Summary

  • APNIC RDAP identifies AS148972 as ZCCCCL-AS-AP and gives the exact name Zero Cirrus Cloud Computing (Shanghai) Co., Ltd.
  • The registration anchors the ASN to a named record; it does not prove that a service is currently operating.
  • At the 1 August 2026 observation boundary, the captured RIPEstat overview reports AS148972 as not announced.
  • The captured announced-prefix list is empty, and routing status reports no currently visible IPv4 or IPv6 space.
  • The same routing-status response reports zero observed neighbours and visibility from zero reporting peers.
  • The historical visible route 103.175.14.0/23 was first seen on 24 December 2021 and last seen on 2 May 2022.
  • Those observations describe the RIPE RIS view available at capture time, not every possible routing context on the Internet.
  • Nothing in the available evidence explains the later silence or establishes closure, outage, transfer, abandonment or permanent inactivity.

The record that remains

AS148972 presents two kinds of infrastructure evidence that are easy to collapse into one another but should remain separate. The first is an administrative identity. APNIC RDAP returns the aut-num record for AS148972, labels it ZCCCCL-AS-AP and includes the exact company name Zero Cirrus Cloud Computing (Shanghai) Co., Ltd. The BTW directory entry carries the same company boundary and projects AS148972 as its network identity. Taken together, those records provide a narrow, stable answer to a narrow question: which named organisation is attached to this autonomous system number in the captured registration material?

The second kind of evidence is operational observation. RIPEstat’s captured responses do not show a live public announcement for AS148972. The overview says “not announced”; the announced-prefix response contains no prefixes; the routing-status response shows no currently visible IPv4 or IPv6 space, no observed neighbours and no visibility from reporting peers. That is not a contradiction in the records. It is a difference in what each record is designed to expose. An ASN can remain registered as an identifiable number resource even when a particular routing observation has nothing current to display for it.

That distinction is the central fact. The registration has not dissolved merely because the captured routing view is quiet, and the registration does not convert the quiet view into evidence of current service. One side preserves identity and administrative continuity. The other reflects route propagation seen through a defined observation system at a defined time. Holding both at once produces a more useful account than choosing whichever one appears more conclusive. Zero Cirrus remains named in the APNIC record; AS148972’s visible route, in the captured RIPEstat view, does not.

What the ASN record establishes

An autonomous system number is useful because it gives routing activity a unique numerical identity. When an origin appears in observed routing data, the ASN allows that activity to be followed without relying only on a brand, website or changing commercial description. When the route disappears from view, the number still gives investigators, operators and counterparties a precise object to query. AS148972 is therefore more than a search term. It is the stable join between the exact holder label returned by APNIC and the origin identity examined by RIPEstat.

The APNIC response makes that join unusually clear. Its aut-num is AS148972. Its name is ZCCCCL-AS-AP. Its descriptive material uses the exact Zero Cirrus Cloud Computing (Shanghai) Co., Ltd. name, and the country code is CN. The record shows a registration event on 22 October 2021 and a later change event on the same date. These details establish the captured administrative record and its timestamps. They allow the ASN, short name and exact company name to be checked against one another rather than inferred through a loose brand resemblance.

Exact-name agreement matters most when the claim remains modest. It reduces ambiguity about which directory company is linked to the ASN. It does not expand the evidence into a corporate profile. The record does not, on its own, say whether Zero Cirrus is offering a current service, how it might be organised, what assets it might use, or how any network might be engineered. Nor does the country code establish a complete geographic operating footprint. It is a field in the number-resource record, not a coverage map.

Registration also does not turn APNIC into a continuous probe of the holder’s network. RDAP answers a question about registered data. It does not test route propagation, contact responsiveness, application availability or commercial activity at the moment of lookup. Its value here is durable identification: even after the captured routing view becomes silent, the number can still be tied to a named record. That is an accountability surface, not a warranty of operation.

What registration cannot certify

A persistent name beside AS148972 may look reassuring because it is concrete, structured and authoritative within its administrative domain. Yet concreteness is not the same as operational proof. APNIC’s record cannot establish that packets currently originate from the ASN, that any public prefix is reachable, or that Zero Cirrus is providing a live cloud or network service. Those questions belong to other evidence. The captured RIPEstat responses, at least at the stated boundary, do not provide affirmative answers to them.

The same restraint applies in the opposite direction. A quiet routing view cannot invalidate the name in the registry. It does not show that the entry has been abandoned, transferred or left without an accountable holder. None of those conclusions appears in the available evidence. A number-resource record and an observed route occupy different layers: one describes the registered association; the other describes activity visible through route collection. They may change on different schedules and for different reasons.

Nor can the exact company name be extended to everything that might resemble it. Similar naming does not establish shared ownership, shared facilities, shared customers or a shared network. The firm boundary here is Zero Cirrus Cloud Computing (Shanghai) Co., Ltd., linked to AS148972. Nothing supports attaching other services, websites, affiliates or physical sites to that boundary. The benefit of the exact record is precision, and precision is lost if the name becomes a licence to absorb unverified associations.

This is why the phrase “keeps Zero Cirrus in the registry” should be read literally. The record keeps the exact company name attached to AS148972 in the captured APNIC response. It does not keep a service online, preserve a route, validate a business or certify a physical deployment. The surviving fact is identification. Everything beyond that requires its own observation.

The running view on 1 August 2026

RIPEstat adds the running-state evidence. At the captured boundary on 1 August 2026, its AS overview identifies the holder as “ZCCCCL-AS-AP - Zero Cirrus Cloud Computing (Shanghai) Co., Ltd.” and reports AS148972 as not announced. This agreement on the holder label is useful because the administrative and routing views are aimed at the same ASN. The operational field then diverges sharply from the durable identity: the holder is named, but no announcement is shown.

A second response narrows the point. The announced-prefix endpoint returns an empty list for AS148972. There is no currently listed IPv4 or IPv6 prefix in that captured result. A third response, routing status, also reports zero current visible IPv4 and IPv6 space. Reading these outputs together avoids leaning on a single status flag. The overview, prefix inventory and routing-status summary all describe the same quiet current observation from different angles.

The routing-status response also shows zero observed neighbours and visibility from zero reporting peers. Those zeros should be interpreted as a coherent absence inside the captured RIPE RIS view. There is no visible origin space from which to derive an observed adjacency in that response, and no reporting peer is shown as seeing the ASN. The fields reinforce the “not announced” state. They do not widen it into a universal claim about every router, private session or network relationship.

Running-code evidence deserves priority when the question is whether a route is visibly operating. A registry label cannot substitute for an announcement. Here the running view is quiet. But running-code evidence also needs a boundary. RIPEstat is reporting what its underlying observation exposed; it is not issuing a corporate-status finding. The correct sentence is not “Zero Cirrus has no network.” It is that the captured RIPEstat view reports no current public announcement for AS148972.

That wording protects the strongest available fact from being weakened by overreach. It says exactly where the observation comes from, exactly which identifier it concerns, and exactly what is absent. It also leaves open the possibility that evidence outside the observed system could differ without suggesting that such evidence actually exists. A bounded statement can be both technically serious and incomplete.

The route with two dates

The quiet current state has a visible history. RIPEstat’s routing-status response records 103.175.14.0/23 as the historical prefix-origin pair associated with the first-seen and last-seen markers for AS148972. The first date is 24 December 2021. The last is 2 May 2022. Those dates provide the only supported chronology for route visibility in the available evidence: a prefix was visible with AS148972 as origin at the earlier boundary, and the captured history stops showing it at the later one.

The chronology is notable because it begins soon after the APNIC registration date of 22 October 2021. That sequence allows a limited observation: the administrative record predates the first-seen route marker, and the route later has a last-seen marker. It does not explain the interval between them, and it does not prove that the registration caused the announcement. The two systems are recording different events. Their dates can be placed alongside one another without being turned into a story of intent.

The prefix itself also needs careful treatment. 103.175.14.0/23 is the route identified in the captured historical response. It should not be described as Zero Cirrus’s only possible address resource, permanent allocation, complete network footprint or current property. The evidence supports a past visible prefix-origin pairing with AS148972. It does not supply a broader inventory of address holdings or a chain of later disposition.

Similarly, “last seen” is an observation boundary, not a declaration by an operator. It states when that route was last visible in the data represented by the response. It does not state when a business decision was made, when a technical change was initiated, whether the route was intentionally withdrawn, or whether another origin or context existed later. The timestamp marks the edge of visibility, not the reason for it.

There is still analytical value in that edge. It separates a route that has never appeared in the captured history from one that did appear and then ceased to be visible there. AS148972 belongs to the latter category on the available evidence. The route’s historical presence makes the current silence a change in observed state, while the unchanged identity anchor makes it possible to follow that change under the same number.

Silence without a cause

The period after 2 May 2022 is where disciplined interpretation matters most. The captured responses show no current route and identify a last-seen date, but they do not explain what happened after that boundary. No operator statement, transfer record, incident notice or technical change record is present. There is therefore no supported causal narrative for the silence.

Several tempting conclusions must be rejected. The data does not establish closure, outage, failure, abandonment, permanent inactivity or a transfer. It does not establish that Zero Cirrus ceased every form of operation. It does not establish that an earlier service continued either. The absence of a visible announcement answers a routing-observation question; it does not answer a legal, commercial or organisational one.

Even the word “withdrawal” can carry more intent than the evidence allows if it is used without attribution. A collector can stop seeing a route for reasons that are not disclosed in the response. The safe description is that 103.175.14.0/23 was last seen with AS148972 on 2 May 2022 and that the current captured view contains no announcement. The transition between those points is observable. Its mechanism is not.

Uncertainty here is not an empty disclaimer. It defines what new evidence would be needed. An explanation would require information capable of linking the routing change to an operator action or a documented administrative event. In the absence of that link, the strongest account stops at chronology. That is more informative than a speculative diagnosis because it keeps the record usable when a later observation arrives.

The edges of collector visibility

Route collectors provide a powerful view of public routing, but every view has a perimeter. A route becomes visible to a collector when the surrounding propagation and reporting relationships carry it into the collector’s observation. The resulting data can show that a route was seen, associate it with an origin ASN and record changes over time. It cannot guarantee that every routing context everywhere is represented with equal completeness.

That perimeter matters for AS148972 because every current field is zero or empty. The captured overview reports not announced. The announced-prefix list has no entries. Routing status shows no visible IPv4 or IPv6 space, no neighbours and no reporting-peer visibility. Together, those fields strongly support silence within the RIPE RIS observation. They still do not support the universal statement that AS148972 appears nowhere in any routing context.

Private routing is an obvious boundary, but it should remain a boundary rather than an invented alternative history. The available evidence does not show a private route, a customer session or a hidden interconnection. It merely cannot rule out activity that would not enter the public collector view. The same applies to a route originated by another ASN: nothing here establishes that 103.175.14.0/23 moved to another origin, but the AS148972 query alone is not a complete account of every possible later use of that address space.

Collector-external activity is another limit. A route might be visible from a place not represented in the captured reporting set, or only under conditions that do not reach the observation. That possibility does not weaken the actual result. It explains why attribution must remain attached to the collector. “RIPEstat reports no current announcement” is a complete and testable finding. “The ASN is absent from the entire Internet” would exceed the measurement.

The zero reporting-peer visibility field gives the boundary a concrete expression. It means the response did not show reporting peers seeing AS148972 at the captured moment. It is not a census of all possible peers, contracts or devices. A reporting peer is part of the observation apparatus; a commercial or physical counterpart is a different category. Treating one as proof of the other would convert telemetry into a business claim.

Careful scope also preserves comparability. If a later check shows a prefix or a reporting peer, the change can be described within the same observation system. If the current state is instead framed as universal nonexistence, any later visibility would appear to overturn an exaggerated claim rather than update a bounded measurement. Infrastructure monitoring works best when each statement carries its field of view with it.

Empty prefixes and zero neighbours

The empty announced-prefix list is the clearest current signal, but it is not a list of everything Zero Cirrus may control. It is the list returned for announcements associated with AS148972 in the captured RIPEstat response. Because it is empty, no current prefix can be attributed to the ASN from that endpoint at that time. It says nothing by itself about registered address resources, unused assignments, private addressing or physical equipment.

The routing-status zeros should be read in the same way. Zero visible IPv4 space and zero visible IPv6 space describe what is currently originated and observed under AS148972 in the response. They do not erase the ASN record, and they do not prove that no addresses are registered elsewhere to the named company. The available records do not provide a complete IP-resource inventory for Zero Cirrus. The ASN query is exact; the company-wide resource picture is not available.

Zero neighbours is equally narrow. In routing observation, a neighbour is an adjacent ASN relationship inferred or exposed from seen paths around the queried ASN. With no visible announcement, there is no current path context in the response from which to present such neighbours. The field does not prove that Zero Cirrus has no contracts, no upstream arrangements, no physical cross-connects or no historical counterparties. None of those things is established either way.

This is why the three current outputs should be treated as mutually reinforcing rather than independently inflated. “Not announced,” an empty prefix array and zeros for visible space and neighbours all point to the same captured condition. They do not add up to a stronger kind of claim about company status. Repetition across endpoints increases confidence in the routing observation, not the breadth of what that observation can legally or operationally mean.

Two clocks: administrative and operational

AS148972 illustrates how infrastructure records can run on two clocks. The administrative clock begins with the ASN’s registration and continues through the availability of its holder record. APNIC shows the registration and same-day change events from 22 October 2021, and it still returns the exact Zero Cirrus name in the captured response. This clock is concerned with an identifiable number resource and the data attached to it.

The operational clock is marked by visibility. For 103.175.14.0/23, the captured history begins on 24 December 2021 and reaches its last-seen marker on 2 May 2022. At the later observation boundary of 1 August 2026, there is no current announcement in the RIPEstat view. The operational clock therefore has a visible interval and then a long unfilled space. No supported event explains that space.

Neither clock can replace the other. The administrative record cannot extend the route’s last-seen date forward. The routing silence cannot backdate the disappearance of the registered identity. Together they show continuity of reference without continuity of observed announcement. That combination is exactly why persistent identifiers matter: the same ASN can be queried across a change in running state, making the discontinuity visible rather than allowing it to vanish with a brand or hostname.

The clocks also impose different update questions. For APNIC, the relevant questions concern whether the holder label, events and accountable record remain accurate. For RIPEstat, the questions concern whether announcements, prefixes, neighbours or peer visibility reappear. A change on one clock does not automatically force a conclusion on the other. A routing reappearance would not by itself validate company operations beyond routing; an administrative change would not by itself prove a route transition.

Why a quiet ASN still matters

A number resource remains consequential even when its current routing view is silent. Uniqueness is part of the reason. AS148972 identifies one autonomous system in the numbering system; it is not an interchangeable label that can be treated as disposable when no route is visible. The record allows past observations, current absence and future changes to be attached to the same identifier.

Accuracy matters for the same reason. If the name associated with a quiet ASN is stale, ambiguous or wrong, anyone investigating a future route, an abuse complaint or an operational handoff begins from a faulty premise. The exact agreement between APNIC and the BTW directory gives a clear present identity anchor, but it does not answer whether every underlying contact and operational detail is current. The value of the anchor lies partly in making those unanswered questions addressable.

Continuity is not the same as constant announcement. It is the capacity to account for a resource across changes. A route may appear, disappear and perhaps appear again; an administrative record may remain stable or change. What matters is that those transitions can be recorded without losing the identity of the number. For AS148972, the visible sequence is registration, a later first-seen route, a last-seen route and a current quiet view. The cause between stages remains blank, but the stages themselves remain linkable.

Security metadata belongs in this accountability frame, though no claim about AS148972’s security controls is supported here. Number resources can be implicated in routing mistakes or abuse even when normal service visibility is low. Accurate holder data and a traceable history make it easier to direct questions and distinguish an expected change from an unexplained one. They do not prevent incidents, certify routing authorisation or prove responsive handling.

Transfer recording is another part of continuity, but the available evidence shows no transfer. That absence of evidence must not be rewritten as proof that no transfer occurred; it simply means transfer is not an available explanation. If responsibility for a resource changes, the administrative trail should make that change legible. Until such evidence is present, the exact APNIC holder label remains the supported attribution for the ASN record, and nothing more should be inferred.

Contact continuity is not contact proof

A named registration creates a point of accountability, but it does not demonstrate that a person or mailbox can be reached today. The captured facts establish the ASN, short name, company name, country code and event dates. They do not establish current responsiveness, staffing, escalation procedures or incident-handling capability. Those are operational properties that require direct and current evidence.

This distinction becomes more important after a route has gone quiet in the observed view. When active paths are visible, counterparties may have multiple technical signals through which to identify an origin and its adjacencies. When no path is visible, the registered record may become the most durable public pointer. Its accuracy and maintenance therefore matter even though the record cannot prove the organisation behind it is actively operating a network.

Contact continuity should also not be inferred from the mere availability of RDAP. A lookup service returning a record means the record can be retrieved. It does not mean every field has been independently tested against the present day. Conversely, no conclusion about unreachable contacts can be drawn because the available evidence contains no contact test. The responsible position is neither confidence nor suspicion; it is a defined missing measurement.

A practical accountability question is whether someone encountering a future AS148972 announcement could trace it to current, responsive responsibility. APNIC’s exact company label supplies the first step. The later steps—contact verification, operator confirmation and incident-response readiness—remain outside the evidence. Preserving that distinction prevents an administrative pointer from being advertised as an operational assurance.

Security metadata without a security verdict

Routing security is another area where absence of evidence can invite overstatement. The captured material does not establish a route-origin authorisation state, filtering policy, validation practice, incident history or security failure for AS148972. It therefore supports no verdict about whether the ASN is secure or insecure. The fact that no route is currently visible to the observed collectors does not answer those questions.

What can be said is more structural. If AS148972 becomes visible again, accurate security-related records would help observers evaluate whether the announcement fits the expected origin and resource history. A new route could be legitimate, mistaken or malicious; the mere fact of reappearance would not decide among those possibilities. It would create a need to compare the new running state with authoritative administrative and authorisation data available at that future time.

The historical 103.175.14.0/23 observation cannot be used as a permanent authorisation statement. It records that the prefix-origin pair was seen between the first and last boundaries in the captured response. It does not say that every appearance was authorised, that the same pairing remains authorised, or that a later different origin would be improper. Those judgments need evidence that the available records do not provide.

A quiet current view also should not be described as safer simply because fewer announcements are visible. Silence reduces what the collector can observe; it does not reveal the condition of systems, credentials, records or unused resources. Security accountability depends on accurate identity, current metadata and observable changes, not on treating absence as a positive control. For AS148972, the appropriate security finding is that the relevant evidence is unavailable, not that a failure or assurance has been demonstrated.

Operational continuity as a set of questions

The gap between registration and current routing visibility leaves a series of operational questions. Is the registered holder information still maintained? Is AS148972 being retained for a future use, held without announcement, or involved in a context outside the captured public view? Is 103.175.14.0/23 still administratively connected to the same party? Has responsibility changed in a way not represented by the available evidence? None of these questions has an answer here, and none should be converted into a claim.

The most important continuity question is narrower: can the sequence of responsibility be reconstructed if the ASN changes state again? A durable identifier makes that possible in principle. The APNIC name anchors the administrative side. The RIPEstat first-seen, last-seen and current-status fields anchor the observed routing side. Missing are the operator-level facts that connect those points and explain the transition.

A future reappearance would not close the gap automatically. It would show new visibility, but investigators would still need to determine the prefix, origin, observation scope and timing. They would then compare the holder record and any available security metadata. If the exact company label had changed, the timing and nature of that administrative change would matter. If it had not, the registration would remain an attribution anchor, not proof that the route was intentionally originated by a currently operating Zero Cirrus service.

The same caution applies if the quiet state continues. Time alone does not turn “not announced” into “abandoned.” A long interval can make continuity questions more urgent, but it cannot answer them. There is no threshold in the available evidence at which an unobserved route becomes a corporate conclusion. The last-seen date remains 2 May 2022 whether it is checked a week later or years later; only new evidence can supply a cause.

Operational continuity also has a public-interest dimension without requiring advocacy. Unique identifiers and accurate records reduce ambiguity when networks change. They help counterparties distinguish a known past origin from a newly observed one and give investigators a consistent object to monitor. AS148972 shows the value of that continuity precisely because the running view is quiet: the number preserves the question even when the route no longer supplies an answer.

The physical layer remains unobserved

BGP visibility is not a photograph of infrastructure. A route announcement can reveal that an origin is visible through observed paths, but it cannot by itself establish who owns a building, where equipment sits, how fibre is arranged, how much power is available or what capacity a service can deliver. The absence of an announcement reveals even less about those physical questions.

Nothing in the available evidence establishes a Zero Cirrus data centre, cloud region, facility, fibre route, power arrangement, coverage area or customer deployment. No capacity, performance, resilience or service-quality measure is available. The company name includes “Cloud Computing,” but a legal or registered name is not evidence of a cloud footprint. Treating it as such would replace network-resource analysis with brand inference.

The same boundary applies to ownership and market presence. APNIC’s exact holder label does not establish beneficial ownership, corporate standing, licences, revenue, customers or current commercial activity. The directory link confirms the existing company entry to which the network evidence is attached; it does not expand the evidence beyond the ASN association. A category label is navigation, not a claim that every characteristic commonly associated with that category has been observed.

Even an illustration of registry identity and sparse routing must remain non-documentary. It cannot be treated as a picture of Zero Cirrus equipment, topology, facilities or coverage. Visual metaphors are useful only when their limits are explicit. The physical and commercial layers remain unavailable because none of the available records measures them.

That absence is central rather than incidental. It prevents the quiet route from becoming a story about an unseen facility or business. The supported account is entirely at the level of administrative identity and captured route visibility. Keeping it there preserves both the technical value of the evidence and the dignity of what is unknown.

A repeatable way to monitor AS148972

A useful monitoring practice begins by fixing the resource and the observation time. The resource is AS148972, not a loose company-name search. Each check should record when the APNIC and RIPEstat responses were obtained, because “current” is only meaningful at a capture boundary. A later response may differ without making the earlier one wrong. It would describe a changed or differently observed state.

The first step is to query the APNIC aut-num record and compare stable identity fields. The ASN should be AS148972; the captured short name is ZCCCCL-AS-AP; the descriptive company name is Zero Cirrus Cloud Computing (Shanghai) Co., Ltd.; the country code is CN. Event fields should be noted without assigning motives to them. If a future response changes the holder label or adds a later event, the change should be recorded exactly and then investigated separately. It should not be interpreted as transfer, closure or continuity until supporting evidence establishes which one applies.

The second step is to read the AS overview for two different purposes. One is identity alignment: does the holder label still correspond to the APNIC record? The other is announcement state: does the overview still say not announced, or has that flag changed? These are distinct checks. A matching label does not make the announcement live, and a live announcement would not by itself prove the rest of the company profile.

The third step is to inspect the announced-prefix response rather than relying only on the overview flag. At the 1 August 2026 boundary, the list is empty. A future non-empty result would be a concrete state change. The prefix or prefixes should be recorded exactly, along with capture time, before any relationship to the historical 103.175.14.0/23 route is asserted. A different prefix would not automatically be a continuation; the same prefix would not automatically explain the intervening silence.

The fourth step is to examine routing status. Current IPv4 and IPv6 visibility, observed neighbours and reporting-peer visibility should be read together. The captured response gives zeros across those fields. If one field changes while others remain quiet, the discrepancy is worth preserving rather than smoothing away. Different fields may update on different evidence, and an emerging route may not immediately produce a complete picture.

The historical markers deserve their own comparison. The present response records 103.175.14.0/23 first seen on 24 December 2021 and last seen on 2 May 2022. A later query might preserve those dates, extend the last-seen marker or present a new visible pair. Any change should be treated as an update to observed history, not retroactive proof of what happened during the unobserved interval. The earlier captured response should remain part of the chronology so that changes in the returned data can themselves be examined.

Cross-endpoint consistency is the fifth step. At present, “not announced,” an empty prefix list, zero visible address space, zero neighbours and zero reporting-peer visibility point in the same direction. If a future overview says announced while the prefix list remains empty, the right response is to note timing and endpoint differences, repeat the query after a measured interval and avoid a premature conclusion. Consistency is evidence; inconsistency is a question, not permission to choose the more dramatic field.

Scope should be written beside every result. The operative phrase is “visible in the captured RIPE RIS view.” That phrase prevents a public-routing observation from being mistaken for a universal scan. If evidence from another observation system is later introduced, it should be named and compared rather than silently merged. Different vantage points may see different paths, and the comparison is useful only when the origin of each observation remains clear.

Administrative and operational changes should then be separated in the timeline. A holder-label event belongs to the administrative side. A prefix appearing or disappearing belongs to the observed routing side. A contact response, if independently tested, would belong to another operational layer. A physical-site finding would belong to yet another. Keeping these layers separate makes it possible to say what changed without claiming that one event caused all the others.

Finally, monitoring should have a disciplined stopping rule. If all current fields remain unchanged, the result is simply a repeated quiet observation. It does not accumulate into proof of abandonment. If a route reappears, the result is a new routing observation requiring identity and security checks. If the APNIC record changes, the result is a new administrative fact requiring its own context. In every case, observation comes before explanation.

Sources