[arin-ppml] Draft Policy ARIN-2026-1: Taking IP To Other Planets (TIPTOP)

[Eric C. Landgraf]

On Mar 25  9:57, arin-ppml-request at arin.net wrote:
> Date: Tue, 24 Mar 2026 15:21:16 -0700
> From: Tony Li <tony.li at tony.li>
> Subject: Re: [arin-ppml] Draft Policy ARIN-2026-1: Taking IP To Other Planets (TIPTOP)
> 
> > Considering appendix A of deepspace-ip-assessment, why
> > does this policy speak to IPv4 at all?
> 
> IPv4 also has a significant advantage in bandwidth overehead.  Deep
> space links are extremely low bandwidth.  Voyager (admittedly not IP)
> gets about 160bps.  Mars rovers, when transmitting directly to Earth,
> get 500bps.  The extra size of IPv6 does make a significant difference
> at these rates.
 
Probably a more human-scale example: AX.25 (X.25 over amateur radio)
typically is run as 1200 baud AFSK over FM (and half-duplex!). I suspect
this is a pretty feasible link speed for "on-demand" interplanetary
networking (as opposed to content networks or sneakernet, which I expect
will be a major fraction of traffic by bandwidth). Do the math: with the
minimum IPv6 MTU, it takes 1s per packet.

That said, I think the implementation solution to save link bandwith
here is large-scale NAT: if your prefixes are easily aggregated, just
strip the prefix! Outside of near-earth orbit if you want to deploy
IPv4, just use the whole space and use NAT64 to address long-distance
(i.e. earth-based) endpoints.

But implementation is out-of-scope for ARIN beyond technical soundness
of policy. And there's no technical reason not to assign IPv4 addresses
for space use, provided we can meet current uniqueness and proposed
aggregation requirements.

As to the policy: I'm generally opposed. I think this policy needs
global development---my understanding is this would go through the ASO
AC (aka NRO NC). Should the ASO designate ARIN as the "RIR for space",
then it would be appropriate to take up this policy proposal. Until
then, organizations can and will go to their most convenient RIR for
address assignment.

	Eric C. Landgraf
	Virginia Tech