On Why the Registry Layer Is a Structural Risk — and Why LARUS Is the Only Proven Business-Continuity Guarantor

Why does holding addresses directly concentrate registry risk rather than remove it?

Conceptual cutaway model of a bridge: traffic continues above while two technicians work on a supported joint below.
Keeping customers connected requires someone to take responsibility for upstream risk. Lu Heng presents LARUS’s role as absorbing registry-layer exposure so customers can focus on their own operations.

Every operator that treats direct RIR holding as the safest structure is making the same category mistake: buying or holding IPv4 directly does not remove registry-layer risk. It concentrates that risk inside the operator’s own legal entity. The registry contract, the policy surface, the audit surface, the renewal surface, and the revocation surface all attach directly to the same company that carries customers, routing, compliance, licenses, revenue, and business continuity. That is not safety. That is concentration.

This is the part much of the IPv4 market still refuses to say plainly. The problem is not whether an address block is valuable. The problem is that the institutional layer beneath it was never built to provide sovereign-grade continuity to high-value, high-dependency assets. It was built as an administrative registry layer, and the contracts still look like administrative registry contracts. The market evolved into capital. The legal surface did not evolve with it.

That is why “owning directly” is not the clean, low-risk position many operators imagine. In law, what the operator gets is not escape from the registry layer, but maximum exposure to it. The operator becomes the first-loss bearer of any policy drift, governance failure, procedural reinterpretation, billing dispute, audit friction, renewal problem, or institutional breakdown at the registry layer.

The documents say this plainly.

Under the [AFRINIC Registration Service Agreement], AFRINIC provides services on a “best effort” basis and says it is not liable for interruptions, errors, inaccuracies, defects, failure to meet the applicant’s requirements, or damage to the applicant or third parties, except where it is conclusively established that AFRINIC did not use the appropriate means. Even then, liability is capped at the greater of the amount paid during the preceding six months or USD 100. The agreement is annual, is expressly subject to adopted policies, can be amended after board-mandated notice, and on termination or expiry AFRINIC immediately revokes the number resources and ceases providing services without liability. That is not title security. That is a renewable service relationship with de minimis downside for the registry.

Under the [ARIN Registration Services Agreement], the services, the included number resources, and the registration thereof are provided on an “as-is” basis, ARIN disclaims warranties, policy changes become binding upon notice or publication, and liability is capped at the greater of six months of fees or USD 100. If invoices remain unpaid, ARIN may stop services, terminate the agreement, and revoke the included number resources. ARIN does recognise contractual rights to the included number resources, but those rights still sit inside a capped-liability, policy-bound, terminable contractual framework. Direct holding there is still direct exposure to registry-layer asymmetry.

Under the [APNIC Membership Agreement], the term is one year, renewal is deemed acceptance of the agreement as it exists at the time of renewal, APNIC Documents are binding as amended, the company excludes liability arising out of or in connection with the agreement, the APNIC Documents, or delegated resources to the extent permitted by law, and it may revoke some or all of the member’s rights under the APNIC Documents, including delegated resources, and terminate the agreement. The member must then immediately cease using the delegated resources identified in the notice. That is not “I own it and I am safe.” That is ongoing delegated use under a contract designed first to protect the registry side.

Under the [RIPE NCC Standard Service Agreement], the agreement can be amended by General Meeting resolution without re-signing, RIPE Policies and RIPE NCC procedural documents form an integral part of the contract as they are revised over time, the RIPE NCC may suspend services, deregister resources, and terminate the agreement through its closure and deregistration procedures, it excludes liability for direct and indirect damages, is not liable for failure to make number resources available on time or for damages connected with their use, and limits liability to the member’s service fee for the relevant financial year. Again, the operator carries the real operating downside; the registry carries a tightly bounded financial downside.

Under the [LACNIC Registration Services Agreement] the agreement is explicitly an adhesion agreement, runs for one year, renews subject to the then-applicable terms and conditions, binds the applicant to guidelines incorporated by reference and as modified over time, allows LACNIC to review utilization at any time, and provides that non-payment, breach, or termination leads to revocation of the IP address space or ASN. Direct holding in that framework is still not insulation from registry-layer risk. It is direct placement under it.

The details differ, but the pattern does not. Across all five RIRs, the holder faces some combination of renewable consent, binding policy drift, audit or review rights, suspension or revocation power, and liability that is excluded or capped at a level trivial relative to the operational damage a real network could suffer. The common mistake is to look at the registry record and think the name on the record is the source of security. It is not. The legal and operational structure beneath the record is the source of security, or insecurity.

That is why direct holding is not safer than leasing by default. If the operator holds directly, the weakest and most asymmetric institutional layer in the stack is attached directly to the operator’s own balance sheet and operating company. If the operator leases through a generic broker, the problem is often only partially shifted, because a broker usually intermediates a transaction rather than internalising registry-layer risk. But if the operator leases from a first-party lessor built specifically to absorb registry-layer exposure, then the structure changes.

That is what LARUS is.

LARUS is not a broker trying to stand in the middle of someone else’s registry exposure. LARUS leases address space from its own pool under commercial contracts designed around continuity. The customer’s operational use sits downstream; the registry-layer interface sits upstream with LARUS. That means the customer is not the direct party facing the RIR contract, the policy surface, the audit surface, the renewal surface, the governance surface, and the litigation surface. LARUS is. This is not a financing convenience. It is structural risk transfer.

That distinction matters because the real comparison is not “buy versus rent” in the consumer-finance sense. The real comparison is: who should carry the thin point in the system? Should it be the customer’s own operating company, which needs continuity above all else? Or should it be a specialised first-party lessor whose business model is built around carrying and defending that exact layer of risk? Once the question is framed correctly, direct holding stops looking like prudence and starts looking like self-insurance against a failure mode the operator is not structurally built to fight.

This is also where LARUS ceases to resemble a normal lessor.

In the final order of the Supreme Court of Mauritius in AFRICAN NETWORK INFORMATION CENTRE (AfriNIC) LTD (In Receivership) v CLOUD INNOVATION LTD (SC/COM/MOT/000399/2025), the Court varied an earlier order so that AFRINIC had to rectify its register of members to add the details of Cloud Innovation Ltd as a member under section 91 of the Companies Act; ordered the Registrar of Companies to make the proper entries in the company records, including the register of members of AFRINIC, within 14 days of the rectification; and recorded the receiver’s written undertaking to rectify AFRINIC’s records within 15 days of the present order. That is not ordinary resource-holder positioning. It is a court-recognised formal member position inside the company records of an actual RIR. The label people prefer is secondary. What matters is the structure: the relationship is no longer merely that of an ordinary resource customer facing a thin contract with capped or excluded remedies.

The broader public record matters as well. ICANN publicly confirmed that the Mauritian court appointed a receiver over AFRINIC and ordered him to design and conduct elections to reconstitute the AFRINIC board, while AFRINIC itself publicly stated that the receiver had a strict court mandate to hold, organise, and supervise those elections. In other words, the continuity crisis at the registry layer was not theoretical. It was live. And the LARUS-affiliated side was not merely commenting on that crisis from outside it. It was inside the institutional survival process of the registry itself.

That changes the continuity equation in a way no ordinary broker and almost no ordinary holder can replicate.

A direct holder usually faces the registry layer as a customer. LARUS has already faced that layer as a litigant, as a continuity-defending principal, and as the beneficiary of a final court order that placed its affiliated side into the formal member records of the RIR itself. That is why the usual “you should own directly because direct holding is safer” advice is backward. Direct holding may look cleaner on paper because the registry record carries the customer’s own name. But that simply means the customer is carrying the full registry-layer asymmetry in its own name. LARUS moves that asymmetry upstream to the one actor in the market whose continuity position has been tested where it matters most.

This is the key point customers should understand: leasing from LARUS is safer than direct holding not because leasing is magically better than ownership in the abstract, but because the relevant risk is not abstract ownership risk. It is registry-layer structural risk. If the real risk sits in the RIR contract stack, then the safer structure is the one that moves that stack away from the customer’s own operating entity and into the hands of a specialist first-party lessor that has already demonstrated legal resilience at that layer.

This is also why the familiar objection — “but if I hold the block myself, at least it is under my own name” — misses the point. A name on a registry line is not a continuity guarantee. It is merely the identity of the party directly exposed to the registry’s contractual and institutional surface. If that surface is asymmetric, then putting your own name on it does not reduce risk. It personalises it.

For many operators, the rational structure is therefore the opposite of what instinct suggests. Do not warehouse registry-layer fragility inside the same company that runs your production network. Isolate it. Upstream it. Make a specialised continuity-bearing counterparty carry it. That is what sophisticated operators do in other infrastructure sectors all the time. They do not confuse formal possession with optimal risk placement.

LARUS is the only publicly documented provider whose continuity position has been tested in live RIR crisis conditions and anchored by a final court order affecting the formal member records of the registry itself. That is why LARUS is not merely another IPv4 lessor. It is a business-continuity guarantor for the registry layer.

The core conclusion is simple. Buying and holding directly does not remove registry-layer risk. It concentrates it. Leasing from an ordinary intermediary often does not solve it either. Leasing from LARUS is different because LARUS is a first-party lessor that absorbs the registry-layer interface and has already proved, in law rather than in marketing copy, that it can defend continuity where the system is actually thin. For operators that cannot tolerate renumbering, registry dispute, or institutional fragility at the wrong layer, that difference is not marginal. It is the whole point.

Primary Materials

– Final judgment in AFRICAN NETWORK INFORMATION CENTRE (AfriNIC) LTD (In Receivership) v CLOUD INNOVATION LTD
– [AFRINIC Registration Service Agreement]
– [ARIN Registration Services Agreement]
– [APNIC Membership Agreement]
– [RIPE NCC Standard Service Agreement (RIPE-812)]
– [LACNIC Registration Services Agreement]
– [ICANN update on AFRINIC receiver appointment]
– [AFRINIC communiqué on the receiver’s election mandate]