[ppml] Proposed Policy: IPv6 Direct assignments to end sites

Marshall Eubanks tme at multicasttech.com
Tue Aug 30 13:33:27 EDT 2005


On Tue, 30 Aug 2005 12:46:15 +0200
 Kurt Erik Lindqvist <kurtis at kurtis.pp.se> wrote:
> 
> On 29 aug 2005, at 16.17, Member Services wrote:
> 
> [Proposal deleted]
> 
> You might want to note that there is no guarantee that this  
> assignment will actually get routed. Just as in some RIR regions you  
> will get less than a /24 v4 space as PI space, and you won't get very  
> far...

Dear Kurtis;

There is no guarantee that an IPv4 /24 will get routed either, nor the
/23's assigned by 2002-3. (I see this as basically the  IPv6 analogue of  2002-3.)

However, having officially assigned address block does certainly help.

What the current proposal really does is set a new expectation of
IPv6 /56's as becoming standard. I would feel more comfortable hearing from a critical mass of IPv6
vendors or SP's that this would not cause problems (it shouldn't, of course, but...). Unless it
does, I see no reason not to support this.


> 
> >         6.5.8.1. To qualify for a direct end site assignment, an
> >                   organization must:
> >
> >                   a) not be an LIR;
> >
> >                   b) be an end site;
> >
> >                   c) be currently multi-homed using IPv6 to two or  
> > more
> >                      separate LIR's. native connections or statically
> >                      configured tunnels may be used to satisfy this
> >                      requirement.
> 
> 
> What's a "statically configured tunnel" ? Does MPLS count? I would  
> not call that a statically configured tunnel. So most L2VPN services  
> are out of the question...
> 

I do not know, but  I suspect that he was trying to include the many sites that
get IPv6 through some sort of quasi-permanent tunnel, which  seems reasonable to me,
at least at this state of deployment. This is a tricky  issue (one could always order
2 T1's for a few months and get an ASN, all to get a 2002-3 address block, and 
then remove them, but that seems like a lot of work and cash
for a v4 /23 if you don't really need it. Now, if I can just set up two tunnels to somewhere, maybe
for free, and then take them down once I get a /48, maybe that is not a high enough barrier to
abuse. 

Here is a suggestion : tunnels (of any sort) should count for multi-homing if you are already IPv4
multi-homed or have IPv4 address blocks assigned to you (though 2002-3 or otherwise), but otherwise
you need direct IPv6 connections. (I.e., if you  are .v4 multihomed, you get a certain presumption
of being truly .v6 multihomed.)

> - kurtis -

Regards
Marshall

> _______________________________________________
> PPML mailing list
> PPML at arin.net
> http://lists.arin.net/mailman/listinfo/ppml




More information about the ARIN-PPML mailing list