Summary

  • An active individual Internet-Draft proposes that IANA set aside IPv6 space for use in space and delegate from it through the existing RIRs. Its preferred design could leave as many as five prefixes per planet, but it does not define “space use”, the block size or the rule that makes one RIR appropriate.
  • Using the existing registry system has a strong case: it preserves a common delegation chain and avoids mission-by-mission number plans. Yet a terrestrial country boundary cannot silently choose the registry for a platform whose owner, controller, gateway, launch state and location may all point in different directions.

The draft chooses the institutions before the constituency

The shortest route to an IPv6 registry for space appears sensible. IANA already delegates address space through five Regional Internet Registries. Those registries already know how to receive a block, write policy with their communities, identify a holder and publish registration data. Why invent a sixth institution when an operating system already exists?

That is the strongest case for draft-kumari-tiptop-address-space-00. The active individual Internet-Draft asks IANA to allocate a block specifically for space use and to delegate from it to existing RIRs. The RIRs would then decide policy and make allocations to network service providers or other subregional registries. Its preferred first arrangement could produce up to five prefixes per planet—one allocation lineage for each RIR—rather than one IANA-managed prefix for each planetary body.

The proposal is early by design. Its own status page says it is an individual Internet-Draft, not an IETF-endorsed document and not a text with formal standing in the standards process. It leaves the size of the block undecided. It leaves the definition of “space use” undecided. It leaves allocation and assignment policy to the RIRs. None of that is a defect in deployed infrastructure, because there is no deployed infrastructure to judge. IANA’s current IPv6 registry still limits ordinary unicast allocations to 2000::/3; the captured record contains no special block for space.

One omission is more basic than a prefix length, however. The draft says an individual or organisation could obtain an allocation directly from its “appropriate RIR (or other) registry”. It never supplies the selector that makes a particular RIR appropriate.

On Earth, “regional” has a working shortcut

RIPE NCC’s public service-region page shows how the current model makes the word regional operational. It describes a territory of more than 75 countries. Organisations providing services in those countries may become RIPE NCC members to receive resources. Organisations providing services outside the region are advised to contact the RIR responsible for their country.

The rule is not merely a coloured map. It connects an applicant, a service footprint, a membership relationship, a body of regional policy and a registry database. Current RIPE IPv6 policy then asks familiar questions: is the organisation an LIR or End User, what is it assigning, where will the use occur, and what must be registered? The answers are contestable, but the first administrative doorway is generally visible.

A space segment dissolves the convenience of that first doorway. Consider a hypothetical craft incorporated under a company in one RIR region, launched by a state in another, operated through a gateway in a third and serving endpoints whose ground stations move with orbital geometry. On a later mission phase it may leave Earth orbit for the Moon or Mars. None of these facts is fanciful as a class of fact; none is asserted here about a real mission. The point is that each could become a plausible administrative selector, and the draft chooses none.

The location of the object would be the most literal answer and perhaps the least stable. The legal home of the holder would be stable but could reward corporate routing around policy. The ground gateway is observable but may change from contact window to contact window. The launch state says little about long-term operation. The customer base may span every region. “Appropriate” is doing constitutional work while looking like an adjective.

Five prefixes are an outcome, not the first decision

The prospect of up to five prefixes per planet makes for a vivid routing question. Multiple aggregates could be harmless in a vast address space, or they could become the first layer of fragmentation. The answer depends on allocation size, routing practice, interconnection, mobility and what a planet means as a topological unit. The draft does not yet provide those inputs, so a verdict about scalability would outrun the evidence.

The prior question is who has authority to create each branch of the tree. If the selector is the holder’s legal domicile, a transfer of corporate control might move the policy relationship while the spacecraft remains on the same orbit. If it is the controlling ground station, a routine gateway change could cross regions without changing beneficial ownership. If it is the mission’s destination, a vehicle in transfer has no settled region. Every choice creates a different handoff, audit and return problem.

Remco van Mook’s 10 September RIPE Labs article makes the timing visible. It says the proposed hierarchy would send policy questions into RIR policy-development processes even though the policy-forming communities have mostly not confronted them. His separate EXPANSE problem statement catalogues the missing registration model, moving-platform problem, jurisdictional boundaries, aggregation risk and discontinuities in RPKI. That second text is also an individual Internet-Draft and explicitly makes no request of IANA. It is a problem inventory, not a competing allocation decision.

Taken together, the two drafts offer a useful division of labour. TIPTOP supplies an institutional shortcut. EXPANSE shows why the shortcut cannot be treated as an answer. The existing registry system is valuable precisely because it binds uniqueness to accountable decisions. Reusing the names of its institutions without defining their new constituency would preserve the hierarchy on paper while leaving its most consequential edge discretionary.

Put a selector record ahead of the policy queue

The minimum useful object is not a treaty for the Solar System. It is a public selector record attached to any future request before allocation policy is applied. It could state:

  1. the legal resource holder and operational controller;
  2. the declared boundary of “space use” for the request;
  3. the RIR or other registry selected;
  4. the public rule and facts that made that registry appropriate;
  5. relevant ground-segment and service jurisdictions without exposing sensitive coordinates;
  6. the mission phase and condition that would trigger review;
  7. the handoff or transfer path if the controlling fact changes; and
  8. the return, retirement and correction conditions.

Such a record would not decide whether domicile, control, gateway or destination should prevail. It would force the authorised community to make that choice visibly and apply it consistently. It would also allow two registries to recognise the same request before both issue space from separate branches.

Privacy and security still matter. A public receipt need not disclose an orbital ephemeris, defence mission, contact schedule or network diagram. It can expose the rule, accountable organisations, coarse mission state and decision history while sensitive technical evidence remains under bounded review. Terrestrial registries already separate public registration data from protected due-diligence files; the design challenge is real but familiar.

Nothing has been allocated

The cleanest reading of the record is also the least dramatic. Two authors have submitted an early address-space draft. A RIPE community contributor has asked RIR communities to engage before the architecture hardens. Another individual draft lists the questions the first text does not answer. There is no IETF consensus, no IANA allocation, no RIPE NCC policy proposal, no applicant and no route to audit.

That is why the selector belongs now. Once prefixes are issued to missions and embedded in flight systems, changing the institutional rule becomes a migration rather than a discussion. The practical advantage of early standards work is not that every early idea survives. It is that the expensive choices can still be rejected while they are words.

Sources