Summary

  • ORELTELECOM is a young Ecuadorian legal entity operating through the established AJnet brand: ARCOTEL granted its internet-access title in December 2024, LACNIC registered AS274252 to ORELTELECOM in November 2025, and the current AJnet site markets fibre service in five Cotopaxi cantons. The public record does not show which legacy AJnet assets, customer contracts, poles, towers or field teams were transferred into the company.
  • Current routing evidence is materially stronger than the old brand footprint but still asymmetric. AS274252 originates one IPv4 /24 and one IPv6 /32 with valid route-origin authorisations. IPv4 is visible only through Ufinet as the direct upstream; IPv6 also appears through AS263238. That second logical adjacency is useful, yet it does not establish separate ducts, towers, road corridors, facilities, power feeds or failure domains.
  • A detailed 2024 Salcedo engineering project describes a 280 Mbps radio link, a proposed 1.5 Gbps GPON node and two outward-running fibre trunks, but it is a design document rather than an as-built acceptance record. ORELTELECOM can substantiate resilience only by publishing current route maps, carrier demarcations, shared-risk analysis, backup-power endurance, spare-cable locations and measured restoration performance for the communities it now markets.

The road fails before the network diagram does

Imagine the first hour after a wet Andean slope gives way above the road from Latacunga toward western Cotopaxi. Mud, rock and trees cover both lanes. A black access cable that followed the road edge is exposed, stretched and then cut by moving debris. The same closure that breaks the line prevents a repair vehicle from reaching the break. Customers farther along the corridor still have powered optical terminals and working Wi-Fi, but their traffic has nowhere to go. This is a resilience test, not a report of a specific AJnet outage. It is nevertheless grounded in a current local hazard: on 9 March 2026, a landslide in Pilaló, in Pujilí canton, caused a total road closure, according to ECU 911's incident account.

The distinction matters because a regional ISP can show two lines on a diagram while both occupy one practical failure corridor. Ecuador's transport ministry described the E30 Latacunga-La Maná road as the link among Latacunga, Pujilí, La Maná and intervening parishes when it declared emergency works after rain, river action and landslides damaged sections between kilometres 56 and 120. The ministry's emergency document said the damage isolated populations. A later national road-condition report still recorded a lower-slope landslide affecting the E30 around kilometre 104 in the Tingo-La Esperanza area. Those are not ORELTELECOM route disclosures. They are evidence that the terrain can remove carriageway, access and any adjacent utility corridor together.

Pangua raises the recovery problem further. Its 2024 territorial plan records 512.32 kilometres of roads and classifies 73.46% of the network as being in poor condition; it also identifies the Quevedo-El Corazón-Latacunga connection as the main road axis. The Pangua plan describes a service territory in which the final kilometres of a repair journey may be slower and less redundant than the fibre itself. An official Cotopaxi provincial-map portal supplies a useful geographic frame, but it does not publish an operator's cable alignment. No public ORELTELECOM record identifies whether its access plant follows E30, local roads, electricity poles, radio sites or some combination.

The opening question is therefore physical, not rhetorical. If the cable beside the blocked road is severed, does traffic turn onto a route that reaches a different upstream through a different valley, pole line and powered site? Can a crew enter from the opposite side? Is spare ADSS cable stored on both sides of the likely isolation point? Are land-access permissions already in place, or does restoration wait for a road authority, pole owner or landholder? Public routing data can reveal who announces a prefix. It cannot answer those questions.

Until ORELTELECOM shows the shared-risk boundaries, the honest conclusion is that route survival after a Cotopaxi landslide remains unknown.

ORELTELECOM is the new legal shell around an older AJnet operation

ORELTELECOM S.A.S. is not simply another spelling of AJnet. It is the legal company to which ARCOTEL granted a 15-year title for internet-access service and the use of non-essential spectrum on 9 December 2024. The granting resolution identifies Ecuadorian tax number 0591763938001 and Jorge Mauricio Aguilera Jiménez as general manager and legal representative. It also states that the original application did not include wireless networks, sets a one-year installation-and-operation period from registration, requires a disaster contingency plan, and obliges the operator to report subscriber, traffic and international-capacity information to the regulator. Those reporting duties do not make the underlying figures public.

Two commercial company profiles broadly corroborate the legal timing while adding little infrastructure certainty. EMIS places ORELTELECOM in Latacunga and dates its establishment to November 2023. Datapower gives the same tax number and formation date. These are secondary signals, not substitutes for company filings, audited accounts or an asset register. They support the proposition that ORELTELECOM is a recent corporate vehicle; they do not reveal whether it owns fibre, leases it, uses infrastructure owned by the representative personally, or merely contracts access from third parties.

AJnet, by contrast, has a longer operational history. A 2023 engineering project at the Technical University of Ambato described AJnet as a natural-person business operating since 2011 under a different tax identity, with fibre in Latacunga and a need for better service monitoring. The Latacunga monitoring study dealt with optical-signal loss, low optical power and equipment failures. It is historical evidence about the predecessor-branded operation, not a current performance audit of ORELTELECOM. The legal title, customer contracts and physical assets may now sit together; public documents do not establish that chain.

The current retail face remains AJnet. Its privacy notice calls AJnet the controller of customer data, gives a Latacunga address on Calle Guayaquil, and refers to installation, support and infrastructure providers. It does not name ORELTELECOM in the visible text. LACNIC provides the clearest bridge: the current autonomous-system registration names ORELTELECOM as the registrant while listing AJnet contact details. That supports an operating connection between the company and the brand, but it still does not define ownership at every layer.

For customers and creditors, the boundary is consequential. The retail name receives the service request; the licensed company carries regulatory obligations; a carrier may supply upstream transit; a pole owner controls attachment; a contractor may perform splicing; and a different owner may hold a tower or building lease. When one of those parties fails, responsibility follows contracts that are not public. ORELTELECOM should therefore be understood as the authorised operator behind a continuing AJnet market presence, while the transfer and control of specific legacy assets remain unverified.

Five cantons are marketed, but the public map resolves only three

AJnet's current retail site promotes “100% fibre-optic” internet and names Latacunga, Pujilí, Salcedo, Saquisilí and Pangua as coverage locations. It also advertises symmetric upload and download, Wi-Fi 6 equipment, technical support and digital television. Those claims demonstrate an active sales surface. They do not say whether every household in each canton is serviceable, whether a named place is reached by owned fibre, leased fibre or radio, or where a network changes from one medium to another.

The operator's public geofeed is more precise about addresses but narrower about geography. It places ORELTELECOM's own IPv4 /24 and IPv6 /32 in Latacunga, alongside small IPv4 blocks in Latacunga, Salcedo and Pujilí. It has no row for Saquisilí or Pangua. A geofeed is an address-location declaration, not a cable map. It can tell content networks where an address is intended to be located; it cannot show a feeder route, a distribution cabinet, a wireless hop, a splice enclosure or a repair boundary. The absence of a row does not prove absence of service, and the presence of a row does not prove a fibre path.

The current brand also operates an AJnet-labelled speed-test endpoint. That is another useful sign of a live retail environment, but it measures the path seen by a participating user at a moment in time. It does not disclose the size of the upstream contract, the number of subscribers sharing it or the amount of capacity available after a link or node failure. A speed test can be excellent while the network has one critical corridor; it can also be poor because of a customer's Wi-Fi even when the backbone is healthy.

The marketed footprint is not trivial. Ecuador's 2022 census profile gives canton populations of 217,261 for Latacunga, 67,493 for Salcedo, 66,980 for Pujilí, 24,356 for Saquisilí and 21,867 for Pangua. The total is 397,957 people, calculated from the INEC Cotopaxi profile. That is a population frame, not a subscriber count or addressable-market estimate. Much of Cotopaxi is rural, household density varies sharply, and an ISP's practical service boundary is determined by pole access, optical budget, tower line of sight, installation labour and the economics of each extension.

The map that a resilience analyst needs is therefore absent. It would show the Latacunga headend or heads, each aggregation node, inter-canton trunks, radio relays, road and pole co-location, upstream hand-off sites, power sources and isolation valves in the topology. It would distinguish active routes from planned routes and owned plant from leased strands. It would also mark the point at which service to Pangua or a rural Pujilí community depends on a single road approach. AJnet's five canton names are valid evidence of commercial intent and apparently current availability somewhere in each named market.

They are not evidence that five geographically independent networks exist.

Salcedo's network history begins with a 280 Mbps radio bottleneck

The richest physical account of AJnet's network is a 2024 Salesian Polytechnic University project to design a GPON network for Salcedo. The full engineering study describes AJnet as serving urban and rural Cotopaxi from Latacunga, then extending to Pujilí and Salcedo. It says the urban Latacunga network used fibre while rural areas used radio, and that Salcedo customers were being served through a radio system whose active equipment remained tied to the Latacunga operation.

The study's account of that radio path is unusually specific. A MikroTik CCR1036 edge router connected over fibre to the El Niagara tower south of Latacunga. From there, a 10.2-kilometre point-to-point link operated at 6,115 MHz with an 80 MHz channel and was described as capable of 280 Mbps. The far end was at a Palama tower, where nine antennas and panels redistributed service to subscribers. This is a credible description of the configuration observed or supplied to the author in 2024. It does not establish that the same radios, frequency, throughput or route remain in service in July 2026.

The 280 Mbps figure is also not a promise of customer throughput. It is a stated point-to-point capability. Protocol overhead, interference, modulation changes, weather, radio alignment and aggregate demand all affect usable capacity. If the link was the only feed to 102 or 103 Salcedo customers, as different passages in the study report, simultaneous usage would have made contention a central operating issue.

Even if the radio was retained only as backup after a fibre build, its restoration value would depend on whether the towers, routers and power supplies were maintained, whether spectrum use was regularised, and whether the backup could carry today's much larger retail plans.

The older Latacunga study adds a customer-facing view of failure without providing current metrics. It records recent interruptions reported by surveyed users, some lasting minutes and others hours, and describes a monitoring effort designed around loss of optical signal, low power and equipment faults. Those findings belong to the earlier AJnet operation and should not be projected onto ORELTELECOM's present service. Their value is diagnostic: they identify the fault types a small regional network must detect, locate and repair.

This history makes Salcedo a useful test case. If the 2024 radio architecture has been replaced by fibre, where does that fibre run and where does it meet the core? If the radio remains as a standby path, is it tested under load, and can it operate through the same power failure that disables the primary node? If both fibre and radio terminate in the same Latacunga router room, they may diversify the access medium without diversifying the facility. If the Palama tower and a roadside cable share the same landslide or power exposure, the second medium may not survive the event that matters.

The public record supplies components, but not the current dependency graph.

A GPON design is not an as-built ring

The Salcedo project proposed a substantial change: two 24-core ADSS trunk cables leaving the Salcedo equipment room in opposite directions, supplemented by four 12-core cables toward nearby zones. It divided the city into Tebaida Norte, Tebaida Sur, Centro, Santa Ana and Palama, with 100 planned network-access boxes across those areas. It modelled optical budgets and selected an OLT, splitters, cabinets and customer terminals. The design is detailed enough to show how fibre could replace radio dependence. It is not a completion certificate.

The grammar inside the study matters. The two trunks “will extend”; routes are “proposed”; the network is simulated; equipment values are estimated. A photograph and coordinates indicate that a Salcedo room and rack existed, while the distribution design remained prospective. The document also says an edge device at the room had “redundancy” toward Ambato in the south and Latacunga in the north through two SFP ports. That claim lacks carrier names, circuit identifiers, hand-off addresses, route traces, fibre ownership and physical alignments.

Its printed port rate contains an obvious unit error, which is another reason not to turn the sentence into a verified capacity claim.

Two cables leaving in opposite directions are not automatically a ring. They may be two branches from one node. They may converge on the same pole line after several streets, enter the same upstream building or use strands in the same carrier cable. A ring requires a closed alternate path and switching behaviour that is demonstrated at the service layer. The public study does not say whether the two proposed trunks reconnect, whether passive split architecture allows automatic protection, or whether customer optical terminals can be reached from both directions. Nor does it establish that the design was built, lit, accepted or maintained.

The legal title reinforces the distinction between permission and plant. ARCOTEL's resolution says ORELTELECOM must register infrastructure and obtain non-telecommunications permissions, including rights to use or occupy public rights of way and construction permits. ARCOTEL separately maintains a registry for physical-infrastructure providers, reflecting the fact that poles, towers and other supports can belong to parties other than the internet provider. Neither source identifies the poles or ducts used by AJnet.

An as-built proof package would look different from the university proposal. It would have dated acceptance records, measured optical traces, splice diagrams, pole or duct permissions, georeferenced cable segments, OLT port assignments, protected-route tests and evidence that a cut on one trunk preserves service through the other. It would state whether the old 280 Mbps radio path was retired, retained as emergency access or repurposed. It would identify any span where both routes cross the same bridge, river, unstable slope or utility structure.

Without that record, the design is valuable evidence of engineering intent and local knowledge, but it cannot carry the stronger claim that ORELTELECOM operates a resilient Salcedo fibre ring.

AS274252 improves routing identity, not physical independence

ORELTELECOM's clearest infrastructure advance is its own autonomous-system identity. The LACNIC registration for AS274252 is active, was registered on 13 November 2025, names ORELTELECOM S.A.S. as registrant and uses AJnet contact details in Latacunga. A separate LACNIC record for 2803:4fd0::/32 assigns the active IPv6 block to the same registrant. These records give the operator control over an origin identity and a large IPv6 address space. They do not identify a fibre route.

RIPE NCC's public observations show AS274252 announcing two prefixes in the period ending 16 July 2026: IPv4 38.97.235.0/24 and IPv6 2803:4fd0::/32. The announced-prefix data therefore matches the operator's geofeed for its own space. Separate IPv4 routing-status data and IPv6 routing-status data show AS274252 as the observed origin. The IPv6 prefix was first seen from that origin in December 2025, soon after registration.

The route-security posture is sound at this narrow layer. RIPE NCC reports a valid route-origin authorisation for the IPv4 /24 and another valid authorisation for the IPv6 /32, each limited to the announced prefix length. That makes it easier for validating networks to reject an unauthorised origin announcement. It does not protect against a backhoe, landslide, failed router, unpaid carrier bill or inaccessible splice.

An ASN still has real strategic value for a regional ISP. It separates the customer network's routing identity from a carrier's address space, supports consistent policy across upstreams, enables route-origin security and can reduce the renumbering cost of changing transit. It creates the possibility of multihoming. It also makes external observation more intelligible: analysts can see who is directly adjacent to ORELTELECOM's origin rather than finding every customer route buried inside a provider aggregate.

But logical control must not be mistaken for physical control. AS274252 can announce the same prefix to two carriers whose circuits occupy the same duct. Two sessions can run on routers in the same rack, through the same optical distribution frame and on the same electricity supply. A carrier can present two commercial services that merge at the same upstream node. Conversely, one upstream ASN can sometimes deliver physically diverse circuits. BGP identifies policy adjacency, not civil works.

ORELTELECOM's ASN is therefore evidence of a more mature network boundary; whether that boundary is reached through independent facilities remains an unanswered engineering question.

IPv4 has one visible door; IPv6 has a second logical neighbour

The observed upstream picture differs by address family. RIPE NCC's neighbour data for AS274252 lists AS52468, Ufinet, for both IPv4 and IPv6. It also lists AS263238 for IPv6 only. In the contemporaneous IPv4 BGP state, all 331 collector paths end with AS52468 immediately before AS274252. The IPv6 BGP state contains 337 paths: 306 reach ORELTELECOM through AS52468, while 31 reach it through AS263238. These are observations from route collectors, not the operator's complete configuration.

For IPv4, the evidence supports a precise statement: only one direct upstream ASN is publicly visible for ORELTELECOM's own /24 at the observation time. It does not prove that there is only one physical circuit. ORELTELECOM could have diverse Ufinet circuits hidden behind the same ASN, a standby that is not announcing, address space supplied by another carrier, or a private failover arrangement. None is visible in the public route state. For IPv6, two direct AS neighbours are visible, but the smaller second path set does not tell us the contracted capacity, preference, stability or physical route of the AS263238 adjacency.

CAIDA's inferred topology provides a deliberately simplified cross-check. Its AS Rank record assigns AS274252 one provider, no peers and no customers, with one IPv4 prefix and 256 IPv4 addresses in its cone. That inference is useful because it agrees with the dominant IPv4 observation, but it compresses relationships and may not reflect the newer IPv6 adjacency. It should not overrule direct route data or be treated as a commercial contract record.

The geofeed reveals a second form of dependence. Four small customer-facing blocks are located in Latacunga, Salcedo and Pujilí but are not separately originated by AS274252. RIPE NCC's network information places the covering routes for 167.250.180.0/29, 200.24.137.96/28, 45.70.198.56/29 and 45.224.150.128/29 under AS52468. This means part of the visible address footprint still sits inside Ufinet-originated aggregates. Those blocks may serve management, transition, customer pools or specific sites; the geofeed does not say.

Address-family asymmetry can turn one physical failure into several customer experiences. A device using ORELTELECOM's IPv6 space might retain a route through AS263238 while an IPv4-only application fails, but only if the second session, local transport and domain-name path all remain available. A customer using one of the Ufinet-originated small blocks could follow a different policy from a customer on 38.97.235.0/24. Network-address translation, resolver choice and application preference can then obscure what has actually recovered.

A successful ping over IPv6 would not prove that the full retail service had survived, while a failed IPv4 test would not prove that every path was down.

This is why a useful failover test has to begin at several customer edges and end beyond both upstreams. ORELTELECOM would need to withdraw or disable the primary hand-off, record convergence for its own IPv4 and IPv6 prefixes, test the carrier-supplied blocks separately, and measure whether real applications work at the reduced state. It should repeat the exercise after removing power from the primary router room, because a second BGP session on the same equipment is not a recovery path. Public observations cannot substitute for that controlled failure test.

The own IPv4 /24 also has an external registration trail. IPinfo's prefix page identifies an ARIN reassignment labelled for ORELTELECOM and dated March 2026. The allocation gives the company an originated block, but not provider-independent proof in the legal sense, and not path diversity. The central result is narrower: ORELTELECOM now controls its visible origin policy, IPv6 has a second logical neighbour, and IPv4 still presents one observable upstream door. The physical doors behind those sessions are undisclosed.

Capacity marketing has outrun disclosed backhaul

AJnet's current site presents plan graphics whose names run from 600 to 1,000, alongside a claim of symmetric upload and download. A prospective customer could reasonably read those as megabit-per-second retail tiers, although the visible page does not publish contention, minimum throughput, busy-hour performance, installation limits or the backhaul assigned to each canton. The promise sits at the customer edge. The public evidence about shared transport remains years older and far smaller in scope.

The 2024 Salcedo design calculated first-year access demand for 143 users: 103 existing customers plus 40 projected additions, each modelled at 40 Mbps with a four-to-one sharing factor. The result was 1,430 Mbps. The document then listed 1,500 Mbps as the capacity assigned to the Salcedo GPON node. That leaves only 70 Mbps of headroom inside the study's own arithmetic, before a larger customer base, higher retail tiers or failure traffic is considered. It is a planning calculation, not a measurement of a live circuit, and the customer count conflicts slightly with the study's 102-customer wording elsewhere.

The GPON line rate is another category. The project refers to standard downstream and upstream capabilities and simulates an optical network. A PON port's line rate is shared among optical-network terminals and is not the same as internet transit. The stated 1.5 Gbps node capacity is not the same as installed OLT switching capacity. A 1,000 Mbps retail tier is not evidence that every subscriber can sustain a gigabit simultaneously. And the 280 Mbps radio link was a different, older access bottleneck. These numbers describe at least four layers: customer product, passive access, node allocation and upstream transport.

ARCOTEL's 2025 tariff notice for older subscribers illustrates why labels matter. It distinguishes a 50 Mbps plan with a one-to-one sharing ratio from a 100 Mbps plan with a two-to-one ratio for the regulated benefit. That notice does not define AJnet's products, but it shows that speed and sharing ratio are separate service characteristics in Ecuador's retail context. AJnet's public plan page does not expose an equivalent ratio for its headline offers.

Installed and usable capacity also diverge during failure. If Latacunga and Salcedo each have a 1.5 Gbps feed but both depend on one upstream interface, the aggregate resilience is lower than the sum. If a backup radio can carry 280 Mbps, it may preserve messaging and basic browsing while being unable to sustain gigabit plans. If IPv6 moves to a second neighbour while IPv4 remains on Ufinet, dual-stack customers may experience partial rather than complete recovery. No public record states transit commits, burst ceilings, busy-hour utilisation, OLT occupancy, backhaul by canton, failover capacity or oversubscription.

The correct capacity conclusion is therefore negative in form but useful in practice. ORELTELECOM appears to be selling a modern fibre proposition and operating its own routed prefixes. It has not publicly shown that the transport behind those products scales with the advertised edge rates, nor that reserve capacity remains usable after the primary path fails. A serious disclosure would publish layer-specific numbers: lit access ports, active customers, node uplinks, contracted transit by carrier, peak utilisation, protected capacity and the reduced service level available in contingency mode.

Power, poles and permits define the real failure domain

Every route in Cotopaxi depends on more than glass. An optical line terminal, edge router, Ethernet switch, radio, air conditioner and customer terminal need electricity. A fibre cable needs poles, ducts or rights of way. A radio path needs towers, clear alignment and powered equipment at both ends. ORELTELECOM's public material names none of the backup-power runtimes at its Latacunga, Salcedo, Pujilí, Saquisilí or Pangua sites.

The Salcedo study describes an equipment room with electrical preparation, temperature control and a sealed environment. Its cost table proposes a UPS among the control equipment. It gives no battery capacity, tested runtime, generator, fuel reserve, automatic transfer arrangement or maintenance record. A proposed UPS line item cannot establish that the room remains powered through a long regional outage. Nor does it say whether upstream carrier equipment and intermediate cabinets share the same endurance.

The local electricity system is actively maintained. In May 2026, distribution company ELEPCO reported preventive work at the Salcedo substation, including replacement of protection relays intended to avoid interruptions. The ELEPCO notice is evidence of utility investment, not a guarantee of continuous power to an ISP node. A separate ECU 911 account of severe weather records fallen poles, electrical wiring and road debris in Pangua after heavy rain. That incident does not identify AJnet infrastructure, but it demonstrates a correlated failure mode: the same weather can obstruct roads, damage shared poles and remove commercial power.

Pole attachment and road access can determine recovery even when the cable itself is owned. ARCOTEL's title requires ORELTELECOM to obtain rights-of-way and construction permissions outside the telecom authorisation. If a cable uses another party's poles, an ISP may need coordination before replacing a damaged span. If a landslide destroys both the road shoulder and the poles, civil reconstruction can precede optical repair. If the alternate route is on the same utility's structures, nominal path diversity may still share a maintenance and storm risk.

Facility independence is similarly unproven. The 2024 document names Latacunga and Salcedo nodes and a Palama tower, but it does not publish current power feeds, flood or slope exposure, access controls, fire protection, spares or carrier demarcations. The ASN data shows routing neighbours, not where sessions terminate. Two carriers meeting in one room behind one breaker are commercially distinct and operationally coupled. Two rooms in separate cities may still depend on one long-haul corridor.

For ORELTELECOM, the most useful resilience map would overlay network and non-network dependencies: cable segments, radio towers, substations, pole owners, bridge crossings, road approaches, site batteries and spare-material depots. It would identify the minimum equipment that must remain powered for each canton to retain local and external connectivity. Public evidence currently supports the existence of power and support dependencies, but no quantified endurance. The failure boundary must therefore remain explicit: there is no public basis for claiming that any ORELTELECOM node can ride through a specified outage duration.

Recovery depends on crews that can reach the break

Regional ISP resilience is often discussed as if failover completes the story. In a mountain service area, failover buys time; labour restores the asset. A crew must receive an alarm, distinguish an upstream fault from an access cut, locate the break, obtain permission to enter, travel with the right cable and closures, make the site safe, splice, test and return customers to the normal path. The duration of each step changes when the road is blocked.

AJnet's site promises technical support and says its technicians are certified. Its privacy notice contemplates installation and support providers. Neither page gives a headcount, depot location, shift pattern, service-level target or emergency stock. A commercial profile's small employee estimate is too uncertain to infer field capacity, especially because contractors may perform installations and repairs. The public record cannot tell whether ORELTELECOM has one internal crew, several canton-based teams or a network of outsourced technicians.

The territory makes that unknown material. Pangua's road plan shows why a crew dispatched only from Latacunga may face long and fragile access. A total closure in Pilaló can split an approach. A damaged E30 can isolate communities and prevent heavy equipment from reaching the same area where cable must be replaced. The correct recovery design may require spare cable, closures, optics, power supplies and radios staged on both sides of known choke points, plus agreements with local technicians who do not need to cross the failure.

Competition can provide customer substitution without repairing ORELTELECOM's network. An ARCOTEL document concerning another operator cites a June 2024 certification of 68 authorised internet providers in Cotopaxi; the figure appears in the 2026 extinction resolution. It is a dated regulatory count, not a list of 68 equivalent fibre choices at every address. In dense Latacunga neighbourhoods, a household may switch or tether. In a rural community reached by one pole line or tower, nominal provider count may offer little practical redundancy, particularly if several networks share the same road and power structures.

Recovery performance should be measured in more than average repair time. The useful distribution includes alarm-to-dispatch, travel, access delay, fault-location, splice, optical testing and full traffic restoration. It should separate urban cuts from mountain-corridor events and planned work from storm damage. It should also state how many customers were isolated and whether a degraded backup preserved essential connectivity. No such ORELTELECOM statistics are public.

The absence of those figures is not proof of weak operations. Small providers often solve faults through local knowledge that never becomes a public report. It does mean outsiders cannot value that capability. In ORELTELECOM's case, local support labour is not merely a customer-service feature; it is part of the infrastructure. The company can turn that invisible asset into credible resilience evidence by publishing coverage by crew base, tested callout procedures, strategically placed spares and anonymised restoration results. Until then, recovery across a closed road is a capability to be demonstrated, not assumed.

What ORELTELECOM would have to disclose to prove resilience

ORELTELECOM has enough public evidence to establish a real, current operating story. It holds an ARCOTEL title. AJnet markets fibre service across five Cotopaxi cantons and maintains current customer-facing web services. AS274252 originates active IPv4 and IPv6 space with valid route-origin authorisations. A detailed local engineering study documents legacy radio infrastructure and a plausible GPON expansion design. These are stronger foundations than a directory listing or an unverified marketing map.

The same evidence also sets a hard boundary around what cannot yet be claimed. No public as-built route map shows the inter-canton network. No carrier contract or demarcation record proves physically diverse upstreams. No acceptance document confirms that the proposed Salcedo trunks and 1.5 Gbps node were built and lit. No report gives current transit capacity, busy-hour use or failover headroom. No backup-power test states how long a node, tower or upstream hand-off survives. No crew record establishes restoration time when E30 or a local road is closed.

A compact disclosure could settle most of these issues without revealing security-sensitive detail. First, ORELTELECOM could publish a simplified physical topology in which routes are generalised to corridors but shared bridges, ducts, pole owners and buildings are marked. Each upstream should be labelled by ASN, hand-off city, facility and whether the local loop is physically separate. The IPv6-only adjacency should be explained: production transit, a peer, a backup, a partial route or an experiment.

Second, the company could classify capacity by layer and state. For each node, it should list access technology, installed port capacity, lit uplink capacity, contracted external capacity, peak observed use and capacity remaining after the largest single failure. The 2024 Salcedo figures should be retired, confirmed or replaced. If a 280 Mbps radio remains as emergency transport, the company should say what services it can sustain and when it was last tested under load.

Third, ORELTELECOM could publish recovery boundaries. That means battery and generator endurance at critical sites, the location class of spare cable and electronics, crew coverage by canton, access agreements and restoration percentiles for major fault types. It should disclose whether Pangua and western Pujilí can be approached and restored from both sides of a corridor. A statement that “support is available” is less useful than evidence that a team, stock and permission remain available after the road closes.

Finally, the company should separate legal, operating and asset ownership. A customer should be able to see that ORELTELECOM is the licensed operator behind AJnet, while understanding which infrastructure is owned, leased or supplied by a carrier. That clarity would improve not only resilience assessment but also procurement, credit analysis and regulatory accountability.

The present verdict is deliberately limited. ORELTELECOM has moved AJnet toward a more autonomous routing posture, and its IPv6 view now contains a second logical neighbour. Its IPv4 /24 still has only Ufinet as a publicly visible direct upstream, and the route, facility, power and labour evidence needed to prove landslide-resistant diversity is absent. The network may be more resilient than the public record shows. For the communities along Cotopaxi's vulnerable roads, however, resilience is the path that still works and the crew that can still arrive—not the second line on an undisclosed diagram.