[ppml]How far is too far?

Leo Bicknell bicknell at ufp.org
Fri Oct 3 11:03:00 EDT 2003


In a message written on Fri, Oct 03, 2003 at 10:38:41AM -0400, Ron da Silva wrote:
> There is no ASN requirement currently for end-user space
> (aka assignments).  So should,
> 
> 1) a change from /20 to /22 for assignments require ASN from some RIR
> 2) should ALL assignments (micro or not) be required an ASN
> 3) should NO assignments (mirco or not) be required an ASN ??

If I were rewriting policy from scratch, my personal view is that:

1) To get an ASN you should have to document that you are multihomed.
   (There are a few possible exceptions, like ISP's may be able to get
   a second as a "customer AS" as many do, and critical infrastructure 
   should have one, even if not immediately multihomed.)
2) Everyone with an ASN should be able to get a /24 with no
   justification.
3) Everyone with an ASN should be able to get a /1-/23, as appropriate,
   with proper justification.  (And no, I'm not sure if I like the 80%,
   or some of the other proposals, so I'll leave it at proper
   justification.)
4) ASN's should be limited to getting space from a single RIR, the RIR
   where the ASN was assigned.  (To insure one entity has full review
   over all the space, when reviews are conducted).
5) Single homed groups with a need for a large amount of space, or
   groups not today connected to the internet but able to demonstrate
   that it is likely they will be in the future should be able to 
   go through a justification process to get space, with a minimum of
   probably a /19.

That said, we're not rewriting from scratch, so a wholesale change
such as I would like is not going to work.  I understand that, and
I'm quite willing to chip away at the problem.

Also, to answer what may be another likely question, I believe that
all RIR's should have the same criteria, so yes, I think what I've
outlined would be good for the world.  (Note, that makes #4 make a
little more sense then it does in today's multi-policy world.)

Note, I think something similar should be true for IPv6 as well,
of course with the boundaries being different, and with 5 being
relaxed since there is A) more space, and B) predicting if unconnected
networks will be connected to the IPv6 backbone in the future is
much harder now, so err on the side that they will.

-- 
       Leo Bicknell - bicknell at ufp.org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request at tmbg.org, www.tmbg.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20031003/7d84d51f/attachment-0001.sig>


More information about the ARIN-PPML mailing list