MUNDO1TELECOM lists a 100 Gbps Santiago exchange port, but its storm-route separation is not public

Summary

• MUNDO1TELECOM's public record now spans three very different businesses: a nationwide fixed-voice concession dating to 2006, internet and subscription-television authority granted in 2023 for twelve named municipalities or districts across six northern and central provinces, and current wholesale or public-sector connectivity visible through reseller approvals and education-ministry payments. Those records establish activity and responsibility, but they do not identify a current as-built fibre route, coastal hand-off, pole owner, duct owner, field depot or restoration sequence.
• AS272054 provides a measurable routing surface. It originates two IPv6 /40s, and every collector path examined reaches the origin through AS269931 immediately before MUNDO1TELECOM. Separately, PeeringDB lists an operational 100 Gbps port at the Azteca Telecom exchange and a presence at CEMO DC in Santiago, while the Santiago Internet Exchange portal labels a 10 Gbps membership in La Vega as not connected. These are useful logical and facility observations, not proof of independent transit contracts or physically separated backhaul.
• The operator markets fibre service, redundancy, multiple national Layer 2 providers and a fully ringed national network, but it publishes no shared-risk route map, busy-hour utilisation, protected capacity, generator endurance, spare inventory or repair-time distribution. The decisive disclosure is therefore not another large port number. It is a route-by-route account of where circuits share enclosures, poles, ducts, bridges, carrier hand-offs and power, followed by evidence that traffic and crews can actually take the surviving path during a storm.

The enclosure is the real test of two routes

A Caribbean squall drives rain sideways across a coastal road. Brown water gathers around a roadside chamber while a technician works from the dry side of the verge, checking a sealed fibre enclosure mounted above the flood line. Two black cables approach from different directions. On a commercial drawing they could be called diverse. Inside the enclosure, however, they may enter the same splice tray; farther inland they may cross the same bridge, climb the same poles or terminate on equipment fed by the same battery string. If water reaches one compromised seal, both drawn paths fail at one physical point.

This is an illustrative resilience test, not a report of a documented MUNDO1TELECOM incident or asset. It is the right opening because the operator's most ambitious public claim is physical. Its company description says that it has multiple IP points of presence, diversity from the NAP of the Americas, multiple national Layer 2 providers, redundant IP cores and a national network that is fully ringed. Each phrase can be true at one layer while being false at another. Two routers in two racks are not two access roads. Two carrier contracts are not two ducts. Two fibre pairs inside one sheath are not two routes. A ring that closes through one coastal carrier room is not protected from loss of that room.

The Dominican Republic has repeatedly supplied the hazard conditions that expose those distinctions. The US National Hurricane Center's final report on Hurricane Fiona records the storm's 19 September 2022 Dominican landfall and damage that included a fatal falling power pole. Its earlier observations describe heavy rain, flooding and destructive winds. The lesson for a fixed network is not that every storm will strike the same province. It is that water, wind, debris, unstable ground, road access and electricity can fail together over a wide area. A route designed only around normal traffic may share precisely the assets that a storm removes simultaneously.

Public evidence does not locate a MUNDO1TELECOM-owned coastal enclosure, submarine landing connection or road-side backhaul. Nor does it identify the carrier that would carry traffic to the Miami facility named in the company's claim. That absence has to constrain the analysis. The relevant question is whether the company can demonstrate that the first and second paths leave each service area through different physical structures, reach different powered sites, meet different carrier failure domains and can be repaired from opposite sides of a blocked corridor.

Until those facts are disclosed, the 100 Gbps exchange entry discussed later is a measure of a port, not of survival.

A 2006 voice concession became a 2023 broadband footprint

MUNDO1TELECOM's legal history begins with fixed telephony rather than residential fibre. INDOTEL's Resolution 113-11 approved the concession contract signed in October 2011 and recited the original June 2006 grant. The authority covered local, national long-distance and international fixed calls throughout the Dominican Republic for twenty years. It also required commercial arrangements with the national telephone providers to which MUNDO1TELECOM connected. That obligation matters because interconnection was part of the service from the start: the company could operate its own switching and customer access while still depending on another network to complete traffic.

The interconnection record became more concrete in 2014. INDOTEL's Resolution DE-003-14 approved an indirect interconnection contract between MUNDO1TELECOM and COLORTEL. It is historical voice evidence, not a current internet-transit map. Nevertheless, it shows the recurring economic structure of a smaller operator: customer service depends on a boundary where another provider supplies reach. The operational quality of that boundary is defined by ports, signalling, circuits, commercial terms and restoration responsibilities that the retail customer does not see.

Broadband authority arrived much later. In Resolution 050-2023, INDOTEL expanded the concession to internet access and subscription television in Santiago, Bisonó and La Canela in Santiago province; Mao, Esperanza and Amina in Valverde; La Vega and Jarabacoa in La Vega province; San Francisco de Macorís in Duarte; Salcedo and Tenares in Hermanas Mirabal; and Moca in Espaillat. The regulator recorded that the request had been filed in 2017 and that legal, economic and technical reviews were completed in 2022. The decision accepted a minimum expansion proposal, but the public resolution does not reproduce a route-level construction plan that a customer could use to identify the actual feeder, backhaul, node or alternate path serving an address.

The distinction between authority and construction is essential. A concession says what service may be offered and where; it is not a certificate that every planned cable was built, every town has two paths, or every path is lit at protected capacity. Resolution 050-2023 also dealt with a competitor's objections. One allegation was that internet had been marketed before the expansion was authorised. The regulator expressly found that allegation unsupported by evidence. It should not be recycled as a fact. What the decision does establish is a dated legal footprint and a remaining term tied to the original concession.

A newer Resolution 044-2026 index page identifies an INDOTEL decision on MUNDO1TELECOM's renewal request for nationwide fixed voice and the same named internet and television areas. The accessible page confirms the existence and subject of that decision, but the preview available for this analysis does not expose the operative clauses. It would therefore be unsafe to invent a renewed term or conditions from the title alone. Current service, routing and payment records demonstrate continuing operations; the precise post-renewal legal term should be taken from the complete signed instrument.

Twelve authorised areas support a wider chain of customers

The retail surface is centred in Santiago. MUNDO1TELECOM's current home page gives an Avenida Francia address, promotes “real fibre optic” service, and combines internet with IP telephone offers. Its internet page advertises included Wi-Fi equipment, installation, round-the-clock technical support and a series of residential plan names. These pages prove an active sales and support interface. They do not publish address-level serviceability, subscriber counts, access-medium percentages, optical split ratios, tower locations or the boundary between owned and leased plant.

The authorised geography is broader than one city but still primarily northern and central. Santiago, Navarrete, Mao, Esperanza, La Vega, Jarabacoa, San Francisco de Macorís, Salcedo, Tenares and Moca form a dispersed set of population centres separated by mountain approaches, river crossings and provincial road corridors. A company can reach them through owned fibre, leased waves, wholesale Ethernet, radio backhaul or combinations of those media. Nothing in the public concession assigns a medium to a town.

A coloured coverage map would not solve that problem unless it also showed the transport route, node type, owner, hand-off and common-risk segments.

Reseller approvals show that MUNDO1TELECOM's infrastructure responsibilities extend behind other retail names. INDOTEL's 2024 AW WIFI decision records a reseller contract under which AW WIFI would resell MUNDO1TELECOM internet in Santiago. The document explains an important division: a reseller answers directly to its customers for availability and quality, while the concessionaire remains responsible for technical rules governing the networks that provide the service. In practical terms, an access or backhaul failure can create two queues at once—end users call the reseller, and the reseller waits on the underlying operator.

That chain is not isolated. INDOTEL's 2025 ADRIEL SOLUTIONS decision records a thirty-six-month agreement to resell MUNDO1TELECOM internet in La Yagüita de Pastor, Bella Vista and La Barranquita in urban Santiago. Those named neighbourhoods add useful service evidence, but the approval still does not reveal whether ADRIEL has an independent access loop, where its demarcation sits, or whether other resellers converge on the same aggregation chassis. Customer counts behind each reseller are also absent.

Public education payments reveal another, socially important footprint. A May 2025 supplier-payment report records MUNDO1TELECOM internet service associated with 206 schools in multiple provinces. A later April 2026 accounts-payable report includes service to 209 schools. These are strong signs of an operating network and recurring institutional demand. They do not disclose per-school bandwidth, service-level targets, alternate access, outage history or the number of schools sharing one upstream junction.

The public map has points but no continuous line

The strongest map evidence consists of named service places and two interconnection locations. The concession supplies municipalities and districts. The website supplies the Santiago office. Peering records add a Santiago facility, and the exchange portal adds a La Vega location. Together they form points on a map, not a route between the points. No public record traces a fibre from Santiago to Navarrete, across to Mao, south to Jarabacoa, east to San Francisco de Macorís or onward to a carrier exit.

PeeringDB's network-facility record places AS272054 at CEMO DC in Santiago. The separate CEMO facility record gives the site as Calle 7 at Calle 12, El Ensueño, with coordinates near 19.443373, -70.697215. It reports 48 VDC and 480 VAC availability and says the facility has diverse serving substations. These fields are valuable for locating a declared interconnection point. They are self-reported directory data, not an engineering inspection, and they say nothing about the route from Avenida Francia, the access carrier, entry conduits, flood elevation, generator fuel or whether two MUNDO1TELECOM circuits enter on different sides.

PeeringDB's Azteca Telecom exchange record likewise identifies an exchange in Santiago and associates it with CEMO DC. The record says the exchange has one facility and lists thirteen networks. That supports local traffic-exchange potential: two Santiago networks may avoid hauling every packet to a remote city. But the exchange's presence at one building can also concentrate risk if an operator treats it as both the primary and alternate hand-off. Exchange membership does not reveal whether transit, peering and customer aggregation share the same router, rack, meet-me cable or power distribution unit.

The Santiago Internet Exchange portal provides a geographically different observation. Its MUNDO1TELECOM member page lists a 10 Gbps port at a La Vega location, with IPv4 and IPv6 exchange addresses, while labelling the member “not connected.” That is not evidence of live alternate capacity. It is evidence of a membership and configured-looking attributes whose operational state is explicitly negative on the public page. The status could change after retrieval, but it should not be promoted into a surviving path without a connected state and observed traffic.

The operator's NAP-of-the-Americas claim adds an international point without filling in the line. A path from the Dominican Republic to Miami necessarily depends on carrier transport and submarine systems, but the public company description does not name a Dominican hand-off building, cable system, landing station, capacity, carrier or protection arrangement. It also does not say whether “diversity from” means two international systems, two national circuits to one system, or two router sessions over one carrier product. A defensible map therefore stops at the declared points and marks every connecting line as undisclosed.

That constraint changes how coverage should be valued. A service point in Jarabacoa and an exchange point in Santiago may be separated by many road kilometres, yet the risk distance could collapse to one bridge approach, one carrier circuit or one aggregation shelf. Conversely, two towns that look close on a provincial map could be reached through genuinely independent corridors. Straight-line distance, municipal boundaries and routing latency cannot decide the matter. Only the operator and its carriers possess the splice plans, circuit orders and access records that can turn geographic separation into an auditable failure-domain statement.

A claimed national ring is not an as-built shared-risk map

A ring is an architectural idea: traffic can move in either direction around a loop. Resilience depends on what the two directions do in the ground and at each node. If eastbound and westbound fibres occupy opposite sides of the same sheath, one cut opens the ring twice. If they use different sheaths on the same overloaded pole line, one fallen pole can still remove both. If they take separate streets but cross the same bridge, enter the same chamber or terminate in one unprotected building, the common point defeats the drawing.

MUNDO1TELECOM's marketing language is unusually categorical. A “100% ringed” national network sounds like an as-built condition, not an aspiration. Yet the company publishes no fibre-kilometre inventory, node list, ring spans, splice schedule, route photographs, carrier hand-offs or shared-risk groups. There is no indication of which circuits are owned, indefeasibly leased, bought as Ethernet services or supplied as radio links. Without that layer, the claim cannot be reconciled with the much narrower regulatory footprint or with the absence of a public map.

Dominican rules illustrate the many parties that can sit inside one apparent route. INDOTEL's passive-infrastructure decision treats poles, ducts, towers and related facilities as shareable infrastructure and recognises fibre backhaul with data-transport characteristics as an essential facility in the interconnection context. An ISP may therefore control the active service while relying on municipal permission, an electricity distributor's pole, another telecommunications provider's duct or a contracted fibre segment. Commercial diversity and physical ownership are different questions.

The national access-network procurement specification gives a useful benchmark for what a buildable design records. INDOTEL's access-network requirements call for detailed topology, component interconnection, coordinates where applicable, third-party transport, management systems, access, transport and international-connectivity layers. They discuss splice intervals and slack, civil works, facility standards, permits and the option of using national fibre transport. These requirements apply to that public project, not automatically to MUNDO1TELECOM. They demonstrate, however, the evidence needed to distinguish a sales phrase from a maintainable network.

A useful disclosure would split every claimed ring into spans and nodes. For each span it would state medium, owner, route kilometres, conduit or pole system, bridge and river crossings, fibre count, lit pairs, splice enclosures and alternate access for crews. For each node it would state the protected equipment, power feeds, battery and generator endurance, cooling, fuel arrangements and spare optics. For each carrier hand-off it would state facility, port, capacity, contractual protection and whether the carrier's two paths remain separated beyond the demarcation.

Confidential coordinates can be generalised without hiding the shared-risk answer.

Maintenance history belongs beside that inventory. A route that was independent when commissioned can lose separation when a road project moves both cables into one temporary conduit, when a replacement pole gathers formerly separate spans, or when a carrier migrates two services onto one optical system. Planned-work notices, emergency reroutes and post-repair drawings should therefore update the risk register. Without that discipline, a true ring can quietly become a common corridor long before the next storm reveals the change.

AS272054 has two IPv6 routes and one visible neighbour

The logical network is newer and more measurable than the physical one. LACNIC's AS272054 registration identifies MUNDO1TELECOM, S.R.L. as the active registrant, with a registration date of 19 January 2022 and Santiago contact information. The separate IPv6 allocation record assigns an active 2803:dd10::/32 block to the company. Registration establishes the right to originate routes and manage address space. It does not show how many routers, circuits or locations are in use.

RIPEstat's announced-prefixes observation, covering the period through 16 July 2026 at retrieval, shows two routes originated by AS272054: 2803:dd10:fe00::/40 and 2803:dd10:ff00::/40. It does not show a currently originated IPv4 prefix in that observation. That does not mean MUNDO1TELECOM has no IPv4 service. Customers may use addresses originated by another network, private addressing with translation, or routes outside the collector's view. It does mean that a claim about current provider-independent IPv4 origin should not be made from this evidence.

The individual routing-status record for 2803:dd10:fe00::/40 shows AS272054 as origin, first observed in September 2023 and visible to all 321 IPv6 collector peers reported at retrieval. The corresponding record for 2803:dd10:ff00::/40 shows the same origin, a first observation in April 2022 and the same reported visibility. These are healthy visibility signals. They reveal neither traffic volume nor the number of physical links carrying the announcements.

The ASN-neighbour summary shows one adjacent ASN: AS269931, Wireless Multi Service Vargas Cabrera, a Dominican network registered in Navarrete. Both /40s tell the same story at path level. In the BGP state for the fe00 route, every sampled collector path examined reaches AS272054 immediately through AS269931; the ff00 route state has the same immediate neighbour. Repetition of AS269931 in some paths is consistent with path prepending, not with a second provider.

CAIDA's AS Rank record independently simplifies the topology to one provider, no peers and no customers. That inference can lag newer relationships and should not override live contractual evidence. It nevertheless corroborates the one-visible-provider picture. The route-origin security view is also incomplete: RIPEstat's validation result for the fe00 /40 reports an unknown state with no validating authorisation at retrieval. Origin validation is a routing-security control, not an availability control, but publishing valid authorisations for both live routes would remove one avoidable uncertainty.

The conclusion must remain bounded. One visible adjacent ASN does not prove one physical circuit; MUNDO1TELECOM could buy two separated links from AS269931. Conversely, two sessions to the same ASN would not prove diverse last miles. Collector data can miss private peering, standby routes and sessions that are suppressed in normal operation. The company can resolve the ambiguity by naming the active and standby providers, facilities and capacities, then describing where their local loops, carrier backbones and international exits cease sharing risk.

Santiago lists 100 Gbps while La Vega says not connected

PeeringDB's network record describes MUNDO1TELECOM as a cable, DSL or ISP network with an open peering policy. It leaves traffic volume and geographic scope undisclosed, and self-reports one IPv4 and one IPv6 prefix. The record includes both an exchange connection and a facility presence. Like every directory maintained from entity submissions, it is most reliable as a statement of what the network has declared, with timestamps, rather than an independently measured inventory.

The striking figure is in the network-exchange record: an operational 100,000 Mbps port for AS272054 at the Azteca Telecom exchange, with IPv4 address 190.110.33.229, IPv6 address 2803:e610:0:4::62 and route-server participation. A 100 Gbps exchange port is material local interconnection capacity. It can reduce latency and paid transit for traffic exchanged with entities on that fabric. It does not mean 100 Gbps of internet transit, 100 Gbps of customer demand, 100 Gbps of protected backhaul or 100 Gbps available after a fibre cut.

The facility relationship is recorded separately. PeeringDB's MUNDO1TELECOM facility entry places the company at CEMO DC from May 2025. The record does not list a second MUNDO1TELECOM facility. That may reflect incomplete disclosure rather than actual concentration, but it prevents a public claim of site diversity. A second router in the same building could protect against a card failure while remaining exposed to the building, meet-me room, access duct and local power boundary.

CEMO's data-centre page markets N+1-or-better uninterruptible power, backup generators, continuous cooling and redundancy for power, cooling and interconnection. These are facility claims, not measured service records for MUNDO1TELECOM's rack. The page does not state the operator's subscribed power feeds, battery runtime, generator fuel endurance, maintenance history or cross-connect paths. CEMO's services page also markets fibre, GPON and high-speed radio implementation, network monitoring and colocation. It does not say which, if any, of those services MUNDO1TELECOM purchases.

The La Vega portal creates a useful contrast. It displays a 10 Gbps port but says the MUNDO1TELECOM member is not connected. The Santiago directory displays a 100 Gbps port as operational. These statuses can coexist: they refer to different exchange organisations and locations. They should not be added into 110 Gbps, treated as alternate transit, or used to infer a Santiago-La Vega ring.

Capacity at an exchange becomes resilience only when the operator shows that customer traffic can reach it through a surviving route, that useful peers or transit remain reachable, and that the port, router and building have enough power and headroom under the failure condition.

Port speed is not usable storm capacity

Capacity claims live at several layers. The retail site uses “MB” in labels for 20 and 100 speed plans. In consumer broadband context the intended measure is likely megabits per second, but the page itself should be quoted as displayed rather than silently converted into a technical guarantee. It does not publish upstream rates for every plan, contention ratios, optical split occupancy, minimum throughput, busy-hour measurements or whether the advertised rate is sustained beyond the access segment.

The exchange port is four orders of magnitude larger than a 20-unit retail label, but that comparison says little by itself. A 100 Gbps interface may carry peering only, may be lightly used, may connect through a smaller transport circuit, or may share a router whose transit links are smaller. PeeringDB marks MUNDO1TELECOM's traffic as undisclosed. The public routing view shows only two originated IPv6 /40s and one immediate neighbour, which is a topology observation rather than a throughput meter. There is no public 95th-percentile graph, committed information rate, burst policy, transit invoice, packet-loss series or saturation record.

School contracts make the missing denominator more important. Two hundred and nine educational sites can represent very different loads depending on per-site speed, concurrency, filtering, caching and school schedules. The payment records establish scale in endpoints, not aggregate bandwidth. Reseller demand is similarly opaque: a wholesale hand-off could serve a neighbourhood from one gigabit port or several protected ports, and the reseller's customers may concentrate peak traffic at the same evening hour as residential subscribers.

Failure-state capacity is the number that matters during a squall. Suppose normal traffic uses two 10 Gbps transport paths and a 100 Gbps exchange port. If one transport path fails and the surviving one is already carrying 7 Gbps, only 3 Gbps of nominal headroom remains before overhead and operational reserve. The exchange port's unused 90 Gbps cannot cross a 10 Gbps bottleneck. If both transport services share one coastal sheath, the surviving capacity is zero. If the alternate path reaches only an exchange and not full internet transit, some destinations remain unreachable even while local peering works.

MUNDO1TELECOM could publish useful capacity evidence without revealing customer secrets. For each service region it could disclose aggregate access capacity, protected backhaul capacity, normal and single-failure busy-hour utilisation, topological oversubscription and the slowest surviving segment. For each exchange and upstream it could disclose port speed, contracted usable capacity, traffic role and whether it remains reachable after each defined failure. Historic month-by-month utilisation bands would show whether expansion precedes congestion.

A table of successful failover tests would turn a nominal port into evidence of capacity that survives.

Quality under load matters as much as raw throughput. The surviving path needs acceptable packet loss, latency and jitter for voice, interactive education and business applications; a congested alternate may keep routing sessions up while making service practically unusable. Measurements should therefore cover the first minutes after reconvergence, the sustained busy hour and the period when repair traffic, monitoring and customer retries add load. A resilience promise is credible only when the alternate carries the services customers actually use, not merely a test packet.

Poles, ducts, water and power define the failure domain

The physical access network begins where the clean facility record ends. Fibre may leave a building through an underground duct, rise onto electricity poles, cross a river on a bridge, enter a roadside cabinet, split through passive optical equipment and terminate at a home, school or reseller. Every transition adds an owner, enclosure, permission and possible repair delay. A local operator's engineering quality is visible not only in optical budgets but in drainage, cable slack, pole loading, labelling, access rights and the ability to isolate water-damaged sections.

INDOTEL's 2024 infrastructure standardisation resolution defines telecommunications infrastructure broadly to include ducts, poles, towers, cabinets, conduits and underground or aerial cable. Its recommended municipal framework requires technical plans, identification methods, permits and agreements when third-party poles are used; it also assigns maintenance responsibilities to infrastructure owners. The resolution addresses national installation and public-space problems, not MUNDO1TELECOM's compliance at a particular site. It confirms why a route cannot be understood from an ASN alone.

Water ingress is a design and maintenance question. A sealed closure above normal standing water may still fail if a gland is damaged, a cable is pulled beyond its bend radius or repeated flooding pumps contaminated water through a weak entry. Underground chambers need drainage and accessible lids; aerial slack loops need secure supports; bridge crossings need protection from debris and scour. Separate approaches into one closure can reduce cut exposure on one street while retaining a common enclosure risk. A credible storm audit therefore photographs and inspects the precise common points rather than merely counting route lines.

Power has the same layering problem. CEMO advertises robust facility systems, but customers also depend on optical line terminals, routers, radio sites, cabinets, school equipment and home optical terminals outside the data centre. Batteries age. Generators need fuel, cooling and safe access. A pole route can remain intact while an unpowered aggregation node makes it useless. The World Bank's Dominican post-storm recovery account describes past widespread damage to electricity, transport and water infrastructure and the restoration of 152 kilometres of transmission lines. It is not a MUNDO1TELECOM outage history; it shows that electricity and road failures are credible correlated constraints.

The missing public figures are straightforward: battery minutes under actual load at each critical node, generator runtime at rated load, refuelling priority, number of mobile generators, fuel stored before storm season, pole-owner contacts, duct access arrangements, closure replacement stock and cable lengths held in each depot. The company should also identify which nodes have remote telemetry when commercial power fails and whether monitoring links themselves use the failed network. Recovery starts with a correct alarm and an accessible spare, not with a large exchange port.

Schools and resellers make restoration order consequential

An outage is not experienced evenly. A residential customer may lose entertainment and remote work. A reseller can lose an entire neighbourhood behind one wholesale demarcation. A school can lose classroom access, administration, testing, security systems and communications at the same time. When many endpoints sit behind one regional junction, the order in which a field team restores spans becomes a public-service decision as well as an engineering one.

The school payment records supply no service-level terms, so the analysis cannot claim that MUNDO1TELECOM owes a particular restoration time or protected path. They do show enough institutional scale to ask how critical sites are classified. Are schools grouped behind shared aggregation points? Which have a second carrier or fixed-wireless backup? Does a storm ticket from a school receive the same priority as a reseller whose single hand-off serves hundreds of homes? Is restoration ordered by customer count, safety, contractual class, accessibility or the number of downstream nodes recovered per splice?

Field labour determines how quickly a technically repairable fault becomes service again. The website promises continuous technical support, but support-desk availability is not crew capacity. No public document gives the number of fibre splicers, climbing teams, vehicles, optical time-domain reflectometers, fusion splicers, generators or kilometres of spare cable. It does not identify crew bases outside Santiago, on-call arrangements, contractor dependence or the ability to reach a break from both sides when a road is flooded or blocked.

Geographic staging is especially important for a footprint that stretches across several provinces. A single spare stock in Santiago may be efficient in ordinary weeks but slow when debris blocks the approach to Jarabacoa or flooding interrupts another corridor. Pre-positioned closures, compatible cable, optics, fuel and safety equipment reduce travel dependence, provided each depot is inventoried and accessible. Cross-trained local contractors can add reach, but only if call-out terms, testing procedures and authority to enter third-party infrastructure are settled before the emergency.

The reseller structure can complicate fault isolation. An end user's Wi-Fi problem may stop at the reseller; an optical feeder fault crosses into MUNDO1TELECOM; an upstream failure may cross again into AS269931 or a facility operator. Each boundary can add a ticket hand-off and a dispute over whether light levels, Ethernet service, routing or customer equipment caused the failure. Shared monitoring, named escalation contacts and pre-agreed demarcation tests reduce that delay. During a widespread event, automated prioritisation should not allow the largest commercial account to obscure a fault that disconnects many public sites.

A credible restoration record would publish anonymised incident classes rather than customer details. It would show detection time, dispatch time, travel delay, repair start, repair completion and traffic restoration at the fiftieth and ninety-fifth percentiles. It would separate fibre cuts, pole failures, power failures, equipment faults and upstream loss. It would record how many sites were behind each event and whether an alternate path carried them.

It would also report exercises conducted before hurricane season: generator starts under load, failover to the second carrier, restoration from spare cable and communication between operator, resellers, schools, power companies and municipal authorities.

MUNDO1TELECOM can make the resilience claim auditable

The public record supports a real regional operator, not a paper-only name. MUNDO1TELECOM has a long-lived telephony concession, an expanded broadband footprint, active retail pages, public-sector service payments, reseller agreements, an active ASN, two well-visible IPv6 routes, a declared Santiago facility and an operational 100 Gbps exchange entry. Those facts give the company economic and infrastructural significance. They also make the remaining opacity more consequential: many customers may depend on paths whose physical boundaries cannot be tested from outside.

The strongest positive evidence is local interconnection. A large port at a Santiago exchange can keep local traffic local, reduce dependency on distant transit for participating networks and give the operator room to grow. CEMO's declared power and colocation capabilities add a plausible professional facility layer. The strongest negative evidence is not a documented failure; none is claimed here. It is the mismatch between categorical redundancy marketing and the absence of route, capacity and recovery disclosures needed to verify the claim.

The company could close that gap with six publications. First, a generalised as-built map should show every core and aggregation node, active span, medium, owner and hand-off, while obscuring street-level security details where necessary. Second, a shared-risk register should identify common ducts, poles, bridges, chambers, building entrances, power substations and carrier segments. Third, an upstream matrix should name provider, facility, port speed, usable commit, routing role and physical local-loop separation. Fourth, a power table should show battery and generator endurance under measured load.

Fifth, a capacity table should show normal and single-failure busy-hour utilisation. Sixth, a restoration report should publish incident categories, downstream impact, spare use and repair-time distributions.

Evidence should be tested, not merely described. A witnessed exercise could withdraw the primary upstream, isolate one exchange router, cut commercial power at an aggregation site and simulate loss of a named fibre span. The results should show which prefixes remain reachable, which schools and reseller hand-offs stay online, what throughput the surviving path sustains and whether alarms remain visible. A second exercise should assume the main road is impassable and demonstrate crew access, spare stock and restoration from the opposite side. Failed tests would be useful if they led to specific remediation dates.

The final network evidence grade is Medium. Legal authority, present-day service, institutional customers, IPv6 origination, a visible upstream relationship, an exchange port and a facility are all supported by public records. The grade cannot be Strong because the as-built access and backhaul routes, coastal carrier exits, ownership boundaries, independent power, protected usable capacity and measured recovery performance are not public. It is not Weak because the routing and facility evidence is unusually concrete for a regional provider.

For customers, the practical procurement question is now precise. Do not ask only whether MUNDO1TELECOM has a ring or a 100 Gbps port. Ask which enclosure, pole line, duct, bridge, building entrance, carrier local loop and power source the primary and alternate paths share; ask how much capacity remains when the largest common component fails; and ask for the last exercise proving that a crew and the traffic can both take the other way around. That is the difference between two lines on a diagram and a service that survives the squall.