[arin-ppml] Legacy space holders were a big part of the community... i.e. all of it.

David Farmer farmer at umn.edu
Wed Aug 27 19:17:35 EDT 2008

On 27 Aug 2008 bmanning at vacation.karoshi.com wrote:

> On Tue, Aug 26, 2008 at 07:14:30PM -0500, David Farmer wrote:
> > On 23 Aug 2008 John Curran wrote:
> > > It is a very relevant point, since almost all of those same legacy
> > > holders were certainly part of the consensus decision in 1993 to
> > > change the basic IP address structure to allow variable sized
> > > blocks (aka "CIDR") and the matching matching address allocation
> > > policies (RFC1466/RFC1518/RFC1519).  These documents state that an
> > > organization should receive sufficient address space to meet two
> > > years worth of organization need, so that we could "delay
> > > depletion of the IP address space".
> > > 
> > > The community of the legacy space allocation era actually already
> > > reached consensus years ago that variable-sized blocks were needed
> > > and that organizational allocations based on two years of need
> > > were most appropriate. They just forgot to return their own extra
> > > space, for reasons unknown, 
> > 
> > With twenty-twenty hind sight, I think maybe this should have
> > happened.  But I can't find anything explicitly calling for this at
> > the time.  Most of the stuff I have found was really focused
> > forward, but that is probably natural, that is where the cliff was. 
> > Do you know of anything calling for what are now classful legacy
> > resources to be repartitioned with CIDR and excess resources to be
> > returned?  Especially from the mid-to late 90's time frame, but even
> > from their early 2000's.  
>  RFC 1797 and the IETF PIER WG archives

RFC 1797 - Class A Subnet Experiment

We this is an experiment, it is related, but it doesn't have a call to do 
anything.  However, with the pointer to IETF PIER WG, I found RFC 1879, 
and RFC 1917.  Thanks Bill, it wasn't a direct reference but it got me there.

RFC1879 - Class A Subnet Experiment Results and Recommendations

It says we can do it but still doesn't call to action.


   The RFC 1797 experiment appears to have been a success. We believe it
   safe to start carving up "Class A" space, if the spaces are delegated
   according to normal IR conventions [7] and recommend the IANA
   consider this for future address delegations."

RFC1917 - An Appeal to the Internet Community to Return Unused

Here we go this is it I know it had to be there some place.  FYI, it is BCP 4 
by the way

"4. Appeal

   To the members of the Internet community who have IP network
   assignments which may be currently unused, the Internet community
   would like to encourage you to return those addresses to the IANA or
   your provider for reapportionment.

   Specifically those sites who have networks which are unused are
   encouraged to return those addresses. Similarly to those sites who
   are using a small percentage of their address space and who could
   relatively easily remove network assignments from active use, the
   Internet community encourages such efforts.

   To those sites who have networks which will never need to connect to
   the global Internet, or for security reasons will always be isolated,
   consider returning the address assignments to the IANA or your
   provider and utilizing prefixes recommended in RFC 1597.

   In those cases where renumbering is required, sites are encouraged to
   put into place a plan to renumber machines, as is reasonably
   convenient, and work towards minimizing the number of routes
   advertised to their providers."

So for the record, Randy, myself (my employer actaully), and many other 
Legacy holders are evil incarnate because we have ignored RFC 1917 and 
BCP 4 and not returned our unused address space. :) 

But even more interesting is what follows:

"4.1 Suggestions to Providers

   Many providers are currently advertising non-CIDR routes which
   encompass a large block of addresses, ie any Class A (0/1) or Class B
   (128/2) space.  Some customers who are only using a percentage of
   their address space (assuming they are subnetting using contiguous
   bits) may be willing to allow usage of the upper portion of their
   assigned address space by their providers other customers.

   This scheme requires certain elements be installed or already in
   place to get the routing correct, but has the potential to gain the
   use of a large number of small networks without growth of the global
   routing tables.  This would require additional measures of
   cooperation between providers and their customers but could prove to
   have both economic advantages, as well as good Internet citizen

So why aren't the "Black Market IP Traders" just not following BCP 4, and 
according to this quote "good Internet citizen(s)"?

And why wouldn't a Liberalized Transfer Policy be better that this? 

Why isn't having this recoded in the registry is a much better idea?

In my gut I still don't like the idea of a Liberalized Transfer Policy, but there 
are lots of things I don't like in my gut that work perfectly fine.  For me 
personally the above are interesing questions, can you help answer them?

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

More information about the ARIN-PPML mailing list