[arin-ppml] Addressing for other planets

[Tony Li]

Hi John,

>> Those aren’t real requirements.  Those would be necessary if the goal was perfect aggregation, which its not (and never is).  If aggregation is not perfect, there will be additional routes.  But that’s actually not relevant to the discussion because you would still be better off than having zero aggregation.
> 
> It’s actually not better off - provider-based allocation is always the optimum case, since addressing follows actual network topology and aggregation is inherent. 


Not true as you fail to capture aggregation that is possible across providers, as it is in this case.


> You are the one proposing a scheme to try and achieve similar aggregation via addressing that’s not topologically aligned with agency network connectivity, and so it will be worse unless you achieve the above conditions and have a significant set of celestial-body relays operational. 


It is exactly aligned with topology.  What you’re advocating is per agency addressing, which has nothing to do with topology.


> Hmm..  Rather than dealing in abstractions, perhaps we can work with actual data?   Is there a network design or operating model for this celestial "shared network”?   


Operational today: NASA’s DSN.  Mars oribiters relaying between multiple (now defunct) rovers.  

Fast forward a few decades and we have a number of colonization sites, hopefully interconnected by land lines and low-Mars orbit satellites, which in turn have an uplink (or two) to Earth.


> I ask because ARIN is capable of dealing with ISP IPv6 allocation requests that are quite sizable and unusual in nature, but we do require some fairly concrete plans to support any such allocation request.  Individual space agency networks are likely to qualify today with sufficiently detailed requests (just as we’ve handled various satellite network requests), but I suspect your Deep Space IPv6 requirements really are really of a nature that need to be handled as a new RIR region – that will allow for allocations to be more expansively sized and for handling the implied operational/policy development requirements. 


This is NOT an allocation request.  This is asking to put in place an architecture to allow for aggregation.

The space allocations need not be expansively sized, but reserving space so that aggregation can happen as things grow would make a great deal of sense.

I don’t think you like aggregation and I don’t understand why.  I don’t want to argue with you anymore.

Regards,
Tony


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20260221/cfdddf96/attachment.htm>