated.

> Central planning
> and control always fails and always provides uneven results.  Needs 
> based testing is central planning!  The current policies are producing 
> uneven results.
> It is time to halt the current needs based allocations.

I'll agree the current policy for measuring need has issues, especially as we move toward IPv4 free pool exhaustion, but that doesn't mean the principle of need is wrong.  If we completely abandon need as a principle what principle would you put in place that would help ensure fair distribution of IPv6 and ASN resources?

I believe the real issue is that IPv4 conservation thinking has warp our sense of what need really means and should be.  I want to see a meaning of need more representative of current IPv6 and ASN policy, or early '90s IPv4 polices, and not the hyper conservation need polices we have for IPv4 today.

I agree that the price of IPv4 addresses will be the primary force determining need for IPv4 going forward, and at least influence IPv6 adoption and demand.  However, that doesn't mean that just anyone can buy IP address, there needs to be some minimal threshold, like actually operating a network, and being responsive to technical and operational issues.  Also, personally I believe there should be an upper limit on how much address space you can buy without a more detailed justification, I've proposed /12 in the past.

In my opinion "free markets with reasonable governance" have some kind of checks and balances placed on the market, and I believe some minimal operational need requirement is one of those checks and balances in this case.

David Farmer               Email: farmer at umn.edu
Office of Information Technology
University of Minnesota
2218 University Ave SE     Phone: 1-612-626-0815
Minneapolis, MN 55414-3029  Cell: 1-612-812-9952 ================================================

More information about the ARIN-PPML mailing list