Summary

  • The strongest corporate evidence points away from a current operator. India's Ministry of Corporate Affairs included AIR ZONE INTERNET PRIVATE LIMITED, CIN U64200DL2015PTC281559, in a 2019 notice of striking off and dissolution. The legal name differs slightly from the AirZone label in internet-number records, but the shared Delhi geography and Pankaj Goswami identity make the connection strong.
  • The number resources remain administratively current. APNIC marks AS134852 and 103.205.144.0/22 active, with Delhi contacts and changes recorded in 2025. IRINN still lists AirZone internet Service Pvt. Ltd. as a Delhi affiliate. Those records show resource stewardship, not a live retail network.
  • Public routing ended later. RIPEstat first saw AS134852 originate a route in April 2016, last saw an AirZone origin on 26 March 2024, and saw no IPv4 route, IPv6 route or neighbouring network for the AS on 10 July 2026. The four /24s in AirZone's /22 were all unannounced at that observation point.
  • One /24, 103.205.146.0/24, was briefly observed behind Excitel's AS133982 after AirZone's own origin disappeared, with a last observation in December 2024. That is evidence of a changed route origin, not proof of an acquisition, customer transfer, address lease or continuing AirZone service.
  • The final network evidence grade is Negative for a current regional-ISP thesis. There is no current public proof of an AirZone licence, orderable service, subscribers, access plant, field workforce, upstream contract, physical route diversity, backup power, outage response or restoration performance. The supported subject is a historical Delhi routing edge whose administrative records outlasted both its corporate record and its visible routes.

The first infrastructure question is whether an operator remains

A small internet provider can be difficult to see from outside. It may have no investor page, no press office and no polished network map. Its useful evidence can instead be a licence, a tariff, an address checker, a customer portal, a route announcement, a maintenance notice or a technician's published contact points. AirZone has traces in several of those administrative and technical categories, but the traces do not converge on a current service.

The corporate record is the place to begin. A Ministry of Corporate Affairs Form STK-7 notice from 2019 is titled "Notice of Striking Off and Dissolution" and includes AIR ZONE INTERNET PRIVATE LIMITED, CIN U64200DL2015PTC281559, at item 376. An earlier Form STK-5 notice includes the same company and CIN among the names proposed for removal. The legal effect matters more than the typography: this is stronger evidence than a commercial company page that has not refreshed its status.

That official record is also a warning against treating every similar online label as identical without a bridge. The company register uses AIR ZONE INTERNET PRIVATE LIMITED. APNIC uses AirZone internet Service Pvt. Ltd. for AS134852 and its address block. The names differ by spacing and by the word Service. The connection is nevertheless strong because the records converge on South West Delhi and on Pankaj Goswami. Secondary company records identify Pankaj Goswami as one of the 2015 directors, while APNIC's administrative contact is Pankaj Goswami at Ramphal Chowk, Dwarka. The original corporate address was also near Ramphal Chowk in Sector 7, Dwarka.

The connection is strong enough to analyze the records together, but not strong enough to erase the naming difference. A current certificate of incorporation, restoration order or operator statement would be the cleanest way to settle whether the AirZone network label was a trading style of the struck-off legal company, whether another company later took over the resources, or whether the registry name simply remained after the corporate name was removed. None of those later documents was established here.

The operating conclusion must therefore start narrowly. Air Zone Internet was a real Delhi company incorporated in 2015, and an AirZone-labelled network received an autonomous-system number and portable IPv4 space in 2016. The corporate name was later struck off. The network continued to appear in public routing for years after that. By July 2026 it no longer did. This sequence supports a historical infrastructure story, not a present-tense claim that households or businesses can buy AirZone broadband.

That distinction changes every other question. It would be premature to ask whether AirZone has enough repair crews or diverse upstreams before proving that AirZone has a current access network, current customers and a current licence. A resilience analysis cannot begin by assuming the service whose resilience is under examination.

The company and the route followed different timelines

Air Zone Internet Private Limited's incorporation and dissolution sit on one timeline. The public internet edge sits on another. APNIC's RDAP record for AS134852 dates the autonomous-system registration to 1 January 2016, calls it AIRZONE-AS-IN, gives the country as India and marks the resource active. APNIC's record for 103.205.144.0/22 dates that portable allocation to the same day and marks it active under the name AIRZONE-IN.

RIPEstat's historical observations show that the number was not merely allocated and forgotten. Its routing-status result for AS134852 first sees a route originated by the AS on 19 April 2016. The long routing-history view shows the four /24s inside 103.205.144.0/22 behind AS134852 for long periods from 2016 into March 2024. It also shows other prefixes at different times, demonstrating that the AS had a wider historical routing role than the allocation still registered in AirZone's name.

The official corporate strike-off notice arrived in 2019. AS134852 nevertheless remained visible well beyond that year. That does not overturn the corporate notice. A dissolved company can leave contracts, assets, receivables and technical configurations that take time to unwind. A network can be operated temporarily by another party, or its resources can remain in use while authority is being regularized. A company can also be restored to a register after a strike-off if the required legal process occurs. The route table cannot identify which explanation applies.

What it can establish is the sequence. The company was listed in the 2019 striking-off and dissolution notice. AS134852 continued originating routes. RIPEstat's last-seen marker for the AS is 26 March 2024, attached to 103.205.146.0/24. On 10 July 2026, the AS overview marked AS134852 unannounced. The announced-prefix result returned an empty list, and the routing-status result counted zero observed neighbours and no announced IPv4 or IPv6 space.

That gap is now measured in years from the corporate action and more than two years from AirZone's final observed route. It is materially different from a maintenance withdrawal lasting minutes or hours. It still is not proof that every old switch, fibre strand or customer account ceased on the same day. Public BGP records the reachability of address space through an autonomous system, not the state of every local device. A network can renumber under an upstream carrier and disappear as an independent origin. Customers can migrate while the old AS is retired. A local cable system can remain powered but isolated from the public internet.

No evidence found here establishes such a migration as an AirZone service continuation. The responsible finding is therefore not that every piece of AirZone-linked equipment vanished in March 2024. It is that AirZone's independent public routing identity stopped then, after the associated legal company had already been struck off, and no current customer-facing evidence closes the gap.

An active APNIC record is not an active broadband line

The most tempting counterargument is administrative freshness. APNIC marks the ASN and /22 active. The ASN record was changed in September 2025, the address record in August 2025, and the incident-response contact in February 2026. IRINN's current affiliate list still includes AirZone internet Service Pvt. Ltd. under Delhi. These are not stale pages untouched since 2016.

They still do not prove retail operation. Internet-number registries answer questions about the administration of scarce number resources. They identify the registered holder, status, maintainers and contacts. They do not certify that a route is visible, that an ISP licence is valid, that customers are connected, that invoices are being issued or that a technician can restore a failed line. The word active in an RDAP status field describes the registration, not the commercial service.

The distinction is visible inside the same evidence set. APNIC says AS134852 is active as a registered resource. RIPEstat says it is unannounced as a public route origin. Both statements can be true. IRINN can continue to recognize an affiliate that retains address space even when that affiliate originates no route. A contact update can reflect housekeeping, a new registry format, an abuse-mailbox refresh or an effort to preserve control of resources. It cannot be converted into a claim that customers returned.

The 2025 changes are especially easy to overread because they occurred after the last observed route. If a network record is maintained, someone appears to retain credentials or authority over it. That is meaningful. It reduces the chance that the record has been wholly abandoned. It may also matter for preventing unauthorized route claims, responding to abuse reports or arranging a later transfer. Yet none of those stewardship functions requires a working last mile.

The current IRINN listing has a similarly bounded meaning. It identifies AirZone among thousands of Indian affiliates and gives Delhi as its location. It does not publish a service territory, subscriber count, licence number, route, facility, exchange port, tariff or support commitment. An affiliate relationship is part of number-resource administration. It is not a substitute for the Department of Telecommunications' authorization list.

This is why the public status should be split into separate statements. Corporate status: the official 2019 notice says the legal company was struck off and dissolved. Number-resource status: APNIC still marks AS134852 and 103.205.144.0/22 active. Routing status: neither the AS nor any part of its /22 was publicly announced at the July 2026 observation point. Retail-service status: unverified, with no current positive evidence found. Combining those statements into one label such as active ISP would discard the most important information.

Four registered /24s now provide no observed public route

The address block is compact and easy to inspect. A /22 contains 1,024 IPv4 addresses before network design, infrastructure use, reservations and assignment policy are considered. AirZone's block runs from 103.205.144.0 through 103.205.147.255. Operators often advertise a /22 as one aggregate, four /24s, or a mixture depending on routing policy and upstream acceptance. Historical observations show AirZone using the four /24s.

The current view is simpler. RIPEstat's prefix overview marks 103.205.144.0/24, 103.205.145.0/24, 103.205.146.0/24 and 103.205.147.0/24 all unannounced. There is no observed origin ASN for any of them. The parent /22 is also not visible as a covering route.

Three of the four /24s share the same final AirZone date in the routing-status history: 26 March 2024. The fourth, 103.205.146.0/24, has an additional chapter. Its routing-status result records Excitel Broadband's AS133982 as its last observed origin, on 19 December 2024. The other three results identify AS134852 as their last origin in March.

That temporary Excitel origin is a signal of operational change, but it is not self-explanatory. An upstream may originate customer space on a customer's behalf. A block may be leased, transferred in use, carried during a migration or announced because of a configuration that later changes. Route collectors show the origin in the path; they do not show the contract authorizing it. The registration remaining in AirZone's name does not establish that Excitel bought the block, acquired the business, served AirZone's former customers or made an error. No such relationship should be declared from one route alone.

The current absence also means the address count cannot be used as capacity. A /22 is installed administrative capacity only in the sense that the allocation exists. At the July observation point, it supplies no observed public reachability. It says nothing about line speed, contention ratio, upstream port size, active subscriber count or the state of customer equipment. Even when the four /24s were visible, 1,024 addresses would not have equalled 1,024 subscribers. Carrier-grade translation can put many customers behind fewer public addresses; business customers can receive several; infrastructure and spare assignments can consume others.

There is no visible IPv6 counterweight. AS134852 announces no IPv6 prefix, and no IPv6 allocation attributable to AirZone was established in the reviewed records. The absence does not prove that every historical customer lacked IPv6 through a tunnel or upstream arrangement. It does mean there is no independent AirZone IPv6 origin to count as current dual-stack service.

Route-origin security is also absent in the present view. RIPEstat's validation checks for the four /24s against AS134852 return unknown, with no validating Route Origin Authorization. Unknown is not invalid. It means relying networks do not receive a cryptographic statement from that system confirming AS134852 as an authorized origin for those prefixes. While the routes are withdrawn, the immediate availability effect is nil, but the missing authorization would matter if a return to routing were proposed.

The last visible upstream was one logical neighbour

Near the end of AS134852's public routing life, route collectors saw one network directly before it in the AS path: AS9730. A RIPEstat neighbour observation for March 2024 identifies AS9730 as the sole left-side neighbour. A BGP-state snapshot from 20 March 2024 shows the same predecessor across the collected paths.

APNIC's record for AS9730 names Bharti Airtel Limited as the registrant and identifies the network as BHARTITELESONIC-AS-IN-AP. This supports a narrow historical statement: immediately before AirZone's routes disappeared, the public path reached AS134852 through one visible Bharti network. It does not reveal the commercial product, circuit count, handoff location, port speed or repair obligation.

One logical neighbour is not necessarily one physical cable. Bharti could have delivered more than one circuit, and multiple sessions could have terminated on separate routers while sharing one AS number. Conversely, two invoices or two router ports can ride the same duct, building entrance, optical span, powered room or upstream chassis. Public BGP is excellent at showing policy adjacency. It cannot certify physical route diversity.

That limitation matters because the planned resilience thesis assumed upstream routes in the plural. The evidence at the end supports one observed upstream ASN, not two. No current AirZone route exists from which to infer any upstream. No public circuit map, colocation listing or service statement identifies a second provider. No independent measurement establishes failover from Bharti to another carrier.

PeeringDB adds an administrative trace but no interconnection map. Its AS134852 network entry was created in October 2020 and updated in October 2022. It names AirZone internet Service and records an open policy, but lists zero facilities, zero internet exchanges, zero IPv4 prefixes and zero IPv6 prefixes. There are no public contacts or exchange LAN entries in the result.

That entry should not be read as evidence that AirZone had no private handoff. PeeringDB participation is voluntary, and many small access providers buy transit without publishing facility details. Its value is negative and bounded: the operator-maintained interconnection directory does not supply a public facility, exchange port or alternate path that could support a current diversity claim.

CAIDA's ASRank result for AS134852 reaches the same current conclusion from inferred AS relationships. It marks the AS unseen, with zero providers, peers, customers, prefixes and addresses in its current graph. CIDR Report labels AS134852 not announced. These are corroborating views, while RIPEstat's time-stamped history remains the clearest basis for the dates.

Licensing evidence does not support a current ISP claim

India's Department of Telecommunications explains that an ISP authorization permits a provider to offer internet access to households, enterprises and other users. It distinguishes nationwide Category A, state-level Category B and district-level Category C scopes, and describes fibre, DSL and wireless broadband as possible delivery technologies. Those are general licensing categories, not an AirZone service description.

The current company-specific evidence is negative. Air Zone Internet does not appear in the Department's UL ISP and UL ISP VNO list as of 28 February 2026. Searches for the legal and network spellings return no AirZone row. The same is true of TRAI's February 2024 ISP authorization compilation.

Absence from those lists needs precise language. It does not prove that AirZone never held an older authorization, that no affiliated operator served a customer, or that no appeal or restoration exists outside the reviewed material. AS134852's visible operation from 2016 shows that the network did more than reserve a number. Yet the current publication cannot support a current licence number, category or service area for AirZone.

IRINN affiliation does not fill that gap. IRINN is India's national internet registry under NIXI. Its affiliate list establishes a number-resource relationship. The Department of Telecommunications controls the ISP authorization that permits sale of internet access. Treating the former as the latter would collapse two distinct forms of authority.

The company status reinforces the caution. The official strike-off notice is not a regulator's finding about network quality; it is a corporate action. But a dissolved legal person cannot casually be assumed to remain the contracting party for retail broadband. A restoration, successor or licensed operating partner would need to be identified before attributing current customer obligations, access plant or support labour to the directory entity.

NIXI's own 2021-22 annual report offers an unusual confirmation. In a table concerning balances with struck-off companies, it names AIR ZONE INTERNET PRIVATE LIMITED, gives the same CIN and records a small trade receivable balance. That entry does not explain the service or settle ownership of the number resources. It does show that NIXI recognized the legal company as struck off while a financial balance remained.

The evidence therefore supports neither a current Category A, B nor C classification for AirZone. It supports Delhi as the administrative geography and a historical ISP function associated with an autonomous system. A reader should not infer national coverage, Delhi-wide coverage or even one orderable address from the generic word ISP in old records.

Delhi is an identity anchor, not a verified coverage map

Every reliable geographic clue points to Delhi, especially Dwarka. The APNIC contacts use C-917, ground floor, Ramphal Chowk, Palam Extension, Sector 7, Dwarka, New Delhi. Corporate records use C-348 near Ramphal Chowk and later secondary presentations also show C-917. IRINN assigns the affiliate to Delhi. These clues justify an India region and a Delhi operating-history frame.

They do not define a service polygon. A registered office can contain administration without a core router. A network contact can work from a residence or storefront while equipment sits in a carrier facility. An address can remain in a registry after a business moves. Commercial IP geolocation pages variously place addresses from the /22 in Delhi, Mumbai or elsewhere, illustrating why inferred address location is not a fibre map.

No current AirZone coverage checker, tariff sheet, installation form or customer portal was established. There is no public list of neighbourhoods, apartment buildings, enterprises, towers, poles, ducts, cabinets, optical distribution points or wireless sectors. There is no evidence that the company served all of Dwarka, all of Delhi or customers outside the capital.

This absence changes how the physical asset should be described. The historical routes are consistent with an access or business-connectivity network in Delhi, but they do not reveal the access medium. Fibre, Ethernet, cable and fixed wireless are all plausible for a small urban provider. The Department's general ISP page lists several of those technologies, but a national technology menu cannot be turned into an AirZone inventory.

The property boundary is especially important in dense urban service. A broadband line may depend on access to a building riser, a rooftop, a basement room or a street cabinet. TRAI's Digital Connectivity Rating framework treats fibre readiness, in-building systems, network redundancy and power redundancy as separate features of a property's connectivity. That framework is not an AirZone assessment. It shows why an ASN and office address cannot answer whether a particular tenant had a diverse, powered path.

The correct location statement is consequently modest. AirZone's corporate and resource records are tied to Delhi and Dwarka. Its exact historical access footprint, facility locations and customer density are not public. Its current service area is not verified. Global would be an especially misleading region because neither the company evidence nor the routing record supports a global access operation.

The physical network behind the old routes remains unknown

When AS134852 was visible, a working customer connection required more than a BGP announcement. Traffic had to travel from a home, office or hosted system into some local access medium, through powered aggregation equipment, toward an edge router and then into Bharti's network. Each element had an owner and a repair boundary. The public evidence identifies only the resource holder and the last visible AS-level neighbour.

AirZone may have owned some customer equipment and local electronics. It may have leased fibre, rooftop space or transit. It may have relied on local cable partners. None of those arrangements is established. The company record discloses no fixed-asset inventory, and no public network diagram identifies who controlled poles, towers, ducts or cross-connects. The office address should not be called a network operations centre without evidence.

The four /24s do not locate the equipment. Routing all four through one AS could have involved one border router or several. Their common final AirZone date is compatible with a shared edge or deliberate withdrawal, but it does not identify the cause. The later Excitel origin for one /24 could reflect a separate handoff, but there is no proof that the other three prefixes or any access plant followed it.

Power is similarly opaque. Every active optical terminal, switch, wireless radio, router and customer-premises device depends on electricity. Battery or generator support can keep the network side alive during a local cut, but the customer's router and building equipment may still fail. CISA's communications dependency primer describes the general dependence of communications facilities on electricity and on transport for generator fuel. It is a useful engineering model, not evidence that AirZone owned a generator or achieved a stated runtime.

Civil works create another generic but relevant exposure. The Department's Call Before u Dig service exists because uncoordinated excavation can damage optical fibre and other buried utilities. If AirZone used buried fibre, a cut could have isolated a street, building or upstream handoff. If it used aerial cable or rooftop wireless, weather, building access and power would create different failure modes. The unknown medium means the article cannot choose one as the historical design.

Field labour is the largest missing operational asset. No staff list, contractor relationship, dispatch area, service-level commitment, spares inventory or after-hours contact was established. A small provider can outperform a large carrier in a compact neighbourhood because technicians know the buildings and cable routes. The same provider can be fragile if one splicer, one vehicle or one person with rooftop access is unavailable. Without a current service or workforce, either description would be speculation.

The physical conclusion is therefore functional rather than visual. A historical AirZone service would have depended on access plant, local power, customer equipment, an aggregation path, an edge and upstream transit. Public evidence does not show where those assets were, who owned them, how much capacity they carried or whether any remains in service under another name.

Installed, usable and recoverable capacity are three different numbers

The AirZone record contains several quantities that look like capacity: one ASN, one /22, four historically announced /24s and one observed upstream near the end. None is a bandwidth measurement. None identifies subscriber load. None shows what would remain during a failure.

Installed capacity begins with assets actually deployed: fibres, radios, switches, routers, ports, power systems and address resources. Public records prove the address resources and historical routing identity. They do not prove the physical plant. The /22 creates room for 1,024 IPv4 addresses, but addresses do not carry bits by themselves.

Usable capacity is smaller and conditional. It depends on access-port speed, signal quality, contention, aggregation links, transit capacity, routing policy, equipment health and customer demand. A provider can advertise every prefix while suffering evening congestion. It can own a large address block and have few users. It can place many subscribers behind address translation. No AirZone traffic graph, port speed, busy-hour utilization, latency series or customer measurement is available to translate the address block into usable throughput.

Recoverable capacity is what survives the first failure. A second router does little if both routers share one power feed. A second carrier does little if both circuits enter through one duct. Spare transit capacity matters because a backup link that can carry only a fraction of normal demand will preserve route visibility while degrading service. CISA's ten keys to resilient local access networks separates redundancy from route diversity and defines diversity as more than one physical path with no common points. The principle applies here as an audit question, not as evidence about AirZone's design.

The historical single observed AS neighbour does not establish recoverable upstream capacity. It is possible that Bharti provided diverse physical circuits under one ASN. It is also possible that AirZone had another route not visible in the selected observations. Neither possibility is documented. PeeringDB adds no facility or exchange record. The current AS has no route at all, so its observable public capacity is zero prefixes rather than a degraded percentage.

The later Excitel announcement for one /24 does not establish recovery capacity either. It occurred after AirZone's own origin disappeared and did not cover the full /22. It may reflect migration rather than failover. By December 2024 that route also vanished. Calling it a backup without a contract, simultaneous observation or operator statement would be unjustified.

This three-part framework prevents two opposite errors. The first is exaggeration: counting 1,024 addresses as 1,024 active lines or treating one current registry entry as a live network. The second is erasure: assuming that an unannounced AS means no physical asset or customer transition ever existed. The evidence supports retained administrative capacity, historical public routing, and no current observed usable or recoverable capacity under AS134852.

The failure paths can be described, but not attributed as incidents

The original resilience question named access cuts, tower or power outages, upstream loss, field-repair shortages, congestion and single-route exposure. Each is a credible broadband failure mode. None is a documented AirZone outage in the evidence used here.

An access cut would affect customers downstream of the damaged segment. Its blast radius would depend on topology. A point-to-point building feed might isolate one property; a cut before a neighbourhood splitter could affect many; a ring could reroute traffic if the alternate side had power and capacity. AirZone publishes no topology from which to choose among those outcomes.

A power failure could occur at the customer, building, cabinet, tower, edge room or upstream handoff. Backup power at one layer does not protect the others. A provider could keep its core router alive while apartment switches fail, or customers could retain local Wi-Fi while the upstream edge is dark. There is no AirZone backup-power runtime, maintenance test or fuel arrangement to assess.

Upstream loss is the only failure path with a historical logical outline. In March 2024, AS9730 was the single visible predecessor. If that adjacency failed and no hidden or alternate path existed, the four prefixes would lose global reachability. The simultaneous disappearance of several AirZone routes is consistent with a shared routing change, but no public notice identifies an outage, non-payment, migration, equipment fault or planned shutdown. The corporate history makes deliberate retirement at least as plausible as a technical failure.

Congestion would be even harder to see. BGP can remain stable while packet loss and latency rise. Without traffic volumes, transit port size, speed tests or complaint records, there is no basis for calling the old network congested or uncongested. The address block's size supplies no answer.

Field-repair shortage is a labour question, not a route-table field. It depends on crew count, skills, vehicles, spare fibre, optical modules, radios, access permissions and escalation rights. A dissolved company record raises a serious question about who could contract for and direct repairs, but it does not prove that a particular technician or successor was absent.

The broadest failure path is withdrawal of the operating identity itself. If a company is struck off, loses or does not renew its authorization, stops announcing routes and leaves no current customer channel, assurance fails before any cable is cut. Customers, if any remained during the transition, would need to know who held their contract, who owned their equipment, whether their addresses would change and who accepted fault reports. Those questions are not answered publicly for AirZone.

The affected population cannot be responsibly counted

A resilient-service article should identify who loses connectivity when the system fails. Here that population is unknown. No current subscriber number appears in a regulator report under AirZone's legal or network name. No customer list, serviceability map or tariff page establishes homes, businesses, schools, clinics or public bodies on the network.

The absence is not proof of zero historical customers. AS134852's long route history and the /22 allocation are consistent with a functioning network. Commercial geolocation services still associate addresses with AirZone because registration data persist. That historical identity may continue to appear in security logs, allowlists and commercial IP products even when the route is absent.

That persistence creates a different affected group: administrators and researchers who rely on attribution. A fraud system may label an old address as AirZone even after another network originates it. An abuse report may go to an address maintained for a resource whose commercial user has changed. A security analyst may mistake an active RDAP status for current traffic. The temporary Excitel origin of 103.205.146.0/24 demonstrates why holder, origin and end user must be separated.

These are real operational consequences, but they are not a substitute for subscribers. The article cannot claim that a Delhi household lost service in March 2024 or that Excitel took over that household. It can say that public routing for AirZone's independent AS stopped and that one part of its registered block later appeared under another origin.

If customers were migrated, the practical risks would have included renumbering, changed carrier-grade translation, revised support contacts, new billing authority and possible interruption during cutover. If the service simply ended, customers would have needed another provider and removal or replacement of premises equipment. If the old access plant was sold, asset ownership and repair authority would have moved. Each scenario requires evidence that is not public here.

The safest count is therefore none at all. The article should not estimate subscribers from address space, reviews, population density or the size of Dwarka. It should state that the affected user population is unverified and that this missing fact is one reason the current regional-ISP category fails.

Small-operator economics explain why records can outlive service

Air Zone Internet was incorporated with a modest capital base according to public company summaries, but the official strike-off notice and NIXI record are more important than the precise old capital figure. A small access provider faces fixed obligations that do not shrink neatly with subscriber count: licensing, transit, address administration, powered equipment, rent, building access, customer support and field repair.

Density can make the model work. One fibre into a multi-dwelling property can spread construction and backhaul cost across many lines. Local knowledge can reduce installation time and help technicians find faults. Direct relationships with property managers can create a defensible neighbourhood position even when national carriers are larger.

The same concentration can create fragility. Losing one building agreement can remove many customers at once. One unpaid transit bill or one unavailable technician can affect a large fraction of the business. A route that depends on one visible upstream can make commercial and technical dependence converge. These are economic mechanisms, not AirZone-specific findings.

AirZone's public history shows the outcome without exposing the mechanism. The company was struck off. NIXI later reported a small receivable associated with the struck-off company. The AS continued routing for several more years. A contact and address registration remained active after routing ended. That is the shape of an unwind, continuation or transition whose contractual details are not visible.

It would be possible to invent a clean story: perhaps another operator kept customers online, perhaps address space was carried temporarily by Excitel, or perhaps the route simply remained after the retail business stopped. The available evidence does not choose among those narratives. The economics section should therefore explain why such mismatches can happen while refusing to label the actual transaction.

The controlled topic Regional ISP economics was removed from the article metadata because the current evidence does not support AirZone as an operating regional ISP. The economic context remains relevant to interpretation, but the strongest durable monitoring subject is Peering and transit: which AS originated the addresses, through which visible predecessor, and when those routes disappeared.

Local support labour was removed for the same reason. Field repair would have mattered to any historical access operation, but there is no public AirZone workforce to monitor today. A topic should describe an evidenced continuing subject, not merely a plausible dependency.

What would be needed to call AirZone operational again

The corporate threshold comes first. A current Ministry of Corporate Affairs master record or court order would need to show that Air Zone Internet Private Limited was restored, or a successor would need to identify itself and its authority over the AirZone network name and resources. The slight difference between the legal and APNIC names should be reconciled rather than ignored.

The licensing threshold comes next. A current Department of Telecommunications ISP authorization should identify the licensee, category and service area. If AirZone operates under another licensee as a virtual provider or local partner, that boundary should be explicit so customers know which company bears regulatory and service obligations.

The service threshold requires something a customer can use: a current official website or portal, tariff, serviceability check, order channel, terms, support contact and installation commitment. A current subscriber count or regulator submission would strengthen the evidence. None should be inferred from the address block.

The routing threshold would be a new public announcement attributable to AS134852 or a documented arrangement under another origin. The current holder should publish or maintain appropriate route authorization, including RPKI where possible. A return of one /24 would not automatically prove the whole /22 or the whole access network had returned.

The resilience threshold is higher. It would require a logical and physical map of upstream handoffs, separate building entries or duct routes, router redundancy, normal and failover capacity, facility power, battery or generator runtime and regular failover tests. Two AS neighbours would still need physical-path verification. One AS neighbour could still be resilient if it delivered genuinely diverse circuits, but that would need proof.

The field threshold would cover dispatch hours, technician coverage, contractor authority, spares, building access and measured restoration time. TRAI's 2024 quality-of-service regulations provide the broader Indian context for measuring access and broadband performance. They do not provide an AirZone score.

Evidence should also settle the 103.205.146.0/24 episode. An operator statement, contract disclosure or registry update could explain why Excitel originated the prefix after AS134852 stopped. Until then, it remains a route-origin observation with no declared commercial relationship.

Meeting one threshold would not satisfy the others. A restored company can have no network. A live route can carry no retail customers. A licence can exist before construction. A tariff can advertise service where no installation is possible. A credible current profile needs the layers to converge.

The lasting asset is the mismatch itself

AirZone matters because its records resist a simple active-or-dead label. The legal company was struck off and dissolved in an official 2019 notice. Its autonomous system continued routing into March 2024. One registered /24 then appeared briefly under Excitel. By July 2026 all four /24s were unannounced. Meanwhile APNIC and IRINN still carried current administrative records.

Each record is useful when read for the question it actually answers. The corporate notice answers whether the named legal company remained on the register at that point. The Department's licensee list answers whether a current authorization under that name is visible. APNIC answers who holds and maintains the number resources. RIPEstat answers what public collectors observed in routing. PeeringDB shows what the operator chose to disclose about interconnection. None can stand in for all the others.

The present-tense regional-ISP thesis fails because the positive operating bridge is missing. There is no current licence, route, customer offer, subscriber evidence, access map, workforce or measured service. The old title's reference to a local connectivity bill would assume a current billed customer. The evidence does not permit that assumption.

The better title follows the chronology. Air Zone Internet's corporate record ended before AS134852's routes did. That statement is specific, verifiable and useful. It leaves room for the possibility that assets or customers moved without inventing who received them. It also explains why the directory entity can remain worth monitoring even when it should not be presented as a current operator.

The final network evidence grade is Negative for current regional-ISP operation. It is Medium for the historical route and resource account: the corporate action, ASN, allocation, Delhi contacts, route dates, final visible upstream and later Excitel origin are supported by strong public records. The grade remains below Strong because the naming mismatch, post-dissolution operating authority, access plant, customer population, contracts and reason for withdrawal are unresolved.

For a customer, the practical lesson is simple. A familiar provider name, an active address registration and an old ASN page do not guarantee that service can be ordered or repaired. For a network operator, the case shows why corporate authority, licensing, resource registration, route origin and field responsibility need to remain aligned. AirZone's public footprint preserves the pieces. It does not show that anyone has put them back together.