Summary

  • APNIC's reclamation problem is not finding unused IPv4. It is converting an old reliance state into a new usable state without surprising operators, customers or buyers who relied on the old record.
  • Reuse becomes legitimate when the registry can show a disciplined state transition: notice, cure, proof of control, dispute isolation, fraud reversal, quarantine, routing hygiene, public provenance and narrow reissue rules.
  • Scarcity makes recovered addresses valuable, but that value is only real after hidden liabilities are priced: stale routes, dirty reputation, reverse-DNS residue, corporate succession, payment failure, legal claims and mistaken recovery risk.

The returned-prefix file

Start with the file, not the slogan. In the IANA recovered IPv4 address-space registry, a line for 160.20.241.0 through 160.20.241.255 records a /24 returned by APNIC in August 2012. Its status is not mythical abundance. It is recovered. Elsewhere, APNIC's returned legacy address-space page lists rows of old prefixes that have come back into the registry's reach, including small 160.238.x/24 entries and older 192.x space. The dry administrative fact is that a number block once attached to a prior holder can lose that state, enter a recovery condition, and eventually become part of a usable supply again.

That is the interesting point. A returned block is not a fresh mineral deposit. It is a piece of operational history whose current condition must be established before anyone should rely on it again. It may carry old routing memory, stale contact records, reverse-DNS residue, reputation damage, out-of-date customer documentation, corporate succession ambiguity, legal dispute or fraud traces. It may also be perfectly quiet, voluntarily returned and easy to place back into circulation. The economic work lies in telling the difference.

IPv4 reclamation is attractive because scarcity makes every clean block valuable. A /24 can matter to a small provider, a hosting firm, an enterprise edge, a public service or a network that needs routable independence. A larger recovered block can influence waiting lists, transfer-market expectations, leasing behaviour and the bargaining position of new entrants. Reuse therefore looks efficient. If an address block is unused, why leave it idle while new demand pays scarcity prices?

The answer is that "unused" is not a sufficient registry fact. A block can be unrouted because it is reserved for disaster recovery. It can be quiet because a network is in migration. It can be poorly documented because a company merged, changed names or lost administrative staff. It can be unpaid because of sanctions exposure, currency controls, bank failure, inheritance confusion or ordinary neglect. It can be fraud-tainted because the wrong party changed the record. It can be disputed because two successors claim the same asset. None of those states is equivalent to clean abandonment.

APNIC's task is therefore not to convert every quiet address into regional stock. It is to run a narrow state machine that protects uniqueness and continuity while allowing dead or properly returned resources to be used again. The registry must be able to say what happened to the old state, what evidence supports the change, who had notice, what cure period was offered, whether a dispute remains, whether operational residue was cleaned, and what limits apply before reissue. Without that discipline, reclamation becomes a tempting word for capital control. With it, reclamation can increase usable supply without making every holder fear that silence, weakness or administrative delay will be read as surrender.

This distinction is especially important in Asia-Pacific. APNIC's region includes large incumbents, small island providers, fast-growing access markets, hosting centres, national registry arrangements, historical resources, public-sector networks and cross-border service chains. A recovered address can move through commercial environments with very different legal and operational assumptions. The same registry action that looks obvious in a dormant record can look arbitrary if it touches a running network, a successor company or a holder caught in a banking or documentation problem. Scarcity raises the value of recovery; it also raises the cost of error.

Reclamation is a state transition

The useful vocabulary is state transition. An IPv4 resource can move from delegated to returned, from delegated to reclaimed, from reclaimed to reserved, from reserved to available, from available to delegated, or into quarantine when usability is uncertain. APNIC itself describes address states for registry accuracy and pool management, including reserved, available, delegated and quarantined conditions. The state language matters because it keeps the debate away from moral theatre. The registry is not deciding whether a holder is virtuous. It is deciding whether the public record can safely move from one operational condition to another.

That seems bureaucratic. It is actually the safeguard. A state transition forces the registry to ask narrow questions. Who was the recorded holder? What contractual or policy basis applies? Is the account closed? Were resources transferred? Did a voluntary return occur? Has a historical resource holder supplied proof of registered use or control? Is there evidence of fraud? Is there a court order, insolvency process or corporate succession record? Has APNIC tried to contact the holder through reliable channels? Has a cure period passed? Are there live routes, reverse DNS, RPKI objects or customer dependencies that contradict abandonment? Is there a competing claim that should freeze the state rather than resolve it by administrative force?

Each question is dull in isolation. Together they protect reliance. An address block is not merely an entry in a database. It can be embedded in firewall rules, bank allowlists, cloud access lists, customer contracts, mail systems, remote-access products, security logs, route filters and procurement documents. If a registry changes the state without adequate process, the economic harm may appear far from the registry portal. A customer may lose reachability, a transfer buyer may inherit uncertainty, a provider may suffer a reputation problem, or a successor company may find that an asset vanished during a corporate transition.

State-transition thinking also protects APNIC. A registry that makes recovery decisions through clear categories is less exposed to accusations that it is using scarcity as a discretionary weapon. If a block is voluntarily returned, the evidence is different from non-payment. If an account closes, the rule is different from a disputed merger. If a transfer was fraudulent, reversal serves record integrity rather than redistribution. If a resource is historically registered but inactive, the registry's demand for proof of control must be evidence-based and proportionate. If a block is in quarantine, the problem is usability, not ownership morality.

This is the narrow registry doctrine applied to scarcity. The registry exists to protect uniqueness, accuracy, contactability, security assertions, transfer records and continuity. Those functions justify a recovery process when the record no longer corresponds to a legitimate controller. They do not justify an open-ended power to decide that a holder uses too little, earns too much, serves the wrong customers or should release capital for regional policy goals. Scarcity narrows the duty because the asset becomes more relied upon, not less.

APNIC's recovery surface is real but bounded

APNIC has several factual hooks for recovery. Its policy material states that when an account is closed, address space and autonomous system numbers held by the member must be returned, and resources not transferred before closure may become subject to the recovered-pool rules if the criteria are met. It also has a policy basis for seeking recovery of unused historical resources for eventual redelegation. The recovered-pool section describes categories of IPv4 that can become available, including returned legacy resources and other IPv4 that becomes available to APNIC, and APNIC's address-state guidance describes reclaimed resources being reserved for a period before becoming available in certain circumstances.

Those hooks are significant. They mean APNIC is not merely a passive museum of old allocations. The registry can receive returns, close accounts, record changed states, reserve recovered resources, quarantine suspect resources and distribute recovered stock under policy. In a scarce market, that has economic force. It affects who can obtain small amounts of address space without buying on the transfer market, how much value sits in dormant holdings, and how confidently recipients can rely on a recovered block.

The hooks are also bounded. They do not transform APNIC into a regional landlord of every idle address. A member account closure is not the same fact as a running company with low utilisation. A voluntary return is not the same fact as silence. A fraud reversal is not the same fact as punishment. A historical-resource evidence request is not the same as compulsory membership. A quarantine period is not a confiscation finding. Each mechanism has its own legal and economic meaning.

That boundedness is what keeps recovery from corroding trust. Holders tolerate registry administration because it protects a common coordination layer. They do not accept that a coordination layer may convert scarcity into a general right to inspect business plans, second-guess reserves or seize underused capital. If APNIC wants a recovered block to be trusted after reissue, it must also make the recovery trusted before reissue. The market will price the difference.

The market's price signal is not abstract. A recipient of recovered space wants to know whether the block will route cleanly, whether its prior holder could reappear, whether the prefix has mail or abuse damage, whether reverse DNS is coherent, whether route filters recognise it, whether the registry record is stable, and whether any dispute flag could interrupt service. A buyer or operator will discount uncertainty. A lender, acquirer or enterprise customer may discount it further. The recovered block's value depends on proof that its old state ended cleanly.

APNIC's public recovered-pool and quarantine language points in the right direction because it treats recovered resources as administratively distinct before redelegation. But the economic standard should be higher than "the registry has stock". The standard should be "the registry can explain the state history in enough detail that a new holder can rely on the block and the old holder, if any, has had a fair route to cure or contest." That is the difference between reuse and churn.

Five different problems should not be merged

Reclamation becomes dangerous when five different problems are merged into one word: unused, abandoned, unpaid, disputed and fraud-tainted. They all may end with a resource returning to the registry. They should not begin with the same presumption.

Unused resources are the most politically seductive category. A quiet route table makes it easy to say that the public would benefit if addresses were moved to someone else. But non-use is not the same as non-control. A holder may keep addresses for redundancy, future expansion, customer migration, regulatory segmentation, security isolation, merger integration, renumbering avoidance, disaster recovery or contractual commitments. It may also be inefficient, speculative or simply stale. The registry cannot know enough from utilisation alone to make a capital-control decision. It may ask for record accuracy and proof of control where policy allows. It should not treat low visible use as forfeiture.

Abandoned resources are different. Here the problem is not that use is low, but that no credible controller can be found or the recorded holder no longer exists in a way that can maintain the registry state. The evidence may include failed contacts, dissolved corporate records, returned correspondence, dead administrative channels, absence of routing over a long period, no successor claim and no payment or service relationship. Even then, abandonment should be a conclusion after process, not a label applied at the first sign of silence. The old state should pass through notice, waiting, reservation and dispute windows before the block becomes ordinary supply.

Unpaid resources are different again. Payment failure can signal abandonment, but it can also signal ordinary administration failure, sanctions exposure, currency problems, bank de-risking, a disputed invoice, insolvency, merger confusion or contact breakdown. In a region as varied as Asia-Pacific, non-payment should not become an instant economic death sentence. If an account relationship ends and policy requires return, APNIC has a basis to act. But the process should distinguish deliberate default from curable failure, and it should preserve enough time for a genuine holder to restore contact or transfer resources rather than lose them by surprise.

Disputed resources require restraint. Two entities may claim succession after a merger, acquisition, insolvency, founder dispute, family-business split or court proceeding. The registry may need to freeze changes, preserve the status quo, mark a dispute and require credible documentation. It should not resolve private ownership by choosing the party it prefers. The narrow role is to protect uniqueness, prevent duplicate or fraudulent changes, and keep public records from making harm worse while the proper decision-maker settles the underlying conflict. In a dispute, the registry's power is most legitimate when it is conservative.

Fraud-tainted resources are the category where registry intervention can be strongest. If a record was changed through false authority, forged documents, account compromise or impersonation, recovery protects the ledger itself. But even here, precision matters. The target is the fraudulent state, not commerce as such. A good-faith recipient, downstream customer or network that relied on an apparently valid state may need notice and transition handling even if the registry must reverse the underlying change. Fraud control is essential; so is false-positive control. A registry that can undo fraud without creating general seizure fear strengthens the market. A registry that treats fraud suspicion as a blank cheque weakens it.

Keeping these categories separate is not procedural nicety. It is the economics of trust. When operators know which state they are in, they can act. They can cure, document, transfer, contest, clean up, price risk or stop relying on the block. When every concern is collapsed into a vague recovery power, no one can price the risk except by discounting the registry environment itself.

The old state must end before the new state can be trusted

A recovered address has two histories. One history looks backward: why did the prior state end? The other looks forward: what must happen before a new holder can rely on the block? The backward history matters first.

For voluntary returns, the end of the old state is comparatively clean. A holder tells the registry it no longer needs the resource, the registry records the return, and any required waiting or quarantine period begins. Even here, the registry should ensure that the person returning the resource has authority and that obvious dependencies are not being missed. But the moral tension is low. Voluntary return is a market-friendly path because it lets holders release surplus without being treated as offenders.

For account closure, the old state depends on contract and policy. If resources must be returned unless transferred before closure, the closure sequence should be legible. The holder should know the consequence, the transfer option, the deadline, the records needed and the effect on related services. The registry should avoid a trap in which a member discovers too late that administrative closure has destroyed a scarce asset. Closing an account is a service-state event; returning address resources is an asset-state event. The two can be linked by policy, but the consequences should be explicit.

For historical resources, the backward history is harder. Many historical allocations predate today's membership assumptions and administrative habits. Records may be old, contacts stale, and corporate structures changed. APNIC's historical-resource maintenance material recognises that data may be updated and that documentation may be needed to prove registered use or control. That is a proper registry concern. But the point of evidence collection is accuracy, not retroactive subordination. A historical holder should not be forced into modern policy categories merely because the registry wants clean stock. If the holder cannot prove control after fair process, recovery may become appropriate. The fairness of that process is the asset's economic foundation.

For insolvency, the old state may end through a liquidator, receiver, court process, asset sale or dissolution. IPv4 addresses have economic value. They may be claimed by creditors, buyers, service customers or successors. A registry should not pretend that an insolvent company is simply absent. Nor should it let a dead record persist forever because insolvency is complicated. The narrow answer is documentation: recognise authorised insolvency representatives, preserve the record while claims are sorted, record transfers when valid, and reclaim only when the resource has no lawful controller or policy basis for continued registration. The registry should not be the bankruptcy court.

For fraud, the old state ends by proving that the apparent state was never legitimate. That is a different kind of transition. The registry may need to restore a prior record, suspend changes, quarantine a prefix or coordinate with affected networks. The more severe the action, the more important the evidentiary standard and the appeal path. Fraud control without appeal can itself become a fraud vector if a determined party can persuade the registry to freeze or reverse a competitor's resource through weak evidence.

Only after the old state has ended should the new state begin. That sounds obvious. In practice, scarcity creates pressure to compress the sequence. A block looks idle; demand is waiting; the registry has a pool; the public value of reuse seems clear. The temptation is to let the future use justify a thin inquiry into the past. That is backwards. Reuse is legitimate because the prior state ended properly. It is not proper because reuse is attractive.

Quarantine prices hidden liabilities

APNIC's policy material recognises quarantine for recovered resources, with the period determined by usability problems such as routability, reachability, security or other concerns. Quarantine is not just a waiting room. It is a pricing device for hidden liabilities.

An IPv4 block can be dirty in several ways. It may have been used for spam, scanning, malware, bulletproof hosting, fraud, evasive VPN services or abusive customer traffic. It may sit on blocklists that update slowly or never. Its geolocation may be wrong. Its reverse DNS may point to an old operator. Its route history may cause filters to mistrust a new origin. Its RPKI or routing-adjacent objects may need cleanup. Its abuse contacts may be stale across many third-party systems. It may be associated in customer documents with a company that no longer controls it. A new holder can inherit all of this even if the registry record is clean.

Quarantine makes the registry pause before it turns a recovered asset into someone else's liability. The pause has a cost: addresses sit unused in a scarce market. But immediate reissue can cost more if recipients receive blocks that are technically available but commercially impaired. A dirty /24 can consume staff time, delay customer onboarding, trigger mail problems, complicate security reviews and damage trust in the recovered pool. The usable yield of recovery is therefore lower than the raw number of addresses returned.

This is where reclamation economics differs from conservation rhetoric. The headline stock may say that a number of addresses came back. The usable stock may be smaller after quarantine, fragmentation, reputation repair, route cleanup and dispute reserves. A registry that reports only gross returns overstates supply. A market that ignores condition overprices recovery. The honest measure is not how many addresses were reclaimed; it is how many can be reused without passing old liabilities to new operators.

Quarantine should therefore have both minimum discipline and maximum transparency. APNIC does not need to publish sensitive abuse investigations or private documents. It should, however, be able to describe categories of condition: routability concern, security concern, dispute hold, contact uncertainty, reverse-DNS cleanup, routing-record cleanup, reputation watch or ordinary cooling period. It should also avoid indefinite limbo. A quarantine state that never resolves becomes another form of dead stock. A quarantine state that resolves too quickly transfers risk to the recipient.

The right economic question is not whether quarantine is pro- or anti-market. It is who should bear the uncertainty. If a block is released immediately, the recipient bears it. If a block is held forever, the market bears scarcity. If APNIC runs a disciplined quarantine with condition notes and release criteria, the uncertainty is reduced before it is sold, allocated or otherwise relied upon. That is a legitimate registry function because it protects continuity and record quality rather than deciding who deserves capital.

Reuse changes the transfer market but does not replace it

Recovered IPv4 can relieve scarcity at the margin. It cannot recreate the old free pool. This is important because overstatement creates bad politics. If reclamation is sold as the answer to IPv4 scarcity, it invites broad recovery powers and expectations that every quiet block should be mined for public supply. If it is understood as a narrow state-transition mechanism, it complements transfers, leasing and IPv6 deployment without pretending to abolish the underlying constraint.

The transfer market remains central because most usable IPv4 value sits with holders that still control their resources. Some may sell. Some may lease. Some may hold for operating need or capital appreciation. Some may return resources voluntarily if the cost of maintaining them exceeds the benefit. Reclamation should not be a coercive substitute for transactions. A registry that makes recovery too discretionary can depress legitimate transfers by making buyers fear future challenge and making sellers fear that engagement with the registry exposes them to pressure.

At the same time, recovery can improve the transfer market by cleaning the edges. Fraudulent transfers, abandoned records, dead holders and unmaintained historical entries create uncertainty for everyone. A buyer wants to know that the seller controls the block. A lender wants to know that the record is stable. A network wants to know that a prefix will not be reclaimed after integration. When APNIC removes false states, closes dead ones and publishes clear provenance for recovered stock, it reduces uncertainty across the whole market.

The discipline is to avoid using recovery as price policy. APNIC should not reclaim because transfer prices are high. High prices are a symptom of scarcity and demand. The registry should reclaim because a resource has entered a recoverable state under clear rules. Once recovered, APNIC can distribute under policy. But it should not let the distribution goal infect the recovery finding. The backward and forward halves must stay separate.

This separation matters for capital control. IPv4 is productive capital for many operators. It supports customers, reduces dependence on intermediaries, provides route independence, preserves legacy reachability and carries balance-sheet value in transactions. A registry that can take it because regional demand is high becomes a capital controller. A registry that can record its return, reverse fraud, close dead records and reissue after quarantine remains a ledger institution. The difference is not ideological decoration. It affects investment, lending, acquisition diligence and operational planning.

The APNIC region is especially exposed to this tension because demand is uneven. New entrants and fast-growing networks may see dormant legacy holdings as unfair. Legacy holders may see recovery language as a threat to acquired rights. Small providers may want access to scarce space without paying large market premiums. Large incumbents may prefer certainty over redistribution. A registry that tries to satisfy all these interests through discretion will lose the narrow legitimacy that makes recovery possible. A registry that applies clear state categories can disappoint some parties without becoming arbitrary.

Waiting lists are the wrong centre of gravity

Recovered-pool distribution naturally raises waiting-list and rationing questions. Who can receive recovered IPv4? How much? How often? Under what eligibility limits? APNIC's recovered-pool policy includes small allocation sizes and eligibility constraints. Those rules matter, but they should not become the centre of the reclamation article.

The more important question is upstream: whether the block has become safely reusable. A waiting list cannot cure a defective recovery. A fair queue does not fix a mistaken abandonment finding. A small allocation size does not cleanse bad reputation. A one-time eligibility limit does not resolve a successor dispute. Distribution rules govern scarcity among eligible recipients. Recovery rules govern whether the resource belongs in the pool at all.

Confusing those stages produces institutional drift. The more compelling the waiting demand, the easier it becomes to view old holders as obstacles. The more visible the queue, the more pressure there is to accelerate recovery. The more sympathetic the applicants, the more tempting it is to relax proof against dormant or weak holders. That is how a ledger becomes a gatekeeper. The registry begins with a coordination problem and ends by choosing whose capital should move because someone else needs it more.

APNIC should resist that logic even when the applicants are deserving. Registry legitimacy comes from neutral coordination, not redistribution by sentiment. A small operator may genuinely need IPv4. A public-interest network may genuinely benefit from a recovered /24. A fast-growing access provider may have a strong case for scarce resources. None of that proves that a quiet holder has abandoned its resource. The registry can run an objective recovered-pool distribution after recovery. It should not let distribution demand supply the evidence for recovery.

This is why the phrase "economics of reclamation and reuse" must be read carefully. The economics are not only about efficient allocation. They are about reliance costs, proof costs, error costs, condition discounts, market expectations and the cost of institutional overreach. Efficiency without due process is unstable because holders will spend more on defensive administration, legal protection, transfer structuring and political resistance. The apparent gain from recovered stock can be offset by a higher risk premium across the registry environment.

Historical resources need continuity, not retroactive punishment

Historical resources are the most sensitive part of reclamation because they often sit outside the ordinary story of modern membership. They may be older than current APNIC procedures, tied to organisations that changed name or structure, or recorded with contacts that no longer work. Some may be unused. Some may be misused. Some may be important operating assets. The registry's job is to maintain accurate data, not to punish age.

APNIC's historical maintenance material recognises that updates can occur without necessarily requiring membership and that documentation may be requested to prove registered use or control. That distinction is vital. A historical holder should be able to establish continuity without being told that the old allocation has become morally suspect because it does not fit present-day scarcity politics. The question should be whether the record reflects a real controller who can be contacted, held accountable for registry purposes and protected against duplicate claims.

The risk of retroactive punishment is not theoretical. Scarcity changes how old records are perceived. What once looked like untidy administrative residue can now look like hoarded value. That shift can make recovery advocates impatient with historical ambiguity. But ambiguity is exactly why process matters. The holder may be a university, a public body, a research network, a company that merged, an operator with old documentation or a successor that uses the block sparingly but legitimately. A rushed recovery can destroy continuity that the registry exists to protect.

At the same time, historical status should not become a shield for dead records. If no controller can be found after serious attempts, if corporate existence has ended without successor, if contact and routing evidence are absent, and if no legitimate claim appears during a fair process, the registry should not keep addresses frozen forever. Historical continuity is not immortality. It is a presumption that old reliance deserves careful handling before the resource changes state.

The balance can be made practical. APNIC can maintain a historical-resource evidence ladder: contact update, authority proof, corporate succession proof, network-use evidence, dispute flag, public notice where appropriate, reserved status, quarantine and eventual availability. The ladder should be clear enough that a genuine holder knows how to respond and a dead record can eventually move. The point is not to make recovery impossible. It is to make it boring, reviewable and resistant to political pressure.

Insolvency and corporate succession are registry tests

Corporate failure is one of the places where reclamation can either preserve value or destroy it. A company may enter liquidation while its IPv4 resources remain valuable. Creditors may expect sale proceeds. Customers may still depend on services. A buyer may acquire network assets. Directors may be unavailable. The corporate registry may show dissolution while the network reality is messier. A regional Internet registry sits at the edge of this process, not above it.

The narrow approach is continuity. APNIC should recognise lawful representatives, require documentation, preserve the status quo when claims conflict, and record valid transfers. If no lawful controller exists and policy supports recovery, then the resource can move toward return, reservation, quarantine and availability. But the registry should avoid treating insolvency as abandonment by default. Insolvency is often a process for preserving and distributing value, not proof that value has ceased.

This matters because IPv4 is now part of corporate economics. It can affect acquisition price, creditor recovery, customer continuity and operational migration. A buyer of a failed hosting business may care as much about address continuity as about servers. A customer may care that its service remains reachable during a sale. A creditor may care whether address assets can be transferred. If APNIC reclaims too early, it may interfere with lawful value realisation. If it waits indefinitely, it may preserve false claims. The registry's legitimacy lies in documented thresholds.

Corporate succession creates a similar challenge without formal insolvency. A merger, spin-off, name change, government restructuring, university reorganisation or family-business split may leave old address records mismatched to current legal reality. The registry should not read mismatch as abandonment when evidence can show continuity. Nor should it accept every claim without proof. A disciplined record-update process is itself anti-fraud infrastructure.

The economic lesson is that recovery starts long before recovery. The better APNIC's ordinary record-maintenance and transfer documentation, the fewer cases fall into the grey zone where reclamation is tempting but risky. Accurate contacts, current authority records, clean transfer logs and visible dispute states reduce the pool of apparently abandoned resources. That may produce fewer dramatic recoveries, but more reliable registry value.

Fraud control must be sharp, not broad

Fraud-tainted resources are the case for strong registry action. If someone obtains a transfer through false documents, compromises an account, impersonates a holder or manipulates stale contact data, APNIC must be able to intervene. A registry that cannot reverse fraud is not protecting the ledger. It is publishing an attack surface.

But fraud control must be sharp. It should target the false act and the state it created. It should not become a general suspicion doctrine that makes any unusual transfer, inactive holder, foreign buyer or leasing arrangement vulnerable to administrative dislike. Scarcity increases both the incentive to commit fraud and the incentive to accuse competitors of fraud. The registry therefore needs evidence standards, audit trails, temporary freezes, appeal channels and proportional responses.

The proportionality problem is practical. Suppose a block is transferred through a document later alleged to be forged. A new holder may already have routed it. Customers may depend on it. A downstream buyer or network may have acted in good faith. The original holder may be real, dead, compromised or disputed. The registry may need to freeze further changes, mark the dispute, investigate, restore records or coordinate a transition. A sudden public-state change can harm innocent parties. A slow response can let fraud harden into reliance. The only credible answer is a known emergency process with review.

Fraud-tainted blocks also need condition handling after recovery. A fraud episode may have involved abusive hosting, hijacked routes, forged route objects or rapid resale. Even when the title problem is fixed, the block's reputation and routing condition may remain impaired. Quarantine should not be seen as punishment. It is part of restoring usability.

This is where registry power is easiest to justify and easiest to overextend. Everyone wants fraud stopped. The danger is that fraud language can be used to smuggle in policy goals that would not survive direct argument: discouraging transfers, discouraging leasing, discouraging offshore holding, disciplining disliked business models or pressuring holders to maintain membership. APNIC's best protection against that drift is specificity. What was false? Which record changed? Which evidence proves it? Which state is being restored? Which parties receive notice? What happens if the finding is wrong?

Sharp fraud control supports markets. Broad suspicion weakens them. The difference is whether participants can predict the registry's response before they invest, route or transact.

Reissue should carry provenance, not stigma

When a recovered block is finally reissued, it should carry enough provenance for reliance and little enough stigma for practical use. New holders do not need a public archive of every private document behind the recovery. They do need confidence that the prefix did not appear from an opaque administrative act that can be challenged tomorrow.

Provenance can be simple. The resource was voluntarily returned, reclaimed after account closure, recovered from a dead historical record, restored after fraud, released from dispute, or made available after quarantine. The relevant dates matter: recovered, reserved, quarantined, released, delegated. The condition class matters: ordinary, security watch, routability watch, dispute cleared, reputation concern resolved. The contact and authority record must be current. Routing-adjacent records should not contradict the new state. Reverse-DNS delegation should not point to the past. If RPKI or route data existed, their transition should not surprise operators.

The goal is not to brand the block as second-hand forever. Most IPv4 is second-hand in a meaningful economic sense now. Transfers, mergers, leases and corporate reorganisations all move address resources through time. The goal is to make the second life reliable. A clean recovered block should become ordinary supply after the required process. A questionable block should not be pushed into use merely because demand is strong.

Provenance also protects APNIC against future disputes. If a prior claimant appears, the registry can show the sequence: contact attempts, evidence requests, cure period, reservation, quarantine and release. If a new holder faces a third-party reputation problem, the registry can show when the state changed and what cleanup was performed. If the community questions recovered-pool allocation, APNIC can distinguish recovery evidence from distribution choice. Transparency reduces the suspicion that scarce resources moved by favour.

There is a balance. Too much public detail can expose security investigations, private corporate documents or personal contact data. Too little detail leaves the market guessing. The right level is state provenance rather than case gossip. Publish the category and dates; protect sensitive evidence; preserve the audit trail for review. That is how a registry can be transparent without becoming careless.

APNIC's narrow power is also its economic advantage

There is a temptation to treat APNIC's narrowness as weakness. If the registry cannot force all idle addresses back into circulation, cannot decide who deserves IPv4 most, cannot set market prices and cannot rewrite private reliance, then what good is it? The answer is that narrow power is precisely what makes the registry useful.

Number-resource markets need a public coordination layer that participants trust. They need uniqueness. They need stable records. They need proof of control. They need fraud resistance. They need clear transfer recording. They need contactability. They need continuity when institutions fail. They need security assertions that are not casually destroyed. They need dispute states that do not turn administrative convenience into private loss. These are not minor functions. They are the conditions under which scarce IPv4 can circulate without constant litigation or operational panic.

If APNIC becomes broader than that, it may gain apparent policy power and lose economic trust. Holders will treat every interaction as a risk. Transfers will be structured defensively. Historical holders will avoid updating records for fear of inviting scrutiny. Buyers will discount blocks that passed through uncertain recovery. Small operators will worry that membership or payment trouble can become asset loss. Governments may see the registry as a pressure point. The scarce asset becomes less liquid because the coordination layer looks discretionary.

By contrast, a narrow APNIC can make recovered space more valuable. If the market believes that recovery follows clear evidence, that quarantine has meaning, that old claims were handled fairly, and that reissue provenance is stable, recipients can rely on the block. That reliability is economic output. The registry has increased usable supply not merely by moving numbers, but by lowering uncertainty.

This is the ledger-not-gatekeeper boundary. The registry record describes and protects reality; it should not invent ownership consequences beyond its coordination mandate. It may record that a resource was returned. It may require proof that a claimant controls it. It may reverse fraud. It may reserve, quarantine and release. It may refuse to publish a state that would create duplicate control. It may not decide that capital should move because the region has better uses for it.

Scarcity makes that boundary harder to hold and more important to hold. When addresses were abundant, recovery mistakes were less costly because alternatives existed. Under scarcity, a mistaken recovery can destroy real value. It can also create a political precedent. The first error may be treated as an exception. The second becomes a tool. The market notices.

What a good APNIC recovery regime would show

A good recovery regime would be legible without being theatrical. It would show resource states clearly enough for holders, recipients and observers to understand what can happen. It would distinguish returned, reclaimed, reserved, quarantined, disputed and available resources. It would publish aggregate statistics that separate gross recovered stock from usable released stock. It would identify the broad reason category without exposing sensitive documents. It would record dates for state changes. It would provide a cure path for curable failures and an appeal path for contested findings.

It would also separate evidence categories. Non-payment would not be treated as abandonment until notice and cure failed. Historical ambiguity would not be treated as disuse until control evidence was requested and assessed. Fraud suspicion would produce a temporary protective state before irreversible action. Insolvency would trigger representative verification rather than automatic recovery. Voluntary return would remain the cleanest path and should be made administratively easy. Disputes would be isolated rather than solved by administrative preference.

For quarantine, a good regime would state what problem is being assessed: routability, security, reputation, dispute, reverse DNS, routing records or other usability concerns. It would have release criteria. It would avoid both immediate reissue of dirty space and indefinite warehouse treatment. It would make clear whether the recipient receives any warning about condition. It would track whether released recovered blocks later produce abnormal complaints, because that is the test of quarantine effectiveness.

For reissue, a good regime would avoid implying that recipients have received charity or politically favoured stock. They have received a scarce resource through a defined process. The new state should be ordinary once conditions are met. Allocation limits and eligibility rules can ration scarce supply, but the registry should not use reissue to make social claims about who is more deserving than lawful holders. Deserving applicants matter at the distribution stage; they do not supply evidence at the recovery stage.

For records, a good regime would keep old and new states connected enough to avoid confusion. A recipient should not discover that third-party systems still associate the block with the prior holder because basic cleanup was neglected. Nor should the public record expose private facts unnecessarily. The practical standard is operational continuity: enough information to route, contact, verify and rely.

The hardest feature may be error correction. A registry that recovers resources should admit the possibility of mistake. What if a holder reappears with credible evidence after release? What if a fraud finding was wrong? What if an insolvency representative was not authorised? What if a block was reissued before a court order was known? There should be a remedy framework that protects the new holder's reliance while addressing the old holder's claim. Hard cases will remain hard. A known remedy is still better than pretending errors cannot occur.

The abandoned block is a public test of institutional character

The abandoned block tests what kind of institution APNIC wants to be under scarcity. One path is expansion. Scarcity is valuable; idle resources offend demand; the registry has community responsibilities; therefore recovery can become broad, discretionary and moralised. That path will find supporters, especially among applicants who need addresses and observers frustrated by old holdings. It will also create fear, because every holder will wonder when administrative weakness becomes evidence of social uselessness.

The other path is narrow continuity. Scarcity is valuable; therefore errors are expensive. Idle-looking resources may be abandoned, but they may also be relied upon in ways the registry cannot see. Recovery is necessary, but it must be evidence-based, state-specific and reviewable. Reuse is good, but only after the old state has ended and the block's condition is known. The registry's power grows more legitimate when it is constrained.

APNIC's own policy surface contains pieces of the narrow path: account-closure return rules, historical-resource recovery, recovered-pool distribution, address-state labels and quarantine for usability problems. The challenge is to make those pieces operate as a coherent economic discipline. A recovered block should not be a trophy. It should be a file that survived proof, notice, condition assessment and release.

That may sound slow. In reality, it is what makes reuse fast enough to matter. Markets move faster when participants trust the rules. Recipients can deploy recovered space faster when they believe the old state cannot credibly return. Holders can update records faster when they do not fear that contact with the registry invites confiscation. Buyers can transact faster when recovery and transfer records are reliable. Small applicants can rely on recovered-pool allocations when the prefixes are not poisoned by old disputes. Due process is not the enemy of reuse. It is the mechanism that turns recovered stock into usable stock.

The returned-prefix file therefore tells a larger story than its rows suggest. A /24 returned by APNIC, a legacy block listed as returned, a reclaimed resource moving through reserved or quarantined state: these are not merely administrative crumbs. They are evidence of how a scarce coordination layer handles time. IPv4 did not stop being valuable when the first holder stopped using it. It did not become public spoil when a contact failed. It did not become clean because demand existed. It became usable again only when the registry could prove that continuity was protected while state changed.

That is the economics of reclamation and reuse. The prize is not the largest possible recovery number. The prize is reliable second-life address space. APNIC should recover what is truly returned, abandoned, unpaid beyond cure, fraud-tainted or legally unclaimed. It should quarantine what is dirty, freeze what is disputed, preserve what is relied upon, and publish enough state provenance for the next holder to trust. If it does that, reclamation adds supply without turning the registry into a capital controller. If it fails, every recovered block will carry an invisible question: not whether it can be routed, but whether the institution that released it might one day do the same thing again to someone else.

Sources and Further Reading