Summary

  • A /16 fixes quantity at 65,536 IPv4 addresses. It does not fix the recognised holder, transfer path, prior allocation history, routeability as one aggregate, number of more-specific announcements, RPKI state, reputation, geolocation, customer encumbrances, remediation cost or time to usable delivery.
  • Clean is not a universal technical state. A block can be absent from one public blocklist and still carry poor private mail reputation, stale geolocation, conflicting route objects, an overbroad Route Origin Authorization, unresolved abuse history or a registration-chain defect. Conversely, a policy-list entry may describe intended network type rather than malicious conduct.
  • Public RIR transfer logs are valuable for completed administrative events but omit commercial consideration and many failed, pending or private arrangements. Broker reports provide timely price observations, yet they cover a provider's own transactions and inventory. Neither source supports a global price for every /16 or a universal quality premium.
  • History matters because address reputation and operational records outlive a change in commercial control. Fragmentation matters twice: a /16 may be sold as one block or many pieces, and it may be announced as one aggregate or many more specifics. Those forms affect buyer fit, transaction cost, route-table burden, cutover work and possible revenue.
  • Regional policy matters because an ARIN-to-APNIC case, a transfer within the RIPE NCC service region and an APNIC historical-resource transfer can require different evidence, recipient conditions, fees, restriction periods and post-transfer object changes. Globally routable addresses are not globally interchangeable administrative instruments.
  • Routing state must be disclosed as a time series, not a screenshot. Expected origins, more specifics, visibility, ROAs, Internet Routing Registry objects, reverse DNS and geofeed information each answer a different question and each has coverage limits.
  • Number Resource Society can standardise a portable IPv4 quality dossier and comparable-price bridge without declaring a block clean, setting a reference price, certifying title or turning selected broker data into a market census.

A /16 is arithmetic, not a commodity grade

CIDR notation makes quantity look like identity. A /16 contains 65,536 addresses because sixteen bits remain available after the sixteen-bit network prefix. If two sellers each offer one /16, a spreadsheet can place the same address count in both rows and calculate a price per address. That normalization is useful. It is not valuation.

Commodity markets develop grades because equal weight does not imply equal deliverability or quality. IPv4 has no widely accepted equivalent of a warehouse receipt tied to a standard grade. The object being negotiated is a bundle: a defined prefix, a private legal transaction, a recognised registration change, a set of operational authorisations, a history of use and a future deployment plan. The bundle can be assembled well or badly.

The prefix itself may differ in ways that do not appear in the slash length. One /16 may have remained registered to the same organisation for decades with a simple corporate chain. Another may have passed through mergers, name changes, dormant subsidiaries and a court process. One may be unannounced and operationally quiet. Another may be covered by thousands of historical observations across more-specific routes. One may have current Route Origin Authorizations aligned with the seller's origins. Another may carry a broad authorisation that must be replaced carefully at closing.

Commercial form also differs. A buyer that needs one aggregate for a global service may value a contiguous /16 and discount a bundle of 256 unrelated /24s. A marketplace serving smaller buyers may earn more in aggregate by subdividing a /16, while accepting longer sales time, more counterparties, repeated transfer work and higher servicing cost. The address count is equal; the product is not.

Intended use changes the answer again. An access network may focus on transferability, routing acceptance and geolocation. A cloud platform may care about customer isolation, private abuse scores and rapid provisioning. A mail provider may assign unusual weight to sender reputation, forward and reverse DNS, warming history and the effects of prior shared use. An investor financing leases may focus on contract coverage, utilisation, customer concentration and substitution rights.

This is why the phrase clean /16 should be treated as a claim requiring a schedule, not as a category requiring no explanation. Clean for what use, measured by which sources, over what period, with what exceptions and as of what timestamp? Without those qualifiers, the word creates a price impression while hiding the work necessary to make the block usable.

The market needs comparability, not false fungibility. Price per address is the starting denominator. The valuation begins when the differences are disclosed.

The public price record has a missing denominator

The most authoritative public records of IPv4 transfers do not normally contain price. RIR feeds identify resources, source and recipient organisations, registry paths, dates and broad transfer types. Field coverage varies. They are administrative records of recognised events, not sale ledgers.

RIPE NCC's public list, for example, identifies the original block, transferred blocks, source, recipient, country codes, transfer type and date. Its policy requires publication of approved changes and distinguishes policy transfers from changes in business structure. APNIC's transfer-log format includes the resource, organisations, economies, source and recipient RIRs, previous delegation date, transfer date and type. ARIN publishes a common-format transfer feed and process statistics.

None of these core formats supplies the agreed purchase price, broker fee, tax, escrow cost, failed bids, remediation reserve or whether the transaction was a sale rather than another permitted commercial arrangement.

This omission is defensible. Consideration can be confidential, and registries do not need to become price-reporting agencies to maintain accurate number-resource records. But the omission limits what researchers and buyers can infer. A row showing one /16 transferred on a date cannot be matched to a public dollar amount unless another reliable source identifies the same commercial case.

Broker and marketplace reports fill part of the gap. IPv4.Global publishes monthly pricing observations and commentary by block size. Its archived reports have described periods when large and small block prices converged, periods when /16 supply pushed larger-block prices down, and wide ranges within a month. Those observations are useful because they come from actual market participation and appear more frequently than formal academic studies.

They are not a global census. The provider sees its own completed sales, auctions, inventory and customer mix. Direct transactions, other brokers, private leases, failed negotiations and off-market corporate restructurings may be absent. Published averages can combine blocks with different histories and terms. An asking price is not necessarily realised consideration. A price reported at contract may differ from the seller's net proceeds or the buyer's all-in cost.

APNIC's annual Internet-address reviews have used identified IPv4.Global data to discuss price over time while separately analysing RIR transfer logs. That combination is informative when its boundary is retained: one source describes a market participant's observed price sample, while the other describes recognised transfer events. Joining the charts does not create row-level price coverage.

The unavailable denominator is the number and address volume of all negotiated IPv4 cases, including completed, rejected, withdrawn, delayed, converted-to-lease and privately settled cases, with comparable all-in prices and condition data. Without it, no one can state the global average premium for a clean /16 or the universal discount for a specific defect.

A responsible valuation can still use evidence. It should label the sample, date, block-size cohort, region, transaction stage and price concept. It should show a range and the reasons for moving within or outside that range. What it must not do is take one broker's monthly figure, multiply by 65,536 and call the result the price of every /16.

Clean is a vector, not a verdict

The word clean often collapses unrelated questions. A useful quality review keeps them separate.

Registration cleanliness asks whether the seller is the current recognised holder, the legal name matches, the account is accessible, the chain of merger or acquisition evidence is coherent, fees are current and no dispute or lock is known. RDAP or WHOIS can show current public registration data. It cannot reveal every private purchase agreement, corporate succession document or pending ticket.

Transfer cleanliness asks whether the exact source-recipient path is permitted now. A resource can have accurate registration and still be subject to a holding period, recipient-use condition, reserved-pool restriction, voluntary lock, sanctions constraint or incompatible inter-RIR route. Approval also depends on complete and authenticated submissions. A clean public record is not pre-approval.

Routing cleanliness asks whether the prefix and expected more specifics are announced by intended origin ASNs, seen with sufficient visibility, free of conflicting active origins and compatible with route filters. BGP observations come from collectors and peers. Absence from one collector is not proof of global absence; visibility at collectors is not proof that every network accepts the route.

RPKI cleanliness asks whether valid Route Origin Authorizations cover the intended origins and prefix lengths without authorising stale or unintended origins. A route can be Valid, Invalid or NotFound under origin-validation logic. Those are technical states, not ratings of the asset or seller. A NotFound route is not automatically malicious. An Invalid route can arise from an origin mismatch or a more-specific exceeding maxLength, including a transition error.

Reputation cleanliness asks how public and private services treat the addresses for a stated use. DNS-based blocklists have different policies and removal processes. Mail providers maintain their own observations. Fraud, proxy, threat-intelligence and cloud-abuse systems may be private. A provider can update at its own cadence and may withhold features to prevent evasion. There is no complete global reputation database.

Geolocation cleanliness asks whether major services map the addresses consistently enough for the buyer's use and whether evidence supports correction. Geofeeds can publish deployment location, optionally authenticated through RPKI, but downstream providers decide whether and when to ingest them. Registration country, holder address, route origin and user location are not the same fact.

Commercial cleanliness asks whether leases, letters of authorisation, customer notices, security interests, liens, rights of first refusal, financing covenants or service commitments constrain delivery. None of these need appear in public registration.

Security cleanliness asks whether credentials, abuse contacts, reverse DNS, route objects and change procedures can be handed over without creating a hijack or outage. A portfolio can pass every historical reputation check and still fail this operational test.

A seller should be able to say that no material exception was found within a defined set of tests as of a date. It should not promise permanent universal cleanliness. Every service changes, every address can be abused after closing, and some relevant datasets are inaccessible to both parties.

The price consequence follows the vector. A buyer discounts expected remediation cost, delay, restricted use and uncertainty. It may pay a premium for evidence and fast delivery. The same finding can matter differently to another buyer. That is why a single grade hides more than it reveals.

History can survive the holder

IPv4 addresses are reused identifiers. Their registration can change while third-party systems retain observations from the previous operator. This creates a reputation tail.

Public research has documented the problem in bounded settings. Studies of public blocklists, cloud address reuse and transferred or leased prefixes show that old behaviour can remain associated with an address after operational control changes. The methods and samples differ, so their percentages cannot be combined into a global contamination rate. Their shared contribution is causal plausibility: an address history can affect a successor who did not create it.

Mail provides the clearest operational example. Google's current sender guidance says that the activity of senders sharing an IP can affect the reputation of all senders using it. It requires forward and reverse DNS consistency and describes IP and domain reputation through Postmaster Tools. It also notes that improvement can take time to appear. That is one receiving ecosystem, not the whole Internet, but it shows why a buyer intending to send mail values history differently from a buyer using the addresses only for outbound web clients.

Spamhaus illustrates why even public list status needs interpretation. Its checker identifies different lists, explains possible impact and directs users through list-specific remediation. Its Policy Blocklist expressly notes that inclusion does not necessarily mean an address is bad; the list can indicate ranges that should not send unauthenticated mail directly. Calling every listed address dirty would turn policy classification into a misconduct finding.

Historical routing matters too. A /16 may have been originated by one stable network, split across customer origins, intermittently hijacked, parked behind an AS 0 ROA, or left unannounced. RIPEstat's routing-history endpoint can show origins and timelines observed by the RIPE Routing Information Service. Its visibility is defined relative to RIS full-table peers at a time. That makes it useful evidence, not omniscience.

Registration history can reveal age and prior delegation in some transfer feeds, but a date is not a moral score. Legacy status may imply a different agreement history or service position; it does not prove cleaner routing. A recent allocation may have excellent records and poor use history. An old block may have a simple chain or decades of complex reassignment.

The valuation question is not whether history exists. It is whether the buyer can identify material observations, attribute them to time and operator where possible, remediate current effects and avoid paying twice for the same risk. A blocklist entry already reflected in a mail-use exclusion should not receive an arbitrary second reputation discount. A past hijack corrected years ago should not be treated like an active conflicting origin.

Evidence should be time-indexed around changes of control. The seller's baseline, the transfer request, the authoritative record update, route cutover and buyer acceptance each need dates. A service observation made between contract and cutover may belong to either party depending on who controlled the route and use.

History affects price because it affects future cost and probability, not because addresses acquire guilt. The buyer pays for expected utility. The dossier must make that utility testable.

Fragmentation changes both product and route

Fragmentation has two meanings in an IPv4 transaction, and both influence value.

Commercial fragmentation occurs when a large block is divided into smaller sale lots. A /16 can mathematically become two /17s, four /18s, sixteen /20s or 256 /24s. Subject to policy and registration constraints, subdivision can reach a broader buyer population and may produce a higher aggregate price per address when small-block demand is stronger. IPv4.Global's reports have repeatedly described changing spreads between large and small blocks and the incentive to split when large-block supply is heavy.

The apparent arbitrage is not free. Each lot may require a buyer, contract, diligence review, escrow, transfer request, fee, registry update and post-transfer cutover. The seller carries unsold inventory and must prevent overlapping claims. More counterparties create more sanctions, tax, authority and payment checks. If the original /16 supports customers, subdivision must preserve exact lease boundaries. A higher gross price can yield a lower or slower net result after transaction cost and risk.

Aggregation value also belongs to the buyer. A network needing one /16 may prefer one contiguous registration and one aggregate announcement. Acquiring 256 /24s from unrelated sources creates 256 chains, transfer paths and potential exception sets. Even if each /24 is individually usable, the bundle is not equivalent to an aggregate under the buyer's routing, automation or governance model.

Routing fragmentation concerns how the address space is announced. A holder may advertise the /16 aggregate, more-specific prefixes, or both. More specifics can support traffic engineering, multi-homing and customer delegation. They also increase route-object, ROA and monitoring complexity. Networks commonly apply prefix-length filters, and acceptance of a route depends on operator policy. The registry's recognition of a /16 does not guarantee global acceptance of every more-specific announcement.

CIDR was designed in part to support route aggregation and scaling. RFC 7020 retains hierarchical allocation as a system goal while explicitly separating actual advertisement from registration. This matters economically: the buyer of a contiguous /16 may be paying for the option to aggregate, not only for 65,536 individually countable addresses.

Fragmentation history can continue after reaggregation. Old route objects, ROAs, customer filters, geolocation entries and reputation observations may refer to the former more specifics. A buyer announcing only the /16 must still review the full covered set. A clean aggregate-level lookup can miss a /24 exception; a single problematic /24 should not automatically condemn all 256 if the buyer can isolate or remediate it.

The valuation dossier should therefore state both the commercial lot and the operational prefix tree. It should show whether the entire /16 transfers in one case, whether any subprefix is excluded or separately encumbered, which more specifics have been announced, which origins were observed, what ROAs permit, and what cleanup is required. Address count alone does not answer any of these questions.

Region changes delivery even when routing is global

An IPv4 address can be announced across regional boundaries. Its administrative transfer still follows regional rules. The difference is not a minor closing formality because it affects who is eligible, what must be proved, when delivery occurs and what operational objects survive.

ARIN's current specified-recipient and inter-RIR procedures require an authorised source, a recognised holder, applicable account standing, a non-refundable evaluation fee, recipient qualification where required, a Registration Services Agreement and completion of any necessary inter-RIR coordination. Current policy includes source holding restrictions and recipient-use demonstrations in stated cases. A /16 whose source cannot prove the registered chain is not equivalent to one whose seller is ready and authenticated, even if both are visible in BGP.

RIPE NCC policy permits complete or partial transfers by legitimate holders, subject to restrictions including a twenty-four-month period for specified scarce resources and voluntary transfer locks. The original holder remains responsible until completion, which occurs when the RIPE NCC updates the registration records. Inter-RIR cases must satisfy applicable commitments of the receiving RIR. A buyer that can receive within the RIPE NCC service region may not qualify for every outbound path on the same timetable.

APNIC distinguishes unused or excess IPv4 transfers, historical resources and changes in business structure. Its conditions can require a detailed use plan, membership or transfer fees and special treatment for addresses delegated from 103/8. In outbound cases, associated sub-assignment, route and domain objects can be deleted. That deletion is manageable when anticipated and disruptive when discovered after a price has been fixed.

The same /16 can therefore have several price concepts. The headline amount may be the seller's gross consideration. The buyer's all-in cost adds facilitator, legal, tax, escrow, registry, financing, remediation and deployment expense. The seller's net subtracts its own costs and delay. A bid conditional on a difficult inter-RIR path should not be compared without adjustment to an in-region completed sale.

Regional labels can also become lazy proxies. A buyer may say ARIN space commands a premium or that another region deserves a discount. A broker's book may observe such patterns in a period, but the label does not identify the cause. The actual drivers may be buyer concentration, recipient requirements, address supply, legacy agreement status, policy restriction, transfer time, currency, tax, documentation, market convention or the intended deployment region. Region is a path variable, not an intrinsic quality of the bits.

The correct comparison matches cases by date, block size, source and recipient path, registration status, condition, urgency, intended use and commercial terms. It preserves unsuccessful cases rather than observing only completed transfers. If no matched sample exists, the valuation should explain the qualitative path adjustment rather than invent a percentage.

RIR public logs help verify that a transfer completed and when. They cannot show all the cases that never reached the log. This selection effect matters: the slowest or most defective blocks may be missing precisely because they did not close. Completed-record averages, even if prices were later attached, would overstate the experience of an unfiltered seller population.

The region effect is real enough to disclose and too heterogeneous to compress into a permanent multiplier.

Routing state is a deliverable, not a background fact

A buyer rarely purchases a /16 merely to admire the registration record. It expects to announce the space, delegate use, secure origins, operate reverse DNS and have traffic accepted. Routing state should therefore be represented as part of delivery.

The first question is current and historical origin. Which autonomous systems have originated the /16 or its more specifics? Over what intervals? With what observed visibility? Did an origin change correspond to an authorised operational transition, a customer, internal restructuring or an incident? RIPEstat routing history can provide collector-based timelines. RouteViews and other public collectors can add perspectives. No finite collector set proves the view of every network.

The second question is prefix shape. Is the aggregate visible? Are more specifics active? Does the buyer intend to preserve them, withdraw them or introduce new ones? Route filters are controlled by receiving networks. A plan relying on very specific announcements needs direct testing and provider coordination, not an assurance that the registry recognises the covering /16.

The third question is route-origin authorisation. RFC 6811 describes validation states based on covering validated ROA payloads, origin ASN and prefix length. A matching authorisation can produce Valid. A covering authorisation with the wrong origin or insufficient maxLength can produce Invalid. No covering authorisation produces NotFound. Local routing policy determines what a network does with the result.

This means RPKI status must be evaluated against the buyer's planned routes, not against the /16 in isolation. A ROA authorising the seller's ASN for the /16 may be correct before closing and wrong after the buyer originates /20s. A maxLength permissive enough to cover /24s can ease delegation while expanding the set of authorised more specifics. An AS 0 ROA can express that covered space should not be routed and must be changed before intended use.

The fourth question is repository and object transition. ARIN explains that its resource certificate is reissued after transfer and advises parties to review or create ROAs, Internet Routing Registry objects and reverse DNS. APNIC notes object deletion for outbound resources. A valuation that assumes these services travel automatically with the commercial contract understates cutover cost.

The fifth question is acceptance. Even technically correct registration, ROAs and route objects do not guarantee that every upstream and peer updates filters on the same day. A buyer can reduce uncertainty with pre-coordination, route tests, staged announcements and monitoring. It cannot purchase universal instantaneous convergence.

Routing state affects price through time and use. A stable aggregate with a simple origin history may be cheaper to deploy for one buyer. A heavily subdivided portfolio with existing customer origins may be more valuable to a lessor that wants continuity. An unannounced block may avoid some operational history but require fresh filtering and geolocation work. No one state is always best.

The sale agreement should define routing acceptance tests. They may include the ability to create intended ROAs, publish required route objects, configure reverse DNS, announce agreed prefixes through designated upstreams, observe them from named collectors and clear active conflicting authority. The tests must name sources and duration. Passing them establishes bounded acceptance, not global reachability forever.

When routing becomes a deliverable, price negotiation improves. The seller can be paid for a prepared handoff. The buyer can reserve for a known exception. Both can avoid treating a post-closing invalid route as evidence that the entire commercial transaction was fraudulent.

Reputation must be sampled by purpose and time

Reputation services can influence deliverability, fraud decisions, content access and customer trust, but they do not share one definition of clean. A proper review starts with use case.

For mail, the buyer can inspect named public DNS-based lists, test forward and reverse DNS readiness, review available provider dashboards and plan gradual traffic warming. RFC 5782 describes DNSBLs as a distribution mechanism while leaving listing and removal policy to the list operator. RFC 6471 recommends transparency, direct removal routes, prompt response and review or expiry suited to the condition. These standards explain why a current listing needs source-specific interpretation.

For cloud or hosting use, the buyer may need private abuse and fraud assessments, customer-isolation controls and a plan for rapid response. Many relevant signals are proprietary. The seller cannot prove absence from a system it cannot query. The contract can require disclosure of known material incidents and cooperation with corrections rather than a warranty covering every undisclosed model.

For consumer-facing services, geolocation can affect language, licensing, fraud controls and access. RFC 9632 defines discovery and optional authentication of geofeed data. It does not require every provider to ingest the feed immediately or accept its contents without review. MaxMind and IPinfo publish correction routes, but their update schedules and customers remain separate. A buyer should test the providers material to its service, not assume one corrected database settles the world.

The temporal window matters. A lookup thirty days before closing can become stale. A scan immediately after cutover can misattribute seller traffic or buyer experiments. The dossier should record baseline dates, transfer completion, route cutover and subsequent observations. Material changes should be assigned to the party controlling use during the relevant interval where evidence permits.

Sampling should also cover the prefix tree. Some systems operate on individual IPs, some on prefixes, some on autonomous systems, domains or behaviour clusters. Checking one address cannot certify the other 65,535. Exhaustive querying may violate service terms, trigger rate limits or reveal little about private systems. A risk-based sample can cover each historical more specific, known service cohort and random addresses, with explicit coverage.

Remediation cost belongs in the price bridge. A current, well-explained listing with a documented removal path may justify a holdback rather than rejection. An unknown private reputation state may justify a deployment reserve. Persistent active abuse from an undisclosed customer is a different risk from a historical policy-list entry. The contract should respond differently.

The buyer should not demand a mystical clean letter. It should demand evidence that known material conditions have been identified, that current control is understood and that the seller will assist with historical correction after closing. Reputation is a changing relationship between addresses, behaviour and observers. The price should reflect the bounded evidence, not a permanent promise.

Build the comparable price before applying the exceptions

A disciplined valuation begins with a reference cohort, not a headline average. The cohort should match the valuation date, block size, transaction form, registry path and price concept as closely as available evidence allows.

Suppose the available evidence includes marketplace sales of /16s during a recent period. The analyst should first separate completed consideration from asks and forecasts. It should identify whether each observation represents one contiguous /16, a larger block priced on a per-address basis, or a package. It should note the source and recipient regions, whether the price includes fees, and whether the observation comes from one provider's book.

The analyst then builds an exception bridge. The bridge does not need a universal percentage for each item. It needs a reasoned estimate of incremental cash cost, delay, restricted utility and residual uncertainty for the subject block.

Dimension Evidence Possible value effect Invalid shortcut
Recognised chain RDAP or WHOIS, agreements, corporate history, transfer receipts Legal and administrative work, delay, failed-delivery risk Current holder name proves every prior right
Transfer path Source and recipient RIR rules, eligibility, fees, locks, pre-approval Time, transaction expense, recipient pool and completion probability Global routeability means global transferability
Commercial fragmentation Exact lot schedule, exclusions, leases and encumbrances Buyer fit, number of closings, gross-versus-net proceeds 256 /24s equal one delivered /16
Routing fragmentation Historical aggregates, more specifics, origins and visibility Cutover effort, filtering, monitoring and aggregation option One current BGP view proves all past and future reachability
RPKI and route objects ROAs, intended origins, maxLength, IRR and reverse DNS Security handoff, invalid-route risk and operational delay A Valid route means the asset is legally clean
Reputation Named public lists, available private tools, incident history and samples Use restrictions, remediation, warming and reserve Absence from one list means universal cleanliness
Geolocation Named provider tests, geofeed and correction status Customer experience, compliance and deployment delay Registry country equals user location
Revenue and use Leases, utilisation, customer rights and concentration Cash flow, vacancy, handback and servicing cost Registered address count equals earning address count
Timing and terms Closing condition, escrow, financing, warranties and holdbacks Present value, execution risk and seller net Contract price equals buyer all-in cost

The bridge should distinguish deterministic expense from risk. A known registry fee can be added. A known object-recreation project can be budgeted. A two-month delay can be valued against financing or foregone service. A probability of unresolved private reputation harm is harder and should be shown as a scenario, not hidden inside a confident point estimate.

Buyer-specific utility must remain visible. If the buyer needs mail-ready addresses, a reputation exception may dominate. If it needs a contiguous aggregate for access infrastructure, fragmentation may dominate. If it plans to lease small blocks, existing subdivision and customer records may be useful rather than costly. There is no context-free clean premium.

The output can be a base range, identified adjustments, scenario reserve and residual uncertainty. It should show seller gross, seller net and buyer all-in values separately. A transaction may be attractive to both parties even when those figures differ materially.

This method does not create perfect price discovery. It creates an auditable explanation of why the subject /16 should not be priced as an unexamined multiple of a market average.

The dossier should survive a skeptical buyer

A premium is credible only when another party can reproduce its basis. The seller's quality dossier should therefore be structured for challenge.

The identity section names the exact prefix and recognised organisation, includes current RDAP or WHOIS output, identifies the registry account and agreement class, and maps legal-name changes, mergers and prior recognised transfers. It states what is not publicly available. A clean chain claim should identify its evidence period and responsible reviewer.

The policy section specifies the contemplated source-recipient route, applicable transfer type, current restrictions, recipient conditions, estimated institutional fees and any pre-approval. It records the date rules were checked. It does not assume that a previous transfer under another path predicts this one.

The prefix-tree section lists every included and excluded subprefix, active lease, letter of authorisation, customer route and separate disposition. Coverage must reconcile exactly to the /16. Overlaps and gaps are stop conditions until explained.

The routing section stores time-indexed observations from named collectors, expected and unexpected origins, more-specific history and visibility. It distinguishes observation from inference. A change in origin can indicate a transfer, customer migration, traffic engineering or hijack; BGP alone cannot choose among them.

The security section captures current ROAs, intended post-closing origins, maxLength implications, route objects, reverse-DNS delegations and credential ownership. It includes the cutover order and rollback condition. Secrets remain protected; the dossier records authority and recovery, not plaintext passwords.

The reputation section names each public or accessible private source, lookup date, sampling method, list type, result, remediation status and limitation. It keeps mail, fraud, proxy, threat, geolocation and policy classifications separate. An unresolved exception includes a cost and owner.

The commercial section lists leases, receivables, liens, financing restrictions, customer concentrations, termination rights, taxes, fees, escrow and holdbacks. It makes clear whether the quoted price buys vacant space, an income stream or a mixture.

The acceptance section defines completion in stages: registry approval, authoritative update, payment release, route and RPKI cutover, DNS readiness, customer migration and post-closing reputation check. A single closing date cannot describe all these events safely.

Finally, the dossier includes a limitations statement. It should say which private reputation systems were inaccessible, which collector views were used, which historical documents were unavailable and which rules can change. That statement increases credibility. It tells the buyer where the reserve belongs.

A dossier does not eliminate negotiation. It converts generic suspicion into specific exceptions. Sellers with strong evidence can defend value. Buyers can compare like with like. Brokers can explain why two /16s in the same month cleared at different levels without pretending to know the entire market.

NRS can standardise disclosure without declaring cleanliness

Number Resource Society can improve this market by defining the dossier and price bridge as open, portable formats. The intervention should target information quality, not price authority.

An NRS profile can define mandatory fields for prefix identity, recognised holder, evidence date, transfer path, commercial lot, historical origins, current ROAs, intended routes, public reputation checks, geolocation status, encumbrances and limitations. Each field should identify source, timestamp, responsible party and confidentiality class. A machine can validate prefix coverage while a human evaluates legal and commercial meaning.

NRS can publish conformance tests. Do included subprefixes cover the /16 exactly? Are excluded portions identified? Do intended routes fit within stated ROAs? Are transfer-rule snapshots dated? Does every reputation claim name its provider and sample? Are seller gross and buyer all-in values separated? Tests can catch omissions without deciding whether the remaining facts deserve a premium.

It can define change-of-control receipts that reputation and geolocation services may accept as evidence. The receipt would not require a provider to erase true history. It would give the provider a reliable date, old and new operator, exact resource set and authoritative registration event when evaluating current attribution.

NRS can support confidential comparable-price reporting. Participants could submit realised consideration, all-in cost, block size, region path, condition categories and timing to an independent aggregation process. Reports must state contributor count, case count, address count, period, inclusion rules and suppression thresholds. Selected participants must never be described as the global market.

It can also publish buyer-specific templates. A mail-use profile would emphasize sending reputation and DNS. A network-aggregation profile would emphasize contiguity, filters and RPKI. A lease-finance profile would emphasize contracts, utilisation and servicing. Common identity and transfer fields would allow comparison while avoiding one universal clean grade.

The exclusions protect legitimacy. NRS should not certify title, approve a transfer, guarantee routeability, order a blocklist removal, set a mandatory benchmark, recommend a specific broker or declare one region inherently superior. It should not demand publication of customer identities, private threat signals or commercially sensitive prices. It should not turn missing private data into an automatic adverse finding.

NRS's holder-focused mission makes this a constructive area for work, but that mission is first-party evidence. The proposed profile, receipts, tests and reporting population are not yet independently demonstrated by the sources cited here. Adoption should be measured through published specifications, interoperable implementations and reports with explicit denominators.

The benefit is institutional rather than promotional. A market with comparable disclosure rewards prepared sellers, lets buyers price known exceptions and reduces pressure on registries to certify matters outside registration. NRS can make evidence portable while leaving price to counterparties.

What the evidence cannot price

No public dataset links every recognised IPv4 transfer since 2013 to realised consideration, block condition, failed negotiations, remediation expense and buyer use. Private direct deals remain largely invisible. Public logs select completed administrative outcomes. Broker reports select provider-observed activity. Academic studies select periods, collectors, lists and methods.

There is no public global premium for a clean /16. The term itself has no standard test. There is no universal discount for a blocklist entry, prior hijack, stale geolocation record, fragmented route history, inter-RIR path or incomplete chain. The effect depends on severity, recency, use and remediability.

Public reputation evidence is incomplete. DNSBLs differ. Mail-provider dashboards show provider-specific data to authorised users. Fraud and security systems can be private. A clean public sample cannot prove absence from every private system. A historical adverse observation cannot prove present misuse.

BGP evidence is also partial. Collector visibility changes. Origin changes have multiple causes. A route can be accepted by some networks and rejected by others. RPKI origin validation addresses origin authorisation, not path validity, business authority, registration chain or reputation.

RIR data provide authoritative registration facts within their scope, not legal title insurance, market price, route guarantee or a complete corporate history. Current policy can change before closing. Inter-RIR availability depends on both sides and the exact direction.

These limits are not reasons to use one average. They are reasons to disclose scenarios and reserves. The analyst can price known cash costs, estimate delay with transaction-specific evidence and show ranges for uncertain use restrictions. It should leave genuinely unavailable denominators unavailable.

The confidence in the article's thesis is high because the mechanisms are documented: registration feeds omit price, broker samples vary, policies differ, routes and ROAs have distinct states, and reputation services apply different criteria. Confidence in any universal coefficient would be low because the necessary matched global panel does not exist.

The valuable /16 is the one whose exceptions are legible

Scarcity makes every globally unique IPv4 address potentially useful. It does not make every /16 fungible.

The buyer is acquiring a path from seller authority to recognised registration, from registration to operational authorisation, and from operational authorisation to a particular commercial use. History, fragmentation, region and routing state change the cost and reliability of that path. Reputation and geolocation can continue to affect use after the registry record changes.

A good seller does not promise universal cleanliness. It presents a dated, use-specific dossier, explains exceptions and accepts bounded obligations to cure or cooperate. A good buyer does not turn every historical observation into a permanent stigma. It prices current effect, remediation, delay and residual uncertainty.

Market reports remain useful when labelled as contributor evidence. RIR logs remain indispensable when used as administrative records. Routing collectors, RPKI, DNSBLs and geofeeds remain useful when each is asked the question it can answer. None should be promoted into a total quality oracle.

NRS can connect these sources through an open disclosure profile, exact prefix reconciliation, event receipts, conformance tests and confidential comparable reporting with explicit denominators. It should never publish a clean seal that encourages counterparties to stop looking.

The price of a /16 begins with 65,536 addresses. The price of this /16 depends on how many of those addresses can be delivered, announced, secured and used by this buyer, through this registry path, on this timetable, with these disclosed exceptions.

That sentence is longer than a benchmark multiple. It is also closer to the asset the parties are actually trading.

Sources