Cl Internet Service Private Limited's 20-subscriber Ghaziabad footprint depends on Netway and unverified field repair

Summary

• The Department of Telecommunications still listed CL Internet Service Private Limited in 2026 with a Category C internet-service authorisation for the Ghaziabad Secondary Switching Area. That establishes a legal operating boundary and permission to provide access; it does not establish a regional footprint, fibre ownership or physical route diversity.
• TRAI's report for the quarter ending June 2025 attributed 20 broadband subscribers to the company. Later reports group small providers together, so 20 is the latest individually visible figure, not a claim about the current total. It is enough to require a sharp downgrade from the former website's claim of thousands of kilometres of dedicated dark fibre.
• APNIC assigns AS149562 and the IPv4 block 103.186.222.0/23 to CL Internet Service. On 10 July 2026, however, AS149562 had no visible prefix or neighbour, while the address block was originated by Netway Internet's AS135223 under a valid route-origin authorisation. The public route therefore supports upstream dependence, not CL-controlled multi-homing.
• The former clinternet.in site advertised 50, 75 and 100 Mbps unlimited fibre plans, leased lines, round-the-clock service and prices beginning at INR500 a month. The domain no longer resolves and the official .in registry says it is available for registration. The old page also contained mismatched testimonials and generic claims, making it unsuitable as proof of present service or network scale.
• No public evidence identifies CL's access route, fibre kilometres, poles or ducts, network room, upstream circuit, backup-power duration, technicians, spares, serviceable addresses, utilisation or restoration performance. The defensible description is a tiny licensed Ghaziabad ISP whose limited visible internet resources depend on another operator, with current customer service and recovery capacity unverified.

Twenty subscribers are a scale test, not a footnote

The most informative public number attached to CL Internet Service Private Limited is not a speed, a route length or a capital figure. It is 20. In its April-June 2025 performance report, the Telecom Regulatory Authority of India listed the company with 20 broadband subscribers and no narrowband subscribers at the end of the quarter. The table was compiled from service-provider submissions. It placed CL near the very bottom of a list extending through more than 1,200 named providers.

Twenty is not zero. It is positive operating evidence. A regulator was receiving a subscriber return under the company's legal name, and the return classified every reported connection as broadband. That is stronger than a business directory, a social-media page or an untested sales form. It supports a narrowly framed conclusion that CL had a small body of internet subscribers when the figure was reported.

It does not support a regional-network narrative. The distinction matters because the company's former website made a much larger physical claim. It said services were delivered over "thousands of kilometers" of dedicated dark fibre running at 10 to 100 gigabits per second. Even without deciding whether that text referred to owned fibre, leased capacity, supplier infrastructure or generic sales language, the magnitude demands corroboration. No route map, fibre inventory, building list, carrier agreement, commissioning notice or audited capacity statement accompanies it.

The 20-subscriber count is also dated. TRAI's later reports no longer identify every very small provider by name. The January-March 2026 report lists providers above 10,000 subscribers in its detailed annexure and groups the remainder. CL's absence from the named annexure therefore does not mean that its count fell to zero. It means the later publication cannot be used to update the company-specific total. Twenty is the latest individually visible public count, not a current customer census.

That limitation does not rescue the broader claim. If CL had expanded from 20 accounts into a substantial regional access network after June 2025, one would expect at least one new operating signal: a working order page, an address checker, a larger visible route, an active company-originated prefix, a network map, new tariffs, customer notices, procurement, permits, job openings or current published contact points. The public evidence supplies none of those. Instead, the company domain has lapsed and the company's own ASN remains invisible in public routing.

The right downgrade is therefore precise. CL was not merely a name on a licence list. The subscriber return and routed address space point to service. But the demonstrated scale is closer to a small cluster, building, reseller arrangement or early local launch than to a regional ISP with a self-operated backbone. The exact access arrangement remains unknown, and none of those possible structures should be selected without direct evidence.

The subscriber series shows motion without a breakout

The 20-account figure is not an isolated oddity. TRAI's April-June 2024 report listed CL with 18 broadband subscribers. Its July-September 2024 report listed 19. The April-June 2025 report then listed 20. At the dates for which the company is individually visible, the series moves by two subscribers across a year rather than by hundreds or thousands.

These figures should not be forced into a monthly growth chart. The reports are quarterly snapshots built from provider returns, and the sequence does not disclose activations, cancellations or temporary suspensions between reporting dates. There may have been more gross additions and losses than the net change suggests. The absence of an individually named figure in a later report also breaks the series. What survives is a scale observation: the named base remained in the high teens or 20 at three separated quarter ends.

That scale changes the meaning of every infrastructure claim. If all 20 customers were households on a 100 Mbps plan, their aggregate headline access rates would total 2Gbps, but they would not necessarily demand 2Gbps simultaneously. If one account were a leased line, it could consume more committed capacity and revenue than several residential accounts. If multiple people shared a building contract counted as one subscriber, the number of end users could be higher. TRAI's row does not answer those questions.

It does answer a more basic one. The former website's language about thousands of kilometres cannot be treated as a measure of a CL-owned retail access network merely because a small number of subscribers was reported. A route of that scale would need a separate economic explanation: wholesale carriage, infrastructure held for another business, leased use of a supplier backbone, a franchise structure or text describing the wider network from which CL bought service. No such explanation is public.

The slow visible movement also makes retention important. At 20 accounts, losing one subscriber is five per cent of the reported base. One prolonged outage in a shared building could therefore affect both service quality and a meaningful portion of recurring revenue. Conversely, adding one apartment block could transform the count quickly. Small-provider economics are discontinuous in this way: a single property agreement, repair failure or upstream dispute can matter far more than a nationwide market statistic.

The series supports neither dismissal nor inflation. A tiny base can be real and worth serving. It can also be too small to fund duplicated routes, specialist staff and large spare inventories without support from another business. That is why Netway's role, building concentration and repair arrangements matter more here than the nominal maximum speed on the old tariff.

The authorisation fixes the outer boundary at Ghaziabad

The legal identity is unusually clear. The Department of Telecommunications' 2026 list of internet-service authorisations includes CL Internet Service Private Limited under licence number DS-11/410/2021-DS-III. It identifies Amit Kumar as director, gives the registered address as 2C/512 Vasundhara, Ghaziabad, and records a Category C authorisation for Uttar Pradesh West, Ghaziabad, effective on 5 April 2022. The same licence appears in the department's January 2025 list, which helps distinguish continuity from a one-off listing.

Category C is a geographic permission, not a quality grade. The department's internet-service guidance explains that Category A covers the national area, Category B a telecom circle or metro area, and Category C a Secondary Switching Area. The Unified Licence agreement uses the same definition. CL's authorised territory is therefore the Ghaziabad SSA, not all of Uttar Pradesh, the National Capital Region or India.

Permission and deployment are separate. The licence allows an ISP to establish last-mile links using fibre, radio or underground copper, to use an authorised cable operator's network for last mile, and to lease bandwidth from another authorised provider. It also makes the licensee responsible for installation, operation, billing and subscriber complaints while permitting infrastructure sharing. Those options explain why an authorisation cannot reveal asset ownership. CL could own distribution fibre, lease strands, buy a managed handoff, use a cable partner's plant or combine those approaches.

The former sales page narrowed the commercial focus further. It described service in Vasundhara and Mohan Nagar and gave the same Vasundhara postcode, 201012, found in the licence list. That is a credible local marketing radius inside Ghaziabad. It is not a coverage map. There were no serviceable-building polygons, street lists, tower coordinates, optical-line-terminal locations or installation-time promises tied to particular addresses.

The location distinction changes how failure should be understood. A Category C licence does not imply a long rural network spread across a state. Vasundhara is an urban area with multiple buildings, utility feeders and possible access routes. A tiny provider there could serve a residential block, a group of small offices or a handful of dedicated circuits without constructing a broad outside-plant network. Such concentration can lower the amount of fibre needed per customer. It can also concentrate every account behind one building entrance, one switch, one power supply or one upstream tail.

No public material identifies a point of presence. The registered office may contain active equipment, or it may be an administrative address. The customer handoff may be in the same building, in a partner facility, at a nearby exchange or at an upstream's node. Treating 2C/512 as a network room would be an invention. The address fixes corporate and service-area context; it does not fix the physical topology.

The old website is evidence of an offer, not of a network

Before the domain lapsed, the CL Internet Service website presented a recognisable retail proposition. It advertised unlimited fibre plans at 50 Mbps for INR500 a month, 75 Mbps for INR552 and 100 Mbps for INR652, with tax extra. It promoted broadband, leased lines, intercom or telephony and IPTV. It said there was no fair-usage limit, no concealed fee and round-the-clock service. It also described fibre reaching the neighbourhood and a dedicated-fibre platform operating at 10 to 100 gigabits per second.

Those claims help reconstruct what the company wanted to sell. They do not show what was installed. A tariff can be published before a building is connected. A 100 Mbps plan can be delivered over a gigabit passive optical network, Active Ethernet, a wireless bridge, a cable partner or an upstream-managed service. "Fibre" can describe the customer's final connection, the feeder, the middle mile or only a supplier path beyond a local wireless hop. The page did not separate those layers.

The page contained strong reasons for caution even when it was accessible. One testimonial said a different provider, FusionNet, installed the connection. Two others praised translation work rather than internet service. The page referred to "Naked DSL" and repeated generic design-agency language. Client images appeared as repeated Canon labels. None of this proves that CL lacked customers. It indicates that parts of the page were adapted from unrelated material and were not reliable enough to validate engineering claims.

That weakness is now compounded by disappearance. On 10 July 2026, clinternet.in did not resolve. The National Internet Exchange of India's official RDAP response for the domain returned a not-found status and said the name was available for registration. The former tariff, support and contact paths consequently failed too. A lapsed domain can result from an administrative oversight, a rebrand, a hosting migration, a billing problem or an end to public operations. It is not proof that customer circuits were switched off.

It is nevertheless an operational failure in its own right. A provider that uses a website for tariffs, support and contact loses a public channel when the domain lapses. Customers may still have telephone or messaging contacts, and network equipment can continue forwarding packets without a company website. New customers cannot use the old page to confirm prices, request installation or find current terms. A third party could also register the former domain, making old links unsafe as an identity signal.

The domain's status creates a simple evidence hierarchy. The old page can establish what was advertised. The licence can establish what was authorised. TRAI can establish a reported subscriber count at a stated date. APNIC can establish control of number resources. Public routing can show how those resources are presently announced. None of those layers, alone or together, can establish fibre ownership, present orderability or restoration performance.

One small address block reaches the internet through somebody else's ASN

CL does possess company-specific internet-number resources. The Asia Pacific Network Information Centre's record for AS149562 names CLISPL-AS-IN and Cl Internet Service Private Limited. APNIC registered the autonomous system in April 2022 and continued to mark it active after a 2025 update. Its IPv4 record assigns 103.186.222.0 through 103.186.223.255 to CLISPL, a /23 containing 512 addresses.

Those assignments are meaningful preparation for an ISP. Public addresses let a provider number infrastructure or customers without relying entirely on an upstream's address pool. An ASN can let the provider express its own routing policy and connect to one or more external networks. Neither resource is proof that the intended routing design was commissioned.

On 10 July 2026, RIPEstat's overview of AS149562 marked the ASN as not announced. Its announced-prefix result returned no prefixes, while its routing-status result showed zero IPv4 and IPv6 address space announced and visibility from none of the service's route-collector peers. The neighbour result returned no observed adjacent networks. PeeringDB's ASN query returned no network entity.

The addresses are not dark. RIPEstat's prefix overview for 103.186.222.0/23 showed the block being announced by AS135223, Netway Internet Pvt Ltd. The /23 had been visible through Netway in its two /24 components for years, and the aggregate appeared in June 2026. APNIC's record for AS135223 identifies Netway Internet. RIPEstat's route-origin validation marked the announcement valid because a route-origin authorisation permits Netway to originate the /23 and its /24 components.

That arrangement is technically coherent. A customer ISP can ask an upstream to originate its address block. Netway may be supplying transit, managed routing, wholesale access or another service. The visible BGP origin means routers elsewhere see Netway, not CL, as the network originating the addresses. It does not prove that Netway owns CL's local access plant, bills CL's subscribers or runs the equipment in Vasundhara.

The public path also does not reveal physical diversity. Netway's observed neighbour table includes several larger networks, but those are Netway's external relationships. They cannot be counted as CL upstreams. CL's /23 may reach Netway over one fibre tail, two tails in one duct, a wireless handoff, a Layer 2 circuit, a shared local switch or a fully managed service. One valid global origin can sit on top of any of those physical arrangements.

This is why the routing evidence supports a downgrade rather than a closure claim. CL's own ASN is unused in the visible routing table, yet its addresses remain reachable through another registered operator. That is consistent with a dependent service. It does not establish whether the 20 reported subscribers remain active, but it prevents the mistake of describing CL as a visibly multi-homed autonomous regional carrier.

A /23 is installed potential, not usable customer capacity

The address block sounds large beside a 20-subscriber count: 512 IPv4 addresses for 20 reported broadband accounts. The comparison is tempting and mostly unhelpful. Address count does not measure bandwidth. Addresses may be reserved for routers, servers, network management, future customers or one-to-one customer assignments. Some may be unused. Subscribers may instead sit behind carrier-grade network address translation, allowing many devices to share fewer public addresses.

The same caution applies to the old site's 10-to-100-gigabit language. A 100Gbps-capable optic, a supplier backbone, a switch backplane and a paid transit commit are different things. A provider can connect a 100 Mbps customer to a gigabit access port while buying far less aggregate upstream capacity, because customers do not all use their headline rate at once. Sensible oversubscription keeps broadband affordable. Excessive oversubscription produces congestion when simultaneous demand exceeds the feeder or upstream link.

No public figure discloses CL's committed upstream rate, burst ceiling, access split ratio, optical budget, port count, traffic profile or busy-hour utilisation. The TRAI quality regulations provide a useful standard for what matters: latency, packet loss, jitter, delivered speed and bandwidth utilisation between customer-serving nodes and the ISP gateway or exchange link. They also establish reporting thresholds and measurement obligations. CL is too small to appear as a named large-provider performance table, so the national reports supply no company-specific result.

The published tariffs can illustrate the economics only if their limitations stay visible. Twenty customers all paying the old INR500 entry rate would produce INR10,000 of headline monthly rental before tax. At INR652 each, the figure would be INR13,040. Those are arithmetic scenarios, not revenue estimates: the June 2025 subscribers may have bought different plans, leased lines may be priced separately, installation fees may apply, and the current account total is unknown.

Even the higher illustration has to cover more than internet transit. A small provider needs customer equipment, fibre or wireless access, an upstream handoff, switching and routing, power, billing, support, field visits, replacement parts, licensing compliance and bad-debt allowance. If the operator owns outside plant, it also faces permits, right-of-way charges, cable placement, splicing and restoration. If it leases most of the chain, capital expenditure falls but recurring supplier dependence rises.

Scale can work in a dense building. Twenty accounts behind one short feeder and one managed upstream can be economical if installation costs are low and support is local. The same customer count spread across distant streets can be uneconomic because every drop, visit and cable fault consumes labour. The public evidence does not locate the subscribers, so neither a favourable building cluster nor a difficult scattered footprint can be assumed.

What can be said is that usable capacity is unverified at every layer. The /23 is registered and globally visible. The old tariffs describe intended access speeds. The licence permits local links. No measurement connects those facts into a demonstrated end-to-end service at a current address.

The ownership boundary runs through licence, access plant, power and transit

The company owns or controls some things with confidence. It is the named licensee. APNIC assigns it AS149562 and the /23. It was responsible for the subscriber relationship reported to TRAI. Those are legal and administrative control surfaces.

The physical ownership surface is blank. There is no public list of CL-owned cable, ducts, poles, towers, rooftop radios, cabinets, optical line terminals, aggregation switches or customer terminals. The Unified Licence permits an ISP to establish fibre or radio last mile, to use an authorised cable network and to lease bandwidth. The most realistic description is therefore conditional: CL must arrange an access path to each subscriber, but it may own, lease or share each segment.

The power boundary is easier to locate. Ghaziabad district's public-utilities page identifies Paschimanchal Vidyut Vitran Nigam Limited as the electricity distributor. PVVNL's Ghaziabad feeder directory lists the 201012 Vasundhara area across several urban substations and feeders. The company does not control that public grid merely because its office and possible network equipment sit within it.

Powering a fibre service requires more than energising the glass. A building switch, optical line terminal, router, media converter, cooling fan and customer Wi-Fi device all need electricity. Passive optical splitters can remain unpowered, but the equipment at both ends cannot. A provider may install batteries or a generator at its network point, while the customer may have no backup at home. The service can therefore fail at the premises even when the provider's core remains online.

Local maintenance is not hypothetical. A PVVNL planned-shutdown notice for 24 June 2026 listed one-hour interruptions on several Vasundhara feeders for line shifting, yard maintenance and tree-branch work. The notice does not identify CL's address or say its equipment was affected. It demonstrates that controlled maintenance is part of the local electrical environment. A credible network design must carry critical equipment through such windows or communicate the interruption.

The transit boundary is Netway. Public routing shows Netway originating CL's addresses; it does not publish the commercial contract, handoff site or restoration commitment. If the handoff fails, CL may need Netway or another transport supplier to restore it. CL can replace customer equipment directly only if it has technicians, access and spares. Every outsourced layer reduces the equipment CL must own while adding an escalation boundary.

This layered ownership is why responsibility cannot be read from a company name. The licensee remains accountable to subscribers even when a pole owner, landlord, electricity distributor, cable partner or upstream controls the failed component. Fast recovery depends on agreements that let the responsible person diagnose, enter the site, replace hardware and escalate the external fault without ambiguity.

A cable cut tests maps and labour before it tests bandwidth

If CL's access is underground fibre, the decisive risk is not the theoretical capacity of a strand. It is whether another excavation crosses the route and whether the provider can find and repair the damage. India's Department of Telecommunications created Call Before u Dig to connect excavators with utility asset owners and reduce damage from uncoordinated work. The service sends location and schedule information to relevant owners, but it does not splice a severed cable or guarantee that every asset has been mapped.

The Telecommunications Right of Way Rules, 2024 cover underground and overground networks, permissions, notices and compensation for damage. They make repair and restoration expenses visible in the compensation framework. The rules clarify rights and processes; they do not eliminate the physical outage between damage and repair.

A small fibre operator needs accurate as-built drawings, route markers, optical test equipment, spare cable, compatible closures, splice trays, a splicing machine, traffic-control arrangements and people who can reach the site. If the damage is inside an apartment or office complex, landlord access and riser keys may become the longest delay. If the plant is leased, the owner may insist on dispatching its own crew.

If the local link is radio rather than fibre, the failure moves upward. A rooftop power supply, antenna mount, radio, Ethernet surge protector or line-of-sight path can fail. Building access, weather, interference and replacement alignment determine restoration. The old website's use of "fibre" is not enough to exclude a radio segment because it did not publish an end-to-end architecture.

No public evidence identifies CL technicians, contractors, shift coverage, a network-operations centre, vehicle fleet or spare inventory. The former page's promise of 24-by-7 service describes availability of help, not the number or location of repair staff. One person can answer a telephone around the clock through forwarding; that does not place a trained splicer beside a damaged feeder.

Small scale can improve or worsen response. A local owner who knows every cable and customer may diagnose a fault faster than a national queue. But one specialist can also be a single point of failure. Illness, simultaneous incidents, travel, locked premises or unavailable parts can lengthen restoration. The relevant evidence would be a response target, staffing arrangement, spare policy and incident history. CL publishes none.

Upstream loss is the most visible common failure

The clearest public dependency is the route through AS135223. If Netway withdraws 103.186.222.0/23, or if its session carrying the route fails, external networks lose a path to those addresses after BGP convergence. Services using the block become unreachable from networks with no alternate route. A still-powered local access network may then look healthy to CL while customers cannot reach the wider internet.

The effect depends on address use. If subscriber traffic uses the /23, the outage can affect browsing and inbound connections. If the block numbers only infrastructure while customers use upstream-assigned addresses, the impact differs. If CL receives a default route over one managed circuit, loss of that circuit can interrupt service even while Netway continues announcing the prefix globally. Public BGP sees the internet-facing announcement, not the state of the private handoff into Ghaziabad.

The valid route-origin authorisation is positive security evidence. It tells participating networks that Netway is permitted to originate the CL block. It reduces the risk that a route is rejected as an unauthorised origin. It does not create a second path. One authorised origin can still be one supplier, one edge, one port and one fibre.

CL's inactive AS149562 is important here. If it were visible through two upstreams, route collectors could offer at least logical evidence of multi-homing. They show no CL-originated route at all. The absence does not prove that there is one physical circuit, because multiple circuits could feed a managed Netway service. It means the public table provides no independent failover path that CL controls through its own ASN.

Netway's larger neighbour set cannot substitute for CL diversity. Netway may have several upstreams and robust internal routing, while CL reaches Netway through one fragile tail. Conversely, CL could buy two physically separate tails into Netway and still appear under one origin. The recovery question sits at the handoff: where are the circuits, do they share conduit or building entrance, what powers both ends, and can the surviving path carry the busy-hour load?

No service-level agreement answers those questions publicly. There is no named second upstream, exchange membership, failover test, path diagram or committed repair time. The correct status is not "single circuit proven." It is "single-provider visible and physical diversity unverified."

Six outages would reveal the actual operating system

The first revealing outage is a customer drop failure. One bent fibre, damaged connector, failed optical terminal or indoor cable can disconnect a single home while every shared component remains healthy. Recovery depends on appointment scheduling, premises access, a replacement terminal and a technician. For a 20-account provider, one fault represents a material share of the reported customer base even though the wider network remains online.

The second is a building or neighbourhood feeder cut. This is the event that exposes concentration. If all accounts share one entrance fibre, splitter or Ethernet switch, one break can remove most of the customer base. A ring matters only if the alternate route enters from a genuinely separate direction and active equipment can switch traffic. Two fibres in one cable do not provide cut diversity.

The third is a power interruption. PVVNL can maintain the distribution system while CL maintains its own backup. If batteries support the network room for two hours but the outage lasts four, service stops at exhaustion. If central equipment survives but customer routers do not, support calls may still rise. A useful disclosure would state backup runtime separately for the upstream handoff, aggregation equipment and customer terminal.

The fourth is upstream loss. The local link may remain synchronised while the Netway handoff, transport circuit or routing policy fails. Diagnostic separation matters: technicians should be able to distinguish optical access, local switching, DNS and external routing before dispatching to the wrong site. The visible Netway origin makes supplier escalation a central part of CL's recovery plan.

The fifth is congestion rather than total loss. Twenty users can saturate a modest link if several transfer large files or stream at once, especially after one of two intended links fails. Customers experience low speed and high latency while monitoring still labels the circuit up. TRAI's utilisation and delivered-speed measures are designed to expose this condition, but no CL-specific results are public.

The sixth is organisational failure. The lapsed domain already shows that a support surface can disappear without a route withdrawal. A telephone can change, a supplier account can lapse, a key technician can be unavailable, or a configuration backup can be missing. Networks are not restored by optics alone. Billing access, escalation authority, current contact lists and documented configurations are operating assets.

These events affect different people. A drop failure affects one subscriber. A common feeder or building switch can affect a cluster. A Netway handoff failure could affect every customer whose traffic depends on it. A domain lapse affects prospective customers and anyone relying on web support, while existing packet forwarding may continue. Because the current subscriber total and address use are not public, impact cannot responsibly be converted into a current household count.

The old tariff makes density and repair the decisive economics

At INR500 to INR652 a month, the former residential plans were priced for a mass-market household, not a bespoke carrier circuit. Such pricing works when many customers share feeder construction, upstream capacity and support. It struggles when each connection requires a long route, repeated civil work or frequent truck rolls.

The June 2025 count makes density the key unknown. Twenty customers in one apartment complex could share one building feed and produce low installation cost per account. Twenty customers distributed across Ghaziabad could require many kilometres of access and expensive visits. The old claim of thousands of kilometres would be economically implausible as a CL-owned network supported only by the visible subscriber base unless substantial wholesale revenue, leased-line revenue, shared infrastructure or external capital existed. None is disclosed.

Local support labour is therefore part of the bill, not an optional service feature. A low monthly rental leaves little room for repeated visits. Good installation quality, remote monitoring and standardised customer equipment reduce future cost. Poor records and mixed hardware do the opposite. A small provider can compete by being nearby, but only if nearby expertise is actually available when a common feeder fails.

Upstream purchasing has the same scale effect. A tiny operator may buy a managed service because operating BGP, peering and redundant transport directly would cost more than the customer base can carry. The Netway-originated route is consistent with that choice. It can be economically sensible and technically stable. The trade-off is dependence on Netway's delivery, escalation and commercial continuity.

The lapsed website weakens acquisition economics. A local provider can sell through referrals and telephone calls, but a non-resolving domain removes the lowest-cost place to publish tariffs, verify identity and capture enquiries. Rebuilding that channel is inexpensive compared with laying fibre. Its absence is therefore notable: it suggests either limited current marketing attention or a move to an undisclosed identity.

None of this establishes insolvency, closure or bad service. Twenty customers can be deliberately small. A managed upstream can outperform a poorly run autonomous edge. A local technician can offer excellent support. The problem is evidentiary: the public footprint provides no present operating facts with which to test those favourable possibilities.

What would move the assessment above Weak

The first requirement is a current service surface. A working company domain, an address-level availability check, current tariffs and a support escalation path would show that the retail offer is still maintained. The service checker should distinguish immediately orderable addresses from areas that require network extension.

The second is a bounded physical description. CL need not publish sensitive coordinates. It could state the number of live buildings or neighbourhoods, whether the last mile is owned, leased or partner-operated, and whether it is fibre end to end. A route-mile range, count of powered aggregation sites and description of the access technology would prevent old marketing language from standing in for installed plant.

The third is an honest upstream statement. CL could say that Netway originates its /23, identify the handoff arrangement at a high level and state whether a second physically diverse circuit exists. If AS149562 is intentionally dormant, explaining why would be more useful than implying independent routing. If it is to be activated, a visible route and documented origin policy would supply evidence.

The fourth is recovery information. Backup runtime, spare-equipment location, repair coverage, supplier escalation targets and a measured failover result would answer the power and labour questions. Monthly availability, median restoration time and longest incident could be published without exposing customer identities.

The fifth is scale reconciliation. A current subscriber count and a clarification of the "thousands of kilometres" claim would settle the largest contradiction. If the phrase referred to Netway or another wholesale backbone, the ownership boundary should be named. If CL owns extensive fibre, a broad route map, fixed-asset disclosure or permit history should exist. If the text was generic, it should be withdrawn.

Until those facts appear, the visible evidence supports only a modest description. CL Internet Service Private Limited holds a live-looking Category C authorisation for Ghaziabad, controlled a small address allocation, and reported a tiny broadband base. Its addresses reach the internet through Netway rather than its own ASN. Its former public domain is now unregistered. The physical access plant, current customers and recovery system cannot be verified.

That is a Weak network evidence grade and a downgrade from the original regional-ISP hypothesis. It is not a declaration that the company never operated, nor proof that every circuit has ceased. It is a refusal to turn a licence, a /23 and old sales copy into a regional network that the public facts do not show.