Summary
- Nayet Internet Service is a verifiable Bangladesh network identity: a BTRC list places it in Narayanganj as a legacy Cat-C ISP, and APNIC records it as the holder of AS133148 and the portable IPv4 block
103.66.228.0/23. - The operating-status hypothesis must be downgraded. RIPE RIS saw
103.66.229.0/24behind AS133148 from January 2023, but its last broadly visible observation was on 6 January 2026; on 10 July 2026 the ASN had no announced space and no observed neighbour. - A December 2025 routing snapshot showed one immediate external neighbour, AS58715. That is evidence of one visible path at that moment, not proof of a sole commercial supplier, a permanent relationship, or physically diverse access to any upstream.
- No public evidence establishes Nayet's current subscriber count, retail packages, access technology, fibre or wireless footprint, pole and duct rights, backup power, spare inventory, field staffing, network ring, second upstream, peering presence or restoration performance.
The broadband bill and the missing route
Imagine a subscriber in Sayedpur or Fakir Bari paying for a month of fixed internet. The name on the receipt may be local. The people who install a drop cable, replace a damaged connector or answer a support call may also be local. Yet the useful service exists only when several separate systems work together: power at the subscriber's premises, an access link to a neighbourhood aggregation point, powered equipment at that point, a backhaul path out of the area, domestic exchange for local traffic, international gateway capacity for the wider internet, and a route that tells the rest of the network where return packets should go.
Nayet Internet Service is a particularly instructive case because the public evidence reaches the middle of that chain and then stops. The Bangladesh Telecommunication Regulatory Commission's ISP list names Nayet Internet Service at Jotma Complex, Sayedpur, Fakir Bari, Narayanganj, and classifies it as an ISP in the old Cat-C system. The APNIC registration for AS133148 names the same organisation, gives a Narayanganj address and records the autonomous system as active in the registry. A separate APNIC record for 103.66.228.0/23 records a portable block of 512 IPv4 addresses under the NAYET-BD name.
Those are meaningful facts. They establish an organisation, a regulatory trace and control of internet number resources. They do not establish that a subscriber can order a connection today. Registration is not the same as operation; an address block can remain allocated while its routes are withdrawn; a historical licence category does not prove that a required conversion was completed; and a route can remain visible while the local customer network behind it is small, idle or carried on other addresses.
The clearest current signal is negative. RIPE's AS overview for AS133148 marked the ASN unannounced on 10 July 2026. Its routing-status record reported no announced IPv4 or IPv6 space, visibility from none of the 327 IPv4 full-feed peers counted at the query time, and no observed neighbours. A current BGP-state query for 103.66.229.0/24 returned no routes. That does not prove Nayet stopped serving every customer. It does mean the strongest public sign of an independently routed Nayet network had disappeared.
The right conclusion is therefore neither that Nayet never operated nor that it currently functions as a regional ISP of known scale. The defensible conclusion is narrower: Nayet operated a publicly visible route until early 2026, but its present routing and retail operating status cannot be confirmed from public evidence. Any assessment of its physical network has to begin from that uncertainty rather than smooth it away.
What the records establish, and what they do not
The BTRC list and APNIC records align on name and location closely enough to identify a genuine Narayanganj operator. APNIC registered AS133148 to Nayet on 11 January 2023. Its /23 address block was registered to the same organisation on the same date, with the registration later changed in July 2023. The APNIC contact record shows that an abuse mailbox was validated in March 2026. Contact validation is evidence that someone could respond to an APNIC confirmation request; it is not a test of network reachability, customer service or licence standing.
The geography is local and commercially dense. The official Narayanganj district profile describes a district of about 683 square kilometres with five upazilas and seven thanas, including Narayanganj Sadar and Fatullah. The preliminary 2022 population census counted 3,909,138 residents in the district. That density can improve the economics of fixed access because more customers may be reached per kilometre of feeder and distribution line. It can also make repair work harder: crowded roads, mixed residential and industrial demand, shared poles, construction, river crossings and congested rights of way concentrate physical risk.
No public map ties Nayet to all of Narayanganj, however. The old Cat-C address places the business in the district, not across every municipality, thana or industrial zone. Bangladesh's 2020 licensing framework replaced the old A, B and C categories with nationwide, divisional, district and upazila/thana licences. The licensing guideline says legacy category licensees were to convert to the upazila/thana structure. No accessible current record reviewed for this article showed which converted service area, if any, belongs to Nayet. The legacy listing is therefore evidence of regulated ISP identity, not a present coverage map.
The routing record is more specific. RIPE's routing history for AS133148 shows one Nayet-originated prefix, 103.66.229.0/24. RIPE's more concise status history reports that the route was first seen on 20 January 2023 and last broadly seen on 6 January 2026. The coarser history records a thin residual observation into the end of January, when only one full-feed peer was seeing it. That is a history of actual global routing visibility, not merely a database registration.
The other half of Nayet's /23, 103.66.228.0/24, appeared in public routing records under another origin ASN during parts of 2023 to 2026. This is an important ambiguity, but it is not a basis for declaring a corporate affiliation, wholesale agreement or transfer. Address-holder registration, route origination and physical service delivery are different facts. A resource holder may authorise another network to originate a prefix, may assign addresses to a customer, or may leave stale records. Settling the operator boundary would require a current letter of authorisation, route-policy record, contract disclosure or confirmation from both parties.
There are further limits. The PeeringDB query for AS133148 returns no network profile. PeeringDB is voluntary, so absence says nothing conclusive about connectivity. IPinfo's AS133148 page labels the network inactive and lists no current originated addresses, which corroborates the routing withdrawal but depends on IPinfo's own observation method. Cloudflare Radar's AS page still recognises the registered name, but a profile shell is not evidence of current traffic volume. None of those sources publishes Nayet's audited customers, revenue, staff, equipment inventory, service-level record or current retail tariff.
An APRICOT 2025 registration list includes a entity who gave Nayet Internet Service as an affiliation. That suggests some organisational presence in the regional technical community in 2025. The same list also contains a similarly named entity affiliated with City Internet.com, which raises questions about operational overlap but proves no ownership or contractual relationship. A conference registration can identify a person and stated affiliation. It cannot establish that a network is in service, where its fibre runs or who controls it.
The evidence therefore supports a real but bounded description: Nayet was a small registered ISP identity in Narayanganj with its own ASN and portable address space; one /24 was globally routed for nearly three years; its own ASN is not globally visible now. Almost every claim beyond that needs fresh operating evidence.
One visible upstream is not a resilience plan
When Nayet's route was visible, it exposed a crucial feature of the service chain. A RIPE neighbour snapshot for 15 December 2025 found one immediate neighbour to the left of AS133148: AS58715. A historical BGP-state snapshot from the same day repeatedly shows paths ending 58715 133148 as viewed from collectors around the world. APNIC identifies AS58715 as Earth Telecommunication, while its PeeringDB profile describes an international-internet-gateway and ISP network with extensive external connectivity.
This is strong evidence that AS58715 was the one externally visible immediate path for Nayet's route on that date. It is not enough to say that Earth was Nayet's only commercial supplier, that the arrangement remained in force after the snapshot, or that every Nayet customer packet followed the same physical circuit. BGP collectors see advertised logical paths. They generally do not reveal whether two circuits share the same duct, bridge, metro aggregation switch, data centre, power feed or field maintenance team. They also cannot see private backup links that were not carrying or advertising the prefix at the measurement time.
Still, a one-neighbour view matters. If Nayet had two independent upstream sessions both propagating the prefix, an observer would usually expect more than one immediate external ASN to appear over a suitable interval, unless routing policy deliberately kept one path hidden. The snapshot therefore provides no affirmative evidence of upstream diversity. That is different from proving there was none.
The distinction is operationally decisive. Two BGP sessions delivered over the same metro fibre are logically diverse but physically common. Two fibres entering a point of presence through the same roadside trench can both fail under one excavation. Two upstreams housed in the same building can both disappear when the building loses power or an aggregation switch fails. Conversely, a local operator might use one public route for business services while residential customers leave through carrier-grade NAT on an upstream's addresses; the public Nayet prefix would then reveal only part of the service.
The current withdrawal raises several possible explanations. Nayet may have ceased independent routing while continuing as a retail reseller. It may be renumbering customers into an upstream's address space. It may have lost a session, licence, commercial agreement or physical backhaul. It may be temporarily dormant. It may have stopped operating. Public BGP evidence cannot choose among those explanations. A direct operational confirmation, fresh route advertisement, current customer installation, current BTRC licence entry or documented migration would be needed.
The route's security posture is also unresolved. A current RPKI validity query finds no validating route-origin authorisation covering AS133148 and 103.66.229.0/24. Because the prefix is withdrawn, this does not make an active route invalid; there is no active route to validate. It does mean that a future reannouncement would not currently gain the cryptographic origin protection a valid ROA could provide. No public Internet Routing Registry route object appeared in RIPE's current WHOIS aggregation either. For a small operator returning to service, publishing correct route authorisation and policy records would make the intended origin easier for upstreams and observers to verify.
The ownership boundary behind a local connection
Bangladesh's licensing design makes the boundary between the retail ISP and the wider network explicit. Under section 22 of the BTRC ISP guideline, ISP licensees connect to licensed International Internet Gateways for international bandwidth and to National Internet Exchanges for domestic inter-operator traffic. The guideline also places transmission within the country's infrastructure-sharing framework. In practice, a local ISP can control the subscriber relationship and parts of the access network without controlling the long-haul fibre, international capacity or exchange fabric on which the service depends.
For Nayet, none of the physical ownership layers is documented publicly. The APNIC /23 proves address-resource registration, not fibre ownership. The historical AS path proves route propagation through one visible neighbour, not ownership of the backhaul carrying the packets. The BTRC address proves a regulatory and business location, not a powered network operations centre. A street cable bearing no label cannot be attributed to Nayet without a permit, asset record or operator confirmation.
A plausible fixed-broadband chain in Narayanganj would begin with customer-premises equipment: an optical network terminal, Ethernet media converter, wireless receiver or router. A drop cable or radio link would reach a distribution point. If the network uses passive optical networking, several customer fibres would share a splitter and feed an optical line terminal at a powered point of presence. If it uses active Ethernet, powered switches could sit deeper in the field. If it uses fixed wireless, tower or rooftop radios, line of sight, licensed or unlicensed spectrum, mounting rights and power at both ends would become material.
The public record does not tell us which of these architectures Nayet used.
That uncertainty should not be filled with the newest-looking design. A /24 can serve a fibre network, a wireless network, enterprise circuits, NAT gateways, servers or a mixture. An ISP licence does not specify whether the last metre is glass, copper or radio. Even the word broadband does not establish fibre to the home. The evidence required is mundane but decisive: dated photographs with identifiable locations, installation forms naming the access technology, optical line terminal or base-station inventories, pole or rooftop agreements, network diagrams, customer-premises device types, or measurements from known active subscribers.
After the access layer comes backhaul. A local point of presence needs a path to an IIG and, for domestic traffic, a NIX connection either directly or through an arrangement permitted by the regulatory structure. The physical path may be leased from a nationwide transmission operator, provided by an upstream, or built within the limited scope the ISP is authorised to control. Every hand-off creates a fault and accountability boundary. The local ISP may see a dark optical interface but have to open a ticket with a carrier. The carrier may then discover damage on infrastructure maintained by a third party.
To the subscriber, all of that still looks like one broken internet connection.
Power cuts across every layer. Passive splitters need no electricity, but optical line terminals, switches, radios, routers, cooling systems and customer devices do. A generator at the core does not help an unpowered active cabinet. A battery at a cabinet does not help if the upstream building is dark. A powered access network does not help if the customer's router has no battery. The relevant resilience measure is therefore the minimum surviving power time across the whole service path, not the largest battery at one site.
The ITU's optical-access topology guidance explains why route separation matters: fibres in different cables but on one route do not protect against a route break, while different routes and different central offices can protect against a wider set of failures. It also recommends sizing battery backup with commercial-power outage rates and repair time in mind. These are useful standards against which to ask questions. They are not evidence that Nayet implemented any particular protection scheme.
Address space is not installed capacity
The number 512 looks precise because a /23 contains 512 IPv4 addresses. It would be a mistake to translate that into 512 customers, 512 simultaneous sessions or any particular amount of bandwidth. One address can sit unused. One can front thousands of customers behind carrier-grade NAT. A household can receive a private IPv4 address and never appear in Nayet's public block. Enterprises may use multiple addresses. Network equipment, servers and point-to-point links consume addresses too. IPv6 could expand the address layer enormously, yet the current public evidence shows no Nayet-originated IPv6 route.
Nor does a route say how much capacity exists behind it. BGP announces reachability, not port speed. The same /24 could be carried over a 100 Mbps circuit, a 10 Gbps hand-off or a congested shared path. A provider can install a high-capacity optical line terminal but buy too little peak-hour transit. It can lease ample upstream capacity while using oversubscribed access switches. It can advertise a retail speed that is reached at quiet times but not when many households stream video in the evening. Installed capacity becomes usable capacity only when every constrained segment can carry the offered load.
Bangladesh's published ISP tariff framework is revealing here. It standardised retail reference packages and discusses a contention ratio of 1:8. Contention is not automatically poor engineering; sharing makes mass-market broadband affordable because not every customer uses the maximum rate at once. But a nominal 1:8 ratio is only one planning assumption. Real quality depends on traffic distribution, upstream headroom, caching, exchange paths, congestion management and whether the provider adds capacity before peak demand overwhelms it.
The national market is large enough for local utilisation to matter. BTRC's internet-subscriber statistics explain that fixed ISP and PSTN figures are updated quarterly because the country has many ISP operators and relatively low monthly churn. BTRC's teledensity series put fixed-broadband penetration at 8.48 per cent in April 2026. These national figures show a meaningful fixed-access market; they do not allocate any subscriber to Nayet.
For a small ISP, the economic problem is to keep the utilisation of expensive shared assets high without letting peak-hour quality collapse. Dense streets can shorten drops and spread the cost of a point of presence over more bills. Yet every extra branch, splitter, switch and customer premises adds a possible trouble ticket. The operator pays wholesale bandwidth, transmission, rent, electricity, licence fees, equipment replacement and labour before retaining any margin. If tariffs are constrained and customers can switch among neighbourhood providers, resilience investments compete directly with short-term affordability.
This is why installed-versus-usable capacity should be audited with time series rather than equipment labels. Useful evidence would include peak and 95th-percentile utilisation by upstream and access port; packet loss and latency at busy hours; optical power levels; port error counts; backup-power runtime under load; oversubscription by aggregation segment; customer count by technology; and capacity-addition dates. No such figures are public for Nayet. Any claim that its /23 represents a certain network scale would be numerical theatre.
What an independent route changes for a small ISP
Running an autonomous system is not a requirement for every neighbourhood access provider. A small retailer can buy an upstream service, use private addresses or addresses supplied by the carrier, and let the carrier originate the public route. That arrangement reduces the expertise and coordination required to operate BGP. It can also make migration harder because the retail provider's customers may need to be renumbered when the wholesale carrier changes.
Portable address space and an independent ASN create the possibility of moving between upstreams while preserving addresses and announcing the same prefix, but only if contracts, router configuration and route authorisations are ready.
Nayet's 2023 move into independent routing therefore represented an operational capability, not simply another registration number. The visible /24 gave the organisation a distinct place in the global routing table. External networks could see AS133148 as the origin, and Nayet could in principle establish policy separately from the upstream carrying the announcement. The history does not reveal why Nayet made that investment or which customers used the prefix. It does show that the organisation progressed beyond an invisible reseller identity for a sustained period.
The value of that independence depends on execution. An ASN with one visible upstream does not by itself create competitive leverage during an outage. Portable addresses do not move automatically when the only session fails. A backup provider must accept the prefix, filters must be correct, a valid path must propagate, and the physical circuit must survive the failure that removed the primary. Engineers need to maintain the configuration even when it is idle. For a small provider, the second circuit and the labour needed to test it can be expensive relative to the revenue protected.
Withdrawal changes the balance again. If Nayet migrated customers onto addresses originated by a wholesale provider, it may have reduced routing overhead while surrendering some portability and public visibility. If it retained customers on the Nayet /24 without another announcement, those addresses would not be reachable from the global internet. If there are no current customers, the carrying cost of an ASN and portable block may still be worth paying to preserve the option of a return. All three situations can produce the same public observation: a registered ASN with no route.
This is why the withdrawal is economically significant without being a verdict on the company. It removes evidence that Nayet currently controls its own external routing. It does not reveal whether revenue continued through another network arrangement. The distinction would matter in a wholesale negotiation, a customer procurement decision and a resilience review. A buyer that requires provider independence would want the live origin and tested alternate path. A household may care only that the connection works and the support number answers. Both are legitimate standards, but they measure different services.
The absence of an independent route also narrows what outside measurements can attribute. When traffic exits through an upstream's addresses, latency and outage data are usually associated with the upstream ASN. Nayet could remain active yet disappear from ASN-level quality dashboards. That possibility is one reason the current routing silence cannot prove closure. It is also why current subscriber tests, invoices, installation records and operator confirmation are necessary to replace the lost routing signal.
Six ways the service chain can fail
1. The access line is cut
The most local failure can be the most labour-intensive. A drop fibre can be bent, crushed or pulled from a pole. A feeder can be cut by road work or building construction. A connector can be contaminated. A shared splitter can fail or be damaged. In an active-Ethernet network, a field switch can lose power or overheat. In fixed wireless, alignment can change, a radio can fail, foliage or new construction can obstruct a path, and severe weather can degrade the link.
The number of affected users depends on where the damage sits. A broken customer drop may disconnect one household. A failed distribution cable can remove a block. A severed feeder can darken every subscriber downstream of an aggregation point. A ring can route around one break only if the ring is complete, both directions are active, the paths are physically separate and the switching logic works. No map or test result shows whether Nayet had a ring, a tree or a chain of single-homed branches.
2. Local or customer power disappears
Fixed broadband is often described as passive because fibre itself needs no electrical amplification over neighbourhood distances. The service is not passive end to end. The access head-end, routers and switches require power, as do many intermediate active devices and the customer's terminal. A small uninterruptible power supply may bridge a short outage but fail during a long one, especially if its battery is old, hot or supporting more load than planned.
Bangladesh businesses operate in a power environment where backup remains a practical concern. The World Bank's Bangladesh enterprise indicators track the share of firms experiencing electrical outages, while daily power-system reporting can show load shedding and unavailable generation. Those sources do not prove an outage at a Nayet site. They establish why a local ISP's generator fuel, battery maintenance and runtime tests are material operating facts rather than decorative specifications.
3. The upstream path is lost
This is the failure most directly connected to Nayet's public record. In December 2025, all globally observed paths in the examined snapshot reached AS133148 through AS58715. If that logical adjacency was the only active external route and it failed, Nayet's public /24 would become unreachable unless a backup began advertising it. The disappearance of the route in January 2026 is consistent with such a loss, but it does not identify the cause.
A second contract is not enough. Resilient upstream design requires separate BGP sessions, correct route filters, authority to advertise the prefix, sufficient capacity on the backup and physical separation far enough toward the customer to survive a shared cut. The failover has to be tested. Otherwise the second line may be a paper reserve that fails on first use because its optics, addressing, route authorisation or cross-connect is stale.
4. Domestic or international paths are impaired
An ISP can retain its local route while access to important destinations fails elsewhere. Domestic traffic may depend on a NIX path; international traffic moves through licensed gateway and cable systems. Congestion or failure at either layer changes customer experience differently. Local caches and domestic services may remain fast while international sites stall, or a domestic exchange problem may make a nearby service take an expensive external detour.
Bangladesh's July 2024 shutdown demonstrated the social consequence of a higher-layer interruption. Cloudflare's traffic analysis observed the national disruption and the staged restoration of broadband connectivity beginning on 23 July. That event was not a Nayet-specific equipment failure, and it should not be used to score Nayet's reliability. It does show that a working neighbourhood access plant cannot guarantee end-to-end service when national gateways or policy controls remove the wider path.
5. Capacity is present but congested
Congestion is a partial failure. The link remains up, support dashboards may remain green, and customers still cannot complete time-sensitive work. Queueing delay rises before a circuit reaches absolute saturation; packet loss triggers retransmissions; video quality falls; interactive applications become erratic. A provider that measures only whether an interface is up can miss the failure its customers actually experience.
The bottleneck may sit in the PON tree, an aggregation switch, an upstream port, a NIX session, a carrier hand-off or the customer's Wi-Fi. Separating these requires measurements at multiple points and at peak time. Cloudflare maintains an internet-quality page for AS133148, but no populated public time series available in the examined view can substitute for operator telemetry or a panel of known subscriber tests. A page label is not a quality result.
6. The repair system runs out of people or parts
A network is restored by a supply chain and a roster. Field technicians need transport, optical power meters, fusion splicers, ladders, safety equipment, connectors, drop cable, splitters, spare optics, switches and compatible customer units. Someone needs authority to access rooftops, poles, cabinets and leased facilities. Someone must diagnose whether the fault belongs to the local operator or a carrier, escalate it with usable evidence and keep the customer informed.
Local support labour is therefore part of physical capacity. A technically redundant path can remain unavailable if no one is able to configure it. A stocked spare is useless if it is the wrong optical class. A skilled splicer cannot repair a buried carrier cable without access permission. Small ISPs gain responsiveness from proximity, but a small team can also be exhausted by simultaneous cuts, storms or power failures. Nayet publishes no staff count, on-call coverage, spare policy or mean time to restore, so its labour resilience cannot be rated.
Who bears the cost of failure
The affected population cannot be counted without a current customer list, but the location indicates the kinds of dependency at stake. Narayanganj is both densely residential and industrial. If Nayet still serves the area, interruptions could affect households, small retailers, workshops, educational users, remote workers and local offices. Larger factories often buy more engineered connectivity, yet small suppliers and staff may still rely on mass-market access. The effect of a failure depends less on a label such as residential or business than on whether the user has an independent alternative.
For a household, the backup may be mobile data, provided mobile networks and the handset have power and capacity. For a shop, a mobile connection might keep messaging and payments alive but not support cloud backups or multiple terminals. For a remote worker, high latency and packet loss can make video calls unusable even when web pages load. For a clinic or emergency service, a consumer connection should not be treated as a resilient circuit without explicit design and testing.
The tariff framework recognises prolonged outage as an economic event. BTRC's published tariff notice says a customer who remains continuously disconnected for five days pays half the monthly bill, for ten days pays one quarter, and for fifteen days pays no monthly bill for that month. The rule protects customers at the extreme end of failure. It does not compensate for lost work, failed sales, travel to find connectivity or repeated shorter interruptions that never cross the continuous-outage threshold.
For the ISP, an outage cuts both ways. Repair costs arrive when revenue is most at risk. Emergency fibre, overtime and transport cost more than planned maintenance. Customers may churn, yet the operator may still owe upstream, rent and licence payments. A prolonged route withdrawal can also make the network harder to observe and trust, reducing the confidence of enterprise buyers and potential peers. That is why operational transparency has economic value even when disclosure is not mandatory.
The route withdrawal makes the present customer question especially sensitive. It would be wrong to infer that every past Nayet subscriber lost service in January. Customers could have been migrated to another address range or upstream-originated network. It would be equally wrong to infer continuity because the APNIC registration remains active. The people who would know are current subscribers, the operator, its current licensed carriers and the regulator. Until one of those sources provides dated evidence, the number affected by any current failure remains unknown.
The recovery evidence that would change the assessment
The most useful next disclosure is not a marketing coverage map. It is a compact, testable description of the live service chain.
First, Nayet would need to establish current legal and operating scope: its converted BTRC licence number, authorised upazila or thana, current retail contact, active packages and a dated statement that new installations are available. A licence proves authorisation, not performance, but it resolves whether the legacy Cat-C listing became a current operating right.
Second, it would need to establish current routing: a visible announcement from AS133148 or a documented explanation that customers use another origin; the prefix and origin; current upstream sessions; route authorisations; route-policy records; and measurements showing propagation. If the service now operates behind an upstream ASN, that should be stated plainly because the public Nayet ASN is no longer a valid proxy for customer reachability.
Third, it would need to establish upstream diversity. The minimum credible evidence would identify two active providers or two independently operated paths, the points of hand-off, contracted capacity, whether both can carry the full critical load, and a dated failover test. The physical route should be traced far enough to show whether the circuits share a pole line, duct, bridge, building entrance, rack, power feed or metro node. Commercial names can be redacted while common-mode risks are still disclosed.
Fourth, it would need to establish access topology and installed-versus-usable capacity. A map can be generalised to protect security while showing points of presence, ring segments, tree branches and high-consequence single points. Equipment counts should be paired with active ports and peak utilisation. A statement such as 10 Gbps installed means little unless the narrowest upstream, aggregation and access segments are also stated.
Fifth, it would need to establish power endurance. Each critical point should have a measured load, battery age, tested runtime, generator or alternate supply where appropriate, fuel and refuelling plan, and an alarm path. The result should be an end-to-end survival time. If the core lasts eight hours but a field switch lasts twenty minutes, the service path lasts twenty minutes.
Sixth, it would need to establish repair capacity. That means on-call coverage, escalation contacts, spare quantities for common failure items, access agreements, fibre-test and splicing capability, and recent restoration times by fault class. Median restoration alone can hide a long tail; reporting the slowest material incidents explains whether difficult cuts are actually manageable.
Finally, evidence should come from operation rather than promise. A route seen by many collectors is better than a planned BGP session. A failover log is better than a claim of redundancy. A battery discharge test is better than a label on a cabinet. A time-stamped customer installation is better than a social-media page that has not been updated. A regulator entry is better than an old category list. Each item closes a different uncertainty; no single document proves the whole network.
A careful status classification
Nayet's public trail is stronger than that of a name with only a business-directory entry. The organisation held portable address space, obtained its own ASN, announced a /24 globally and remained visible through hundreds of routing peers for most of 2023, 2024 and 2025. Its abuse contact was validated in 2026, and a stated representative appeared in a 2025 regional conference registration. Those facts justify treating Nayet as a historically operating network identity rather than a purely speculative company name.
But the original regional-ISP hypothesis does not survive as a present-tense operating claim. The ASN is unannounced. The last widely observed route ended in January 2026. There is no current visible upstream, no PeeringDB profile, no published service map, no verified package list, no customer measurement panel and no current licence-conversion evidence in the sources examined. The historical route showed one immediate neighbour in the sampled period, while no evidence demonstrates a second logical or physical path.
This is a downgrade to historically routed, present operation unverified. It is not a finding that the business has closed. It is not a claim that no customer receives service. It is not a declaration that AS58715 caused the withdrawal. It is an evidence boundary: the observable network that once made Nayet independently reachable is absent as of 10 July 2026, and the remaining records cannot reveal what, if anything, replaced it.
The distinction matters to anyone assessing a local broadband bill. A price and speed are only the visible edge of a chain of assets, contracts and labour. Nayet's own history shows how much can be learned from one route: the dates it was reachable, the address space used, the one visible neighbour near the end and the point at which global visibility disappeared. It also shows what a route cannot answer: who owns the poles and fibre, how many customers remain, whether a backup exists, how long batteries last, and how quickly a crew can reach a broken feeder.
Until current operating evidence appears, those unknowns should stay unknown. The next credible update would be simple and concrete: a current licence area, a live installation, a visible or documented routing arrangement, a tested second path, a physical topology description, measured power endurance and recent repair performance. That evidence would turn a registration history into an operating profile. Without it, Nayet remains a useful case study in the gap between owning network identifiers and proving a resilient local connectivity service.