AS274266 has one visible upstream, while FIBERCOM Dominicana's bridge and repair routes remain undisclosed

Summary

• FIBERCOM Dominicana has a small but verifiable public footprint. Dominican trade-name records connect a 2020 installation-and-repair business with a FIBERCOM Dominicana commercial name registered in 2025; LACNIC then registered AS274266 and an IPv6 /32 to FIBERCOM DOMINICANA SRL in November 2025. Those records establish names, dates and internet-number resources, but they do not identify a concession, customer count, service area, cable route, network facility or inventory of owned plant.
• Global routing collectors currently show AS274266 originating 38.74.56.0/22 and 2803:bfd0::/32 through one visible direct neighbour, ACOLME TECH SRL's AS273217. Both address families experienced a large wave of withdrawals on 14 July 2026. The observations prove a narrow public routing relationship and recent volatility; they do not prove a customer outage, its cause, the contracted transit rate, or whether the two companies reach each other over physically diverse paths.
• Santo Domingo Norte is divided from the central city by the Isabela river and has repeatedly documented flood exposure. Road bridges, drains, pole lines and water crossings are therefore practical failure boundaries for a local provider, yet no public FIBERCOM map shows a bridge attachment, buried duct, aerial span, splice point, alternate crossing or repair staging site. Until the company identifies those dependencies and publishes measured capacity and restoration evidence, the strongest conclusion is not that the network lacks resilience, but that its physical resilience cannot be independently assessed.

After the rain, the decisive network record is under the bridge

The rain has stopped, but the Isabela is still high and brown. Beneath a road bridge at the northern edge of Santo Domingo, two technicians in reflective waterproof clothing stand beside a protected conduit transition. One cleans mud from a chamber lid while the other traces a cable sheath toward the abutment. Traffic moves overhead. The immediate questions are tangible: does this cable belong to FIBERCOM Dominicana, is it leased from another operator, and is there another crossing if water, debris, construction or a vehicle strike damages this one? This is a resilience stress test, not a report of a documented FIBERCOM incident.

Its purpose is to expose the information that a route announcement cannot supply.

The geography makes the test legitimate. Santo Domingo Norte occupies the urban territory north of the Isabela river, a boundary described by the Dominican municipal-information portal. A provider whose registered address is in that municipality may serve only a nearby block, a cluster of neighbourhoods, or a wider market. The public evidence does not say. But any network that connects customers across the river, crosses tributaries or follows the large avenues around Villa Mella, Guaricano and Sabana Perdida must negotiate structures, rights of way and drainage conditions that are invisible in internet routing data.

Those conditions are not abstract. On 28 April 2025, the public-works ministry announced maintenance on the Jacobo Majluta bridge over the Isabela and on a nearby bridge near Guaricanos. The MOPC account described cleaning, resurfacing and work around bridge elements. It did not identify telecommunications cables or FIBERCOM. It nevertheless establishes that crossings in the company's only publicly registered locality are active civil assets subject to intervention. A cable attached to a bridge may depend on the bridge owner for access, traffic management and permission to work; a cable in a roadside duct may depend on drainage and excavation controls; an aerial approach may depend on poles owned by an electricity or communications company.

Heavy rain can join those dependencies. The Dominican presidency reported that strong flooding affected families in Santo Domingo Norte on 9 April 2026, with emergency assistance delivered after water entered vulnerable areas. The official flood report is evidence of recent local exposure, not of network damage. An older municipal account of Hurricane Irma recorded 3,362 displaced people across Sabana Perdida, Villa Mella, Guaricanos, Higüero and La Victoria. The municipal report shows that flood risk is distributed across several communities rather than confined to a single bank.

Water does not need to submerge optical fibre to interrupt service. It can fill a handhole, move a poorly secured duct, undermine a pole, block the road used by a repair truck, cut power to an active cabinet, or prevent a contractor from reaching a carrier demarcation. Debris clearance can sever unmarked plant. A bridge inspection can restrict access exactly when a crew needs it. If the main and backup paths share the same abutment, chamber, pole run or upstream hand-off, two logical routes can fail as one.

The scene under the bridge therefore defines the central uncertainty. FIBERCOM's public internet records show that the company can originate addresses. They do not show the first metre of its local network, the last metre before its upstream, or the civil structures between them. A credible resilience account would name the relevant crossings without exposing customer security: route A uses one protected corridor, route B uses another; each has separate power and carrier demarcation; inspection and repair authority are assigned; spare cable and closures are staged on both sides.

In the absence of that account, the bridge is not proof of weakness. It is the place where a claim of independence would have to become physical.

A field-installation trade name became a routed company

FIBERCOM Dominicana's documentary history begins before its autonomous system. An October 2020 publication by the Dominican National Office of Industrial Property records the commercial name JOSANTFER FIBERCOM under registration 595872. The activity is described as telecommunications installation services and the repair and installation of wired connections, with Jonatan Santos Fernandez named as applicant. The ONAPI 2020 bulletin is unusually relevant because it anchors the enterprise in field work rather than in an aspirational technology label.

Five years later, ONAPI published a new commercial-name entry for FIBERCOM DOMINICANA. Registration 880149, applied for by Jonatan Santos Fernandez and Carolin Daniabel Jerez Gonzalez, describes the activity as the sale of telecommunications services. The August 2025 ONAPI bulletin establishes continuity in a name and applicant while marking a shift from installation work toward selling service. A commercial-name registration is not a corporate asset schedule, a telecom authorisation or proof that the 2020 operation transferred every tool, cable and contract to the current limited-liability company.

The clearest current identity record comes from LACNIC. Its AS274266 registration names FIBERCOM DOMINICANA SRL as the registrant, gives a Santo Domingo Norte address on Avenida Los Restauradores, and records the resource in late November 2025. A separate LACNIC record for 2803:bfd0::/32 assigns the IPv6 block to the same company and address on the same date. These are direct evidence that the company holds internet-number resources and an autonomous-system identity. They are not evidence that it owns the fibre carrying those addresses, controls the poles beneath that fibre, or has subscribers using the entire allocation.

The company's domain is nearly contemporaneous. Verisign's registration record for fibercomdominicana.com shows a creation date of 25 October 2025 and GoDaddy name servers. As observed on 17 July 2026, the public domain redirects to a generic parked landing page. That condition matters because the obvious public surface does not provide plans, a service map, support hours, company terms, outage notices or a network description. A parked domain does not establish that the company is inactive; service can be sold through personal contact, social media, referrals or another brand. It simply leaves the public company record thinner than the routing record.

The sequence is consistent with a small field-services enterprise formalising into an internet provider: wired installation in 2020, a broader commercial name in 2025, a domain in October, and LACNIC resources in November. It is a plausible reading, not a proven corporate genealogy. The records do not disclose incorporation capital, audited revenue, employees, subscriber contracts, a regulator's operating title, asset transfers, insurance or creditor claims. Nor do they show whether FIBERCOM is primarily a retail provider, a wholesale connectivity seller, an installer that now originates addresses, or a mixture of those roles.

That distinction is essential for infrastructure analysis. An installer may know the local pole and conduit environment exceptionally well while owning little plant. A retail provider may control customer installation and support while buying all middle-mile transport. A routed company may hold address space and policy authority while its only physical hand-off is in another operator's cabinet. The trade-name records establish useful human and commercial continuity. The LACNIC records establish network identity.

What remains absent is the legal and operational bridge between the two: who owns which assets, who grants access to them, and who must act when a cable beneath a real bridge needs repair.

Santo Domingo Norte is the only public geographic anchor

The address attached to FIBERCOM's LACNIC resources is the strongest company-specific geographic fact in the reviewed record. It points to Avenida Los Restauradores in Santo Domingo Norte, postal code 11201. An address can locate administration, a home office, a network room, a warehouse or merely a contact. The record does not label it as a headend, point of presence, customer centre or equipment site. It would be unsafe to turn a registration address into a cable route or service boundary.

The surrounding market is substantial. The National Statistics Office's current municipal profile reports a combined 2022 population of 674,274 for Santo Domingo Norte and its La Victoria municipal district, and 80,160 fixed-internet accounts in 2024. The ONE Santo Domingo Norte profile provides context for demand, not a FIBERCOM subscriber figure. Accounts are not necessarily households, the total covers all providers, and the statistic does not identify technology, local market share or the blocks served by this company.

The municipality's scale also changes the meaning of “local.” Dense neighbourhoods can support short fibre drops and efficient truck rolls, yet unplanned street layouts, crowded poles and drainage channels can make extensions difficult. La Victoria and peripheral communities introduce longer distances and lower density. Sabana Perdida, Villa Mella and Guaricano each face different road approaches and watercourses. A provider can be locally important while serving a tiny portion of the municipality; conversely, a modest legal entity can lease wholesale access and cover a broad area without owning a long backbone.

No public company page resolves that ambiguity. There is no reviewed coverage polygon, list of neighbourhoods, installation tariff, postcode checker, customer count, fibre-route map or facility list tied to FIBERCOM Dominicana. The absence must be interpreted narrowly. It does not prove that such information has never been given to customers or the regulator. It means an outside reader cannot determine whether the registered address sits inside the service footprint, at its edge, or away from operating equipment.

Regulatory context supplies another boundary without filling in the company details. The Dominican government's concession service page explains that providing public telecommunications services requires a concession process and supporting information such as an expansion plan. No indexed FIBERCOM concession resolution was located in the materials reviewed for this article. That is not evidence of unauthorised operation: an authorisation may sit under a different legal name, be pending, be available in a record not readily indexed, or the company's role may not match an assumed retail model. It is a disclosure gap that the company or regulator can close with a resolution number and scope.

A useful comparison is INDOTEL's 2025 registration of TRUENET as a reseller in urban Santo Domingo Norte. The TRUENET resolution identifies the upstream supplier and explicitly limits the authorised role. Another 2025 SERVITELECOM resolution names neighbourhoods including Ponce, La Mina and Brisa del Norte in Guaricano. Neither record describes FIBERCOM. Together they show that locally specific public documentation can distinguish a reseller from its wholesale provider and tie a small operator to defined communities.

For FIBERCOM, even a short statement would materially improve the evidence: the company serves named sectors; it operates as a concessionaire or reseller under a cited decision; its core hand-off is in a stated facility or municipality; local access is owned, leased or mixed. Such disclosure would not require publishing sensitive coordinates. It would turn an address into a service geography and establish which physical questions are properly directed to FIBERCOM rather than to an upstream, pole owner or wholesale access provider.

The public map ends at an address and a parked domain

A network map can reveal too much when drawn at cabinet or household resolution, but the present public record reveals too little to test basic resilience. The map that can be reconstructed has only a few nodes. One is the LACNIC contact address in Santo Domingo Norte. A second is the logical autonomous system, AS274266. A third is its visible upstream, AS273217. Farther away sits a public exchange attachment associated with that upstream. Between the address and the autonomous system there is no published fibre line, radio hop, carrier hotel, street cabinet, power site or river crossing.

This is especially important in a city where civil works and water management intersect. In April 2026, the water authority announced an intervention at Cañada Juan Valdez, explaining that dense settlement, accumulated waste and limited drainage had contributed to flooding before the channel reached the Isabela. The CAASD account concerns drainage, not telecommunications. It nevertheless illustrates why a line drawn as a straight geographic connection can be misleading. A short route may cross a channel, use a congested culvert corridor, or enter a road segment where excavation and flood clearance are recurring activities.

Dominican infrastructure rules acknowledge that multiple operators can depend on the same physical supports. INDOTEL's passive-infrastructure sharing regulation covers access to and shared use of passive facilities, seeking efficient deployment and less unnecessary duplication. That legal framework makes several ownership arrangements possible. FIBERCOM could own a cable on someone else's poles, lease a strand, rent duct space, buy an end-to-end wholesale service, or combine those approaches. A shared structure can reduce construction costs while concentrating failure and access authority.

The current domain offers no correction. There is no diagram showing a core site, upstream hand-off, service sectors, redundant ring or maintenance contact. The public routing interfaces also do not carry geographic meaning at street scale. A route collector that sees AS274266 through another autonomous system can show a logical adjacency. It cannot tell whether the hand-off is in Santo Domingo Norte, La Caleta, another Dominican facility or a remote interconnection. It cannot reveal whether IPv4 and IPv6 share a port, chassis, fibre pair and power feed.

An appropriate public map would use zones rather than exact plant coordinates. It could identify a northern service zone, one or more core zones, upstream hand-off zones, the Isabela as a risk boundary, and the broad corridors of primary and alternate transport. Each link could state owned, leased or wholesale; aerial, buried or radio; and primary or standby. Crossings could be labelled by responsible structure owner and authorised repair party without giving a chamber number.

The map should also show where diversity ends: two local paths may merge at one bridge, two upstream circuits may enter one building, or two address families may share one router.

The absence of that map has a commercial consequence. Customers cannot distinguish a genuinely separate backup route from a second service riding the same poles. Enterprise buyers cannot price the risk that an upstream or structure owner controls restoration. Lenders and equipment vendors cannot see which investments would remove a shared point of failure. FIBERCOM itself loses the ability to demonstrate that a small operator can possess strong local engineering even when it buys wholesale transit.

The appropriate conclusion is disciplined. There is no public evidence of a FIBERCOM cable crossing the Isabela, and none should be invented from the registered address. There is also no public evidence that all customers are confined north of the river. The river and its bridges are analytical boundaries because the locality makes them plausible and official records document maintenance and flooding. The company's actual map could show no river dependency, one well-protected crossing, or multiple diverse crossings. Until it is published or independently surveyed, all three remain possible.

AS274266 is real, recent and unusually exposed

The logical network is more visible than the physical one. RIPE's routing history for 38.74.56.0/22 shows AS274266 originating the IPv4 aggregate, with the /22 and four constituent /24 announcements becoming visible from 1 April 2026. ARIN's registration view for 38.74.56.0 identifies the larger parent space under Cogent's 38/8 allocation rather than a public, specific FIBERCOM registration. The prudent formulation is therefore that AS274266 originates the prefix in BGP; the reviewed allocation record does not establish that FIBERCOM owns that IPv4 block.

IPv6 has a cleaner registry chain. RIPE's routing history for 2803:bfd0::/32 shows the LACNIC-registered block first observed at the end of January 2026, disappearing for part of March and returning from 1 April. By mid-July both address families were globally observable, but observation was not stable across all collectors.

The most consequential recent event occurred on 14 July. RIPE's IPv4 update record records 2,564 announcements and withdrawals in the 1-17 July interval, including a broad withdrawal wave around 14:02 UTC on 14 July. The equivalent IPv6 update record records 3,911 updates and a similarly concentrated withdrawal wave. Limited announcements later returned to the collector view.

Those figures describe control-plane messages, not minutes of retail downtime. One underlying change can generate many observations as routes propagate. A collector can lose a path while customers retain service through a route the collector does not see. Maintenance, upstream policy, equipment reload, filtering, address-family changes and physical failure are all possible explanations. No public incident notice from FIBERCOM or its upstream was located to assign cause. It would therefore be wrong to call the event a fibre cut or a confirmed customer outage.

It would also be wrong to dismiss the event. The withdrawal wave shows that a young network's reachability changed sharply across external observation points only three days before this assessment. It creates a concrete question for the operator: what happened, which customers or services were affected, how long did restoration take, and did any alternate path carry traffic? A brief incident statement could separate planned routing work from an avoidable failure and provide the first public restoration metric.

Route-origin security remains another visible gap. RIPE's current validation queries return “unknown” for both the IPv4 origin pair and the IPv6 origin pair, with no validating route-origin authorisation in the returned records. That does not make the announcements invalid; “unknown” is distinct from invalid. It means relying networks cannot use a valid authorisation to confirm that AS274266 is permitted to originate those exact prefixes.

For a small operator, publishing valid origin authorisations is a relatively bounded improvement compared with building a second physical route. It reduces one class of routing ambiguity and demonstrates control over address policy. It does not create capacity, prevent an upstream failure or diversify a bridge. The withdrawal event, route history and security state together portray a network that is genuinely routed and externally visible, but too recent and too lightly disclosed to support strong claims about continuity.

One visible upstream is a logical dependency, not a route survey

RIPE's neighbour view for AS274266 identifies a single direct left-side neighbour: AS273217. Current IPv4 BGP state shows the path reaching FIBERCOM through that autonomous system. Current IPv6 BGP state also places AS273217 immediately before AS274266 in the observed paths. Across both address families, the public control-plane evidence therefore converges on one direct upstream relationship.

LACNIC identifies AS273217 as ACOLME TECH SRL. Its registration record gives a Santo Domingo Este address and dates the autonomous system to 2024. RIPE's neighbour view for AS273217 shows AS272112 on its upstream side and AS274266 among its downstream neighbours. This produces a comprehensible public chain: FIBERCOM announces to ACOLME; ACOLME in turn reaches a larger provider.

The chain does not reveal a contract. FIBERCOM may buy full transit, a managed connection, address transport, or another service. It may have a second private or backup arrangement that current collectors do not observe. AS273217 may carry the session directly or through a reseller. The absence of another visible neighbour is evidence of public concentration, not proof that no contingency exists.

Nor does the chain reveal geography. The AS274266-to-AS273217 adjacency could cross a short patch lead in a shared rack, a metropolitan fibre, a wireless link, a leased circuit or multiple carrier segments. IPv4 and IPv6 could use the same session or separate sessions on the same port. Two sessions could terminate on different routers but enter the building through one duct. A physical survey needs demarcation sites, carriers, route corridors, building entrances, power feeds and equipment ownership; an AS path supplies none of them.

This boundary is particularly important after the 14 July withdrawal wave. If both address families depend on one direct neighbour and one physical interconnection, an upstream port or transport fault can remove both. If they depend on different physical circuits but the upstream changes policy, both can still disappear from collectors. If FIBERCOM maintains local services during a global withdrawal, customer experience may vary by destination. Without traffic, incident and topology evidence, the same public symptoms fit several very different causes.

Single-upstream operation can be an economically rational stage for a new regional ISP. A second full transit service adds recurring cost, router ports, cross-connects, operational complexity and the risk of poor routing policy. The first resilience investment might instead be a diverse transport path to the same upstream, a standby circuit, better local caching, spare optics, longer backup power or stronger field response. The optimal sequence depends on traffic, service commitments and the failure modes already observed.

What cannot be inferred is that one visible upstream automatically equals one fragile cable. A strong single-upstream design can use two diverse circuits to separate routers and facilities under a tested failover arrangement. Conversely, two upstream AS numbers can still share one bridge, one duct or one powered room. FIBERCOM's public evidence establishes the commercial and routing concentration clearly enough to ask for more detail, but physical weakness must be proven with physical evidence.

The most useful disclosure would separate logical from physical diversity. It would state the number of external BGP neighbours, number of transport circuits, number of independent building entrances, whether IPv4 and IPv6 share sessions, and the broad failure domains each circuit crosses. It could report tested failover time without naming exact addresses. That information would allow customers to assess the one-upstream choice without mistaking a route table for a civil-engineering plan.

A 40 Gbps exchange port upstream is not FIBERCOM capacity

Public exchange data adds one attractive number to the chain, but it must be kept in its proper place. PeeringDB's network search for AS274266 returns no public FIBERCOM network entry. This means FIBERCOM has not exposed an exchange profile through that directory as observed; it does not prove that the company has no peering, facility presence or private interconnection.

ACOLME does have public exchange data. PeeringDB's network-exchange record for AS273217 lists an operational 40,000 Mbps connection at PIT DOMINICANO with both IPv4 and IPv6 addresses and route-server participation. The PIT DOMINICANO exchange record associates the exchange with NAP del Caribe in La Caleta, near the eastern side of metropolitan Santo Domingo. This is useful evidence that FIBERCOM's visible upstream participates at a domestic interconnection point.

It is not evidence that FIBERCOM owns, leases or can consume 40 Gbps. The port belongs to ACOLME's public profile. It may serve all of ACOLME's customers, internal traffic and peers. Actual traffic can be far below port speed, while contracted transit and traffic engineering can limit a downstream independently. FIBERCOM's hand-off may be 1 Gbps, 10 Gbps, a rate-limited service, or another value; no public record reviewed here states it.

The exchange port also does not prove end-to-end domestic traffic retention. Some destinations may be reached at the exchange, others through ACOLME's upstream. Content delivery may depend on caches or private links not shown. The number says nothing about the transport from FIBERCOM's equipment to ACOLME or from ACOLME to the exchange. A 40 Gbps port at La Caleta cannot protect a 1 Gbps access hand-off, a single fibre on a bridge, or a router with limited public evidence forwarding capacity.

Capacity has at least four relevant layers. Installed capacity is the physical rate of ports, optics and radio channels. Contracted capacity is what a buyer is entitled to use. Usable capacity is what remains after overhead, congestion, policy and equipment limits. Survivable capacity is what remains during the loss of the largest link, node, facility or power source. The public 40 Gbps figure belongs only to the installed port layer of FIBERCOM's upstream at one exchange.

For customers, the missing denominator is equally important. A 10 Gbps hand-off can be ample for hundreds of light users or congested under a smaller base of high-speed plans. Headline retail speeds do not add arithmetically because customers are not all active at once, but peak concurrent demand, video traffic, cloud backups and business commitments determine the experience. FIBERCOM publishes no reviewed plan speeds or customer count on its parked domain, so even a verified upstream contract would not yield a defensible contention estimate.

The exchange evidence should therefore be read as a potential efficiency advantage, not as a FIBERCOM performance claim. A domestic upstream with an exchange presence may shorten paths to participating networks and reduce expensive transit. Whether that benefit reaches FIBERCOM depends on routing policy, the downstream contract, physical transport and available capacity. The correct next evidence would be FIBERCOM's hand-off rate, peak 95th-percentile traffic, packet loss, latency to key domestic and international destinations, and throughput remaining after a circuit failure.

A bridge can combine conduit, traffic, flood and repair risk

Physical concentration is rarely a single entity. Consider a hypothetical FIBERCOM cable that approaches an Isabela crossing along a road, enters a conduit at the bridge, exits into a pole line and terminates in a powered cabinet. The bridge is one dependency, but the route also depends on the access road, chamber drainage, attachments, pole condition, cabinet power, permissions and an upstream circuit beyond the cabinet. A failure at any one may remove service; a flood or civil intervention may affect several together.

INDOTEL's general internet-access service regulation sets a national framework for continuity, quality and relations among providers and users. It does not publish FIBERCOM's topology. Its relevance is that service continuity is an operating obligation even when infrastructure is shared or supplied by others. A retail customer's claim normally runs to the provider, while restoration may depend on a chain of structure owners, carriers and contractors.

The quality regime adds time discipline. INDOTEL's quality-of-service regulation describes continuity and reporting expectations, including rapid notification for specified network events and periodic reporting duties for providers above relevant subscriber thresholds. The reviewed public record does not establish whether those thresholds apply to FIBERCOM or whether it has filed any particular report. The regulation nevertheless provides a useful operational standard: detection and communication begin before a field crew finishes the repair.

A bridge route needs explicit authority. The operator must know who can close a lane, open a chamber, access an abutment, attach a new cable, enter a private lot, isolate power and approve an emergency splice. If FIBERCOM owns only the customer-facing service, it may need the wholesale carrier to dispatch. If it owns the cable but not the structure, it may still wait for a permit. If a contractor holds the splice equipment, travel time and after-hours terms become part of network availability.

Flooding changes the sequence. A crew may detect optical loss instantly but be unable to reach the break until water recedes. The fastest path to the suspected fault can be the same road closed by authorities. Spare cable stored at the central office may sit on the wrong side of a flooded crossing. A portable generator is useful only if it can be transported to the cabinet and connected safely. An alternate fibre route that shares the same chamber can be damaged by the same water and debris.

These risks can be managed without owning every asset. Contracts can specify response times, emergency access, spare holdings, escalation contacts and restoration priority. Cables can be installed in separate structures or on different banks. Network design can keep a local ring alive even when global transit is lost. Power can be monitored remotely, and batteries can be tested under load. Crews can rehearse entry from both sides of a likely isolation point. What matters is control and verified response, not a simplistic owned-versus-leased label.

FIBERCOM's early trade-name activity suggests installation and repair capability may be part of its identity. That could be a material advantage: local knowledge and direct labour often shorten restoration. But a 2020 activity description does not prove the size, employment status, equipment or present availability of the 2026 field team. The public record lacks support hours, dispatch locations, repair targets, vehicle access, splicing inventory and recent incident results.

The bridge test can be answered with bounded evidence. Publish a high-level route-risk register; identify the owner and repair authority for each critical crossing; record whether alternate routes share approaches; state tested battery endurance; disclose spare strategy and escalation times; report restoration performance in ranges. Such information would let enterprise customers understand the service without exposing precise cable coordinates. Until then, the bridge remains a compound dependency whose relationship to FIBERCOM is plausible but unproven.

Installed capacity and usable capacity remain undisclosed

The address resources set an upper boundary on identifiers, not throughput. An IPv6 /32 is vast enough for extensive hierarchical assignment, but it has no intrinsic bit rate. An IPv4 /22 contains 1,024 addresses before exclusions and subnetting, yet address count does not reveal subscribers because of private addressing, carrier-grade translation, static assignments or unused space. Announcing four more-specific /24s alongside an aggregate can support traffic engineering, but it does not multiply transport capacity.

The same distinction applies to autonomous-system visibility. A route seen by many collectors may carry little traffic; a route seen by fewer collectors may still serve customers through a major provider. Update counts measure routing messages rather than payload. The July 14 withdrawal wave therefore provides strong evidence of a control-plane event but no direct measure of lost gigabits, affected accounts or revenue.

FIBERCOM's missing capacity evidence begins at the customer edge. The parked site publishes no current retail speed tiers, oversubscription policy, service-level commitment or technology description. The record does not say whether last-mile service is fibre-to-the-home, fixed wireless, Ethernet, resold access or a combination. Each medium has different constraints: optical split ratios and power budgets, radio spectrum and line of sight, switch-port rates, or wholesale product limits.

The middle mile is equally opaque. There is no public transport rate between a local aggregation point and ACOLME, no interface counter, no optical-light budget, no radio modulation, and no route length. There is no statement of whether a second circuit exists in standby or whether its capacity is reserved. The current BGP views show public reachability through one neighbour, but a BGP session can ride a very small link or a large one.

The statistics regulator offers a template for the kind of operating data that exists at national level. INDOTEL's 2024 statistical-indicator rule requires service information to be broken down by dimensions including municipality, technology and speed. That does not make FIBERCOM's figures public or prove that every reporting obligation applies to it. It shows that customer, geography and technology can be measured separately rather than collapsed into a single coverage claim.

Useful disclosure would begin with ranges. FIBERCOM could state installed external capacity, peak utilised capacity and capacity available after loss of the largest circuit without exposing commercially sensitive prices. It could give active subscriber bands, the proportion served by each access medium, typical busy-hour loss and latency, and the number of customers behind each major aggregation node. It could state whether capacity is symmetric across IPv4 and IPv6 and whether domestic peering remains available when international transit is impaired.

Measurement must also distinguish reserved from tested backup. A standby link listed in a contract may not carry the current route table, customer load or IPv6. Batteries with a nominal eight-hour rating may deliver less under heat and age. A radio bypass may have clear line of sight in the dry season but suffer interference or lose access to a shared tower during an emergency. Survivable capacity is credible only after a periodic failover exercise or a real event with recorded results.

In economic terms, the disclosure would illuminate where the next peso should go. If busy-hour utilisation is low but one transport corridor dominates, route diversity may matter more than a larger port. If physical paths are diverse but the upstream is congested, contracted capacity or a second transit provider may matter more. If both are adequate but repairs take too long, spares and local labour may deliver the largest improvement. The present evidence cannot rank those choices because installed, usable and survivable capacity are all undisclosed.

Field labour and repair authority are part of the network

Regional internet service is often described through electronics and fibre counts, but restoration is performed by people operating within contractual and geographic limits. FIBERCOM's documentary origin in wired installation and repair makes labour especially relevant. If those skills remain inside the company, it may diagnose and splice local faults faster than a reseller waiting for a distant carrier. If work is outsourced, quality may still be excellent, but response depends on the contractor's coverage, priority and stock.

The public materials identify Jonatan Santos Fernandez repeatedly, first in the 2020 trade-name record, again in the 2025 commercial-name application and in the LACNIC contact. This continuity suggests a hands-on founder or operator, but the record does not establish an employee count or round-the-clock duty roster. One experienced person can be an asset and a concentration risk at the same time. Illness, simultaneous faults or a permit negotiation can overwhelm a very small team even when technical competence is high.

Local support labour has three distinct functions. The first is customer installation and fault triage: checking power, optical levels, Wi-Fi conditions and drop cable. The second is outside-plant repair: locating a break, obtaining access, deploying a safe crew, splicing and protecting the closure. The third is network operations: recognising a route withdrawal, isolating an upstream or equipment problem, communicating with carriers and restoring policy. A company can perform one function internally and buy the others.

Santo Domingo Norte's size and traffic make dispatch geography material. A crew based near Los Restauradores may reach Villa Mella quickly under normal conditions but face delay across congested avenues or flooded approaches. A failure near La Victoria can require a longer journey. A river or bridge intervention can separate the crew from the fault. The customer experience therefore depends not only on mean repair time but on where vehicles, spares and authorised personnel are positioned relative to likely isolation boundaries.

Repair authority can dominate repair skill. A technician may identify a damaged third-party fibre but lack permission to enter the chamber. A provider may have replacement cable but no approved bridge attachment. An upstream may own the optic and require its own engineer. The pole owner may prohibit work until electrical danger is cleared. Contracts should assign responsibility before the event, with named escalation paths and access commitments that survive nights, weekends and storms.

Incident communication is another labour task. The July 14 routing event offered a chance to state whether maintenance was planned, whether services were affected and when full visibility returned. No public FIBERCOM status page or notice was found on the parked domain. That does not show that customers received no messages; phone or private channels may have worked. A durable status surface would nevertheless reduce uncertainty for enterprise customers and provide a history against which restoration promises can be assessed.

Evidence of local capability need not expose staff identities. FIBERCOM could publish support hours, after-hours escalation, broad crew zones, target acknowledgement and restoration ranges, number of equipped field teams, access to a fusion splicer and optical time-domain reflectometer, and the locations of spare holdings at municipality level. It could report the share of incidents resolved internally versus escalated to an infrastructure supplier. It could also conduct a crossing-loss exercise and record the elapsed times for detection, dispatch, access, splice and route recovery.

Such disclosure would connect the company's earliest commercial description to its present routed identity. Installation skill becomes infrastructure value when paired with authority, stock, transport, monitoring and repeatable response. Without those elements, the strongest technician cannot restore an inaccessible third-party route. With them, a small local operator can outperform a larger provider whose nearest crew is far away. FIBERCOM's public record leaves that competitive question open.

What evidence would prove an independent crossing

The current record supports a precise conclusion. FIBERCOM Dominicana exists as a named commercial operation and current holder of AS274266 and an IPv6 /32. Its prefixes are visible globally through one direct upstream, ACOLME. A significant withdrawal wave affected external observation on 14 July 2026. The company's only public geographic anchor is in Santo Domingo Norte, a large municipality north of the Isabela with documented bridge maintenance and flood exposure. None of those facts establishes the route, ownership, capacity or recovery design of its local plant.

The first missing item is a service and facility map at safe resolution. It should name neighbourhood or municipal zones, core and aggregation zones, upstream hand-off zones and major physical boundaries. Each transport segment should be classified as owned fibre, leased strand, managed circuit, wholesale service or radio. The map should show whether an Isabela or other water crossing exists, but it need not reveal exact chambers, poles or customer addresses.

The second item is a shared-risk statement. For each primary and alternate path, FIBERCOM should identify common bridge approaches, ducts, poles, buildings, power feeds, carriers and contractors. An alternate that rejoins the primary before the hazard boundary is not fully independent. A second upstream terminating in the same chassis and conduit removes some policy risk but not the largest physical risk. Conversely, two diverse circuits to ACOLME can provide meaningful physical protection even though external route collectors still show one neighbour.

The third item is capacity evidence. Installed port speeds should be separated from contracted, busy-hour and survivable throughput. The company should publish a dated test in which the primary path is withdrawn and the remaining path carries representative IPv4 and IPv6 load. Latency, loss, route convergence and available throughput should be recorded. The 40 Gbps PIT DOMINICANO port associated with ACOLME should be treated only as upstream context unless a contract and end-to-end measurements tie a defined share to FIBERCOM.

The fourth item is repair evidence. A crossing register should identify the structure owner, cable owner, maintenance party, emergency-access authority, spare type, crew origin and target response. A drill should test a failure after heavy rain, when a normal approach is unavailable. The result should record time to detection, correct fault location, dispatch, site access, temporary restoration and permanent repair. If a wholesale provider controls any stage, its contractual target should be included.

The fifth item is regulatory and commercial clarity. A public concession, resale or other operating reference would establish the service role and authorised geography. A working company site should give legal identity, plan terms, support channels, coverage language and outage communication. Valid route-origin authorisations for the announced IPv4 and IPv6 prefixes would close a separate, readily remediable routing-security gap. None of these steps alone proves physical resilience, but together they make responsibilities auditable.

Finally, the July 14 event deserves a short factual post-incident account. The company or upstream should state whether the change was planned, what failed or changed, the scope of customer impact, the restoration time and whether backup arrangements worked. If no customers were affected, that is valuable evidence too. External update counts cannot answer those questions, and restraint about causation should not become an excuse to leave the event unexplained.

FIBERCOM may already have strong local engineering, a protected route and a capable repair team. The available evidence does not contradict that possibility. It also leaves open a more concentrated design in which one upstream, one transport path and one small crew carry most of the service risk. The difference lies beneath the logical route: in the bridge attachment, duct, splice, power feed, contract and dispatch path. A credible independent crossing is proven when its alternate survives the same civil event, reaches a functioning hand-off, carries measured load and can be repaired by an authorised crew.

Until FIBERCOM publishes that evidence, its routed identity is visible, but its physical escape path remains undisclosed.