[ppml] 2002-3: Micro-Assignments for Multihomed Networks

[Richard Jimmerson]

Hello Forrest,

> Is it possible for someone to tack on all of the discussion 
> from 2002-7 onto this proposal.  No need to rehash everything 
> that's already been said.  

Because of the similarities between 2002-3 and 2002-7,
the discussion that has taken place on this list for 2002-7
will be included in the discussion summary for 2002-3 at the
upcoming ARIN XI public policy meeting.

Best Regards,

Richard Jimmerson
Director of Operations
American Registry for Internet Numbers (ARIN) 


> -----Original Message-----
> From: owner-ppml at arin.net [mailto:owner-ppml at arin.net] On 
> Behalf Of Forrest
> Sent: Friday, March 21, 2003 12:01 PM
> To: ppml at arin.net
> Subject: Re: [ppml] 2002-3: Micro-Assignments for Multihomed Networks
> 
> 
> 
> On Fri, 21 Mar 2003, Member Services wrote:
> 
> > New text:
> >   If an end-user is not multi-homed, the minimum block of IP
> >   address space assigned by ARIN is a /20. If assignments 
> >   smaller than /20 are needed, end-users should contact their 
> >   upstream provider.
> > 
> >   If an end-user is multi-homed, and has an ARIN assigned
> >   ASN, the minimum block of IP address space assigned by
> >   ARIN is a /22.  If assignments smaller than a /22 are
> >   needed, end users should contact their upstream provider.
> 
> 
> 
> I'm not sure I see the point in reducing the minimum 
> allocation from /20 
> to /22.  Are people going to come back 3 years from now and want the 
> proposal changed again to make the minimum /24?  One of the 
> arguments for 
> changing the minimum allocation is that people currently lie 
> about their 
> needs and uses in order to qualify for a /20.  I think with 
> this proposal, 
> people that could get by with a /24 are going to find 
> creative ways to 
> waste addresses to qualify for a /22.  Whether someone 
> receives a /24 or a 
> /22, its only one added prefix to the global routing tables, 
> so why not 
> give people what they REALLY need and conserve IP space.
> 
> I strongly feel there needs to be a mechanism added that 
> enables IP space 
> to be reclaimed if an organization ceases to be multihomed.  
> 
> Is it possible for someone to tack on all of the discussion 
> from 2002-7 
> onto this proposal.  No need to rehash everything that's already been 
> said.  
> 
> Forrest
> 
> 
> > 
> > Problem Summary:
> >   Many end-user organizations are choosing to multi-home
> >   for reliability reasons.  At the same time, many are
> >   using technologies such as NAT, or load balancers that
> >   reduce the need for external IP space.  These groups
> >   are forced today to take one of two actions:
> > 
> >      1) Use IP space from one of their upstreams on both
> >         connections.  This can lead to load balancing
> >         issues, and also makes the end-user more dependent
> >         on the ISP who assigned the space.  The ISP's
> >         business problems, for instance could force downtime
> >         and/or renumbering.
> > 
> >      2) "Waste" address space (often by not using the 
> >         technologies that conserve it) in order to qualify
> >         for a /20 under the current policy.
> > 
> >         In order to allow people to both conserve address
> >         space, and reap the benefits of multi-homing the
> >         minimum size assignment for those who do multi-home
> >         should be made smaller.
> > 
>