Summary
- U2RED’s own service descriptions reveal a mixed physical network: line-of-sight radio customers depend on elevated shared nodes, while urban fibre customers depend on pole-mounted distribution boxes with limited ports.
- Public routing observations show AS273894 actively announcing three IPv4 and two IPv6 prefixes with one visible neighbouring network, but they do not establish independent upstream circuits, distinct fibre paths, backup-power endurance or measured customer throughput.
- The decisive rural-resilience metric is therefore the recovery clock: how long power, batteries, backhaul, roads, technicians and spare equipment can sustain or restore each affected site, evidence that U2RED’s 99% and 99.5% availability statements do not yet disclose.
The repair clock starts before the technician reaches Guamalito
A repair trip from El Carmen towards a rural relay serving the country around Guamalito is a clock, not a speed test. No public incident record establishes that this exact journey occurred, so it should be read as the practical test implied by U2RED’s geography rather than as a reconstructed outage. The clock could begin with an alarm at an elevated radio node, a customer call from a village, or a loss of commercial power. Before a technician touches a radio, the operator has to identify whether the fault sits in the customer equipment, the shared access node, its power supply, a backhaul hop, a fibre segment or the upstream connection.
The next variable is access: whether the road is passable, whether travel is safe, whether the right spare is in the vehicle and whether a second person or climbing support is required.
That sequence matters because U2RED does not present itself as a dense urban fibre carrier. Its current public site markets both fibre and wireless connections, including service for rural and remote areas, and promises round-the-clock support. Its company history says the business began as a family initiative in El Carmen intended to reach Guamalito and Catatumbo villages. It also claims more than 3,500 satisfied customers, presence in more than 13 municipalities and 99% service availability. Those statements describe commercial reach and ambition. They do not say how many powered radio sites exist, where spares are held, how many simultaneous field failures can be handled, or how long a site continues operating after the grid disappears.
The operating environment makes those omissions material. Norte de Santander’s government said in February 2026 that violence in El Carmen had affected mobility, commerce and education, while describing official measures to restore security and access; the report was not about U2RED and does not prove that any of its crews were delayed, but it shows why field mobility cannot be assumed. A separate 2026 account of new Catatumbo road projects describes work on the Tibú–El Tarra–Convención axis, an initial 7.8 kilometres on the La Mata–Convención corridor and community-road agreements that include El Carmen and Hacarí. Investment is good news, but the need for continuing intervention is itself evidence that travel time across the footprint is not a constant.
This is the central distinction in assessing a rural internet provider. An access speed is a rate available while the chain is intact. Resilience is the probability that the chain remains intact, plus the time required to repair it when it does not. A 200 Mbps residential plan or a 10 Gbps enterprise offer can be commercially meaningful without telling a reader anything about battery endurance at a ridge site. Conversely, a small provider may recover quickly through local knowledge and close customer relationships even without the engineering depth of a national carrier.
U2RED’s family and local origin could be an advantage in dispatch and terrain familiarity. Yet public evidence has to show that advantage in service-class measurements, fault distributions and recovery percentiles before it can support a resilience conclusion.
The repair clock therefore begins with several unanswered timers. How long does each radio node operate without grid power? Is there automatic switchover, and is fuel required? Which sites share a backhaul path? How many technicians are on call in the west of Norte de Santander versus Villa del Rosario? What portion of interruptions is cleared remotely, and what portion waits for a road trip? Until those timers are disclosed, U2RED’s rural claim is credible as coverage but unproven as sustained availability.
U2RED’s legal footprint is clearer than its physical boundary
The legal identity behind the network is unusually traceable for a small regional provider. The LACNIC registration for AS273894 names U2RED S.A.S as the registrant, gives an El Carmen address and records registration on 1 August 2024. It also assigns administrative, technical and abuse roles to Álvaro Javier Uribe. That is current evidence that the company controls the public routing identity and that a named contact carries operational responsibility. It is not evidence that the contact owns the company, that the company owns every tower or fibre strand it uses, or that the autonomous system encompasses every retail circuit sold under the U2RED name.
Chamber records supply a longer, if incomplete, corporate trail. An Ocaña Chamber of Commerce February 2021 bulletin lists U2RED under tax identifier 901393892-7 and registration 36386, with a 2021 renewal and reported assets of COP 50,726,213. That asset figure is historical accounting information, not a present valuation and not a measure of network investment. A March 2023 chamber bulletin again records renewal under registration 36386 in El Carmen. A July 2024 bulletin records a change in a secondary economic activity from engineering consultancy to road freight transport. The change may reflect an ancillary business activity; it does not by itself describe network construction, vehicle ownership or field logistics.
There is also evidence of a continuing local fiscal presence. El Carmen’s 2024 management report includes U2RED S.A.S and identifier 901393892 in municipal industry-and-commerce tax entries for 2023 and 2024. This corroborates operation in the municipality but does not reveal revenue, subscriber mix or profitability. A current company-information rendering describes U2RED as active, located in El Carmen and classified as a microenterprise, with telecommunications among several registered activities. Because that page is a secondary presentation of registry data, a fresh official certificate would be the stronger authority for current legal representatives, capital, shareholders and establishments.
The public origin story contains a smaller but useful conflict. U2RED’s website says it was born in 2013; its LinkedIn company page says 2016. Neither date should be silently promoted into a precise incorporation date. One may refer to the first family service, another to a later commercial phase, and the S.A.S may have been incorporated later still. The surviving public records establish continuity from at least 2021 and a routing registration in 2024, but the complete incorporation and ownership history would require the company’s current certificate and historical filings.
This boundary discipline matters for infrastructure analysis. U2RED is clearly the operator associated with AS273894 and the retail brand. It may own, lease, share or obtain access to different parts of the physical chain: poles, towers, rooftop positions, ducts, fibre, transport and upstream capacity. Nothing in the public corporate material identifies a parent company, a controlling shareholder, a tower affiliate or a wholesale network partner. Nor does it show that the legal company owns the two addresses published for customer contact.
The conservative conclusion is that U2RED S.A.S operates the service and routing identity; the ownership and contractual boundary around each physical asset remains undisclosed.
That distinction protects against two opposite mistakes. It would be wrong to treat every mast visible in a company photograph as a U2RED-owned tower. It would also be wrong to assume a microenterprise has no durable assets merely because public accounts are sparse. What is needed is an asset schedule or a sufficiently detailed network statement that separates owned, leased and shared sites; company-built and wholesale fibre; access and transport; customer-facing equipment and core routing. Until then, legal continuity is stronger than physical-asset attribution.
A thirteen-municipality promise spans two different access systems
U2RED’s footprint is best read as a set of named service markets rather than a continuous coverage polygon. The company’s history page lists ten places in Norte de Santander—El Carmen, Convención, Guamalito, Aspasicas, El Tarra, Ocaña, San Pablo, Hacarí, Villa del Rosario and Ragonvalia—and three in Cesar—Río de Oro, Ayacucho and La Mata. Some are municipalities, some are population centres or localities, so the phrase “more than 13 municipalities” should not be converted into a jurisdiction count without checking each name.
The list nonetheless establishes an operating arc from Catatumbo and western Norte de Santander towards the Cúcuta metropolitan edge, with a southern-Cesar extension.
The company’s contact page publishes addresses in Villa del Rosario and Guamalito. Those are useful service-presence anchors: one lies beside the Cúcuta metropolitan market and the other near U2RED’s stated origin area. They are not identified as network operation centres, core sites, warehouses or points of presence. A customer office may contain network equipment, but its street address cannot be treated as a backbone node without separate evidence.
U2RED also embeds a public coverage map. It is a commercial orientation aid at regional zoom. It does not publish surveyed radio contours, fibre polylines, tower coordinates, route ownership, signal thresholds or dates for individual segments. A pin or shaded area on such a map can support the proposition that a market is offered service; it cannot establish that every home is reachable, that a radio path has line of sight, or that two markets are connected over physically separate transport.
The mixed access design explains why a single footprint label conceals two different systems. In a town centre, fibre can run along streets and terminate at distribution boxes before a customer drop. Outside that urban fabric, fixed wireless can bridge distance between an elevated node and customer premises that can see it. Both may ultimately feed the same transport and upstream connection, but their local failure modes differ. Fibre is exposed to pole damage, cuts, connector faults, capacity at distribution points and rights of access.
Radio is exposed to power at the shared node, alignment, interference, vegetation, weather, equipment failure and line-of-sight limitations. One customer may experience a local drop failure while another loses service because a shared node or backhaul hop fails.
Regional road evidence gives the service list physical meaning without drawing a false route. El Carmen’s 2024 territorial implementation report records public work and studies associated with tertiary roads. It shows that rural access is an active development concern in the home municipality. It does not locate U2RED infrastructure. The route from a depot to a relay may use primary, secondary and tertiary roads, and the final approach may be on private or unpaved land. A regional road project can reduce average travel time while leaving the last kilometres to a site unchanged.
The footprint is therefore commercially plausible but physically unresolved. The most useful missing disclosure would not be a more colourful map. It would be a dated service-area account that separates active fibre zones, radio sectors and planned expansion; identifies the number of active shared sites by municipality; and gives coarse transport corridors without exposing security-sensitive coordinates. That would allow readers to understand where U2RED has installed service, where it can sell only after a line-of-sight survey, and where an address is simply near a named market.
Until such a disclosure exists, “13+” should be treated as a presence claim, not as a measure of network density. Two providers can name the same thirteen places while having radically different operating surfaces. One may have deep fibre coverage and several independently fed sites in each town; another may rely on a small number of long wireless chains. U2RED’s own technical description provides enough detail to show that its surface is real and mixed, but not enough to calculate how much of the advertised geography remains usable after any one site, road or transport path fails.
Radio coverage is a chain of sightlines, powered nodes and backhaul
U2RED’s most revealing infrastructure document is its service-connection description. For radio customers, it says the company uses point-to-multipoint links from users to nodes in strategic locations, including sites above 1,600 metres above sea level. Installation occurs only when customer equipment has direct line of sight to one of those nodes. That language establishes a shared fixed-wireless access layer, elevated siting and a pre-installation visibility test. It also exposes the chain that availability depends on.
At the customer end, a receiver must have power, remain aligned and preserve a clear enough path to the serving node. At the shared site, sector radios, switches, routers and backhaul equipment need continuous power and environmental protection. The site must then reach another node, a fibre handoff or a core location. A high site can improve coverage, but elevation does not create independence. If several sectors and a backhaul radio share one cabinet, battery bank, mast or access road, a single fault can affect many customers.
If two radio paths leave the site but converge on the same powered intermediate node, their visible separation may not survive that node’s failure.
The published description does not state how many nodes are active, their coverage sectors, power source, battery chemistry, designed autonomy, generator availability, refuelling plan or preventive-maintenance interval. It also does not identify licensed versus unlicensed spectrum, channel widths, contention ratios or interference measurements. Those omissions do not imply poor engineering. They mean the public record cannot move from “this architecture exists” to “this architecture remains available for a stated duration under loss of grid power.”
Historical energy evidence explains why the distinction deserves attention. A UPME electricity-coverage map for El Carmen reported full coverage for the municipal centre but 70.3% for the dispersed zone in its survey baseline. The map is dated and describes households, not telecommunications sites; it cannot establish the present supply at any U2RED node. A broader UPME Norte de Santander rural energisation report estimated that 70.14% of surveyed rural homes were connected to the public grid, 1.96% used generators and 27.37% had no electricity service at the time. Again, this is historical regional context, not a reading from a tower meter. It shows why a rural network cannot treat grid availability as uniform across its territory.
Power resilience has several layers. A site can have grid service but still experience interruptions. A battery can bridge short failures but lose capacity with age and temperature. A generator can extend runtime but requires fuel, testing and access. Solar can reduce fuel dependence but adds storage and charge-control constraints. Remote monitoring can shorten diagnosis but cannot replace a failed power supply.
The relevant public measure is not whether “backup power” exists in the abstract; it is how long critical equipment remains powered at each site class under a defined load, how often that endurance is tested, and whether backhaul and upstream endpoints have equal or longer endurance.
Line of sight brings a second set of dependencies. A successful installation survey proves a usable path at that moment. It does not disclose fade margin, seasonal vegetation growth, interference or whether the backhaul uses a similarly exposed link. Nor does “point-to-multipoint” imply that a customer can switch automatically to another node. The customer antenna may be aimed at one sector, and alternate coverage may require a visit or realignment. Redundancy has to be specified at each layer: customer-to-sector, sector-to-aggregation, aggregation-to-core, core-to-upstream and power at every shared point.
U2RED’s radio access is therefore more than a marketing label. Its own text provides a credible physical mechanism for reaching difficult terrain. But the same mechanism concentrates risk in elevated shared nodes whose power, backhaul and maintenance arrangements are not public. A resilience case would need a node inventory by site class, tested power autonomy, backhaul path count, failover behaviour and outage history. Without those, the best-supported conclusion is that U2RED operates a real rural radio chain and that the chain’s duration under stress is unknown.
Urban fibre still runs through poles, boxes and scarce ports
Fibre sounds more self-contained than radio, but U2RED’s description makes its street-level dependencies unusually visible. The company says fibre is broadly available in urban sectors, not across all rural territory. It describes distribution boxes placed in the street near customers, with limited slots, commonly installed on electricity poles, and a single fibre strand running from the box to the customer. That account supports a genuine last-mile fibre architecture while warning against treating an urban coverage label as unlimited capacity.
The first constraint is physical reach. A town can be “covered” while an individual street lacks a nearby box, an available port or a permitted pole route. The second is shared concentration. A distribution box aggregates several customer drops, and an upstream splitter or feeder may aggregate several boxes. Damage to a pole, closure of a box, a feeder cut or contamination at a connector can affect more than one household. The third is restoration logistics. A fibre repair may require locating the fault, reaching both sides, obtaining pole access, preparing cable, splicing strands, testing optical levels and confirming service.
A local team with the right equipment can restore quickly; a team short of splicing capacity or replacement cable can be delayed even when the fault is geographically close.
Pole mounting also creates a relationship with electricity infrastructure without proving dependence on grid current. Passive fibre in a street box does not need power merely because the box is on an electricity pole. Active equipment elsewhere does. The important point is shared physical exposure and access: the same storm, vehicle strike, vegetation event, civil work or pole replacement can disturb multiple utilities. U2RED does not publish whether its feeder routes share one side of a road, cross rivers at common structures, follow a single corridor between towns or use another carrier’s transport.
It is therefore impossible to infer path diversity from the existence of fibre alone.
The scarcity of ports has an economic meaning. A box with no free slot may require augmentation before another customer can be connected. That makes the advertised footprint wider than immediately usable installed access capacity at a specific address. It also means the customer count cannot be translated into network utilisation without knowing how many ports, split ratios and feeder strands are installed, active, reserved or defective. A network can have spare optical bandwidth but no physical port at the closest box; it can also have open ports while its shared upstream capacity is busy at peak hours.
The published office locations add no route certainty. Villa del Rosario and Guamalito are customer-presence anchors, not declared fibre hubs. Nor can the commercial map settle the matter: it offers geographic orientation without precise linework or asset status. The public evidence supports urban fibre within the named service area, but it does not show an inter-town fibre backbone owned by U2RED. Some transport may be leased, delivered over radio or bought at a handoff. Any of those arrangements can be sound; each has different repair authority and escalation time.
The strongest next disclosure would divide capacity into three physical layers. At access, U2RED could report active and spare distribution ports by service zone, without exposing customer addresses. At feeder level, it could report installed versus lit strands and whether major zones have independent paths. At transport level, it could state whether each market reaches the core over owned fibre, leased fibre or radio, and whether “independent” paths avoid the same poles, ducts, bridges, powered cabinets and upstream handoff. Those facts would turn a fibre label into a resilience account.
Until then, urban fibre should be judged like the radio network: as a chain of shared dependencies. It likely offers higher and more stable access rates where installed, but its continuity still depends on physical route, aggregation equipment, power at active sites, repair labour and upstream transport. U2RED’s candid reference to limited ports is valuable because it distinguishes service-area ambition from address-level readiness. It also makes clear that the company’s installed footprint, sellable inventory and usable peak capacity are three different quantities.
The 10 Gbps ceiling is an offer, not measured usable capacity
U2RED publishes a broad commercial range. Residential service is advertised over fibre or radio, depending on coverage, with plans extending to 200 Mbps. Enterprise channels are described as dedicated and symmetric from 30 Mbps to 10 Gbps, with a 99% service-level commitment. Point-to-point links are offered as dedicated, redundant and monitored. These are product definitions and design promises. They are not measurements of installed backhaul, current traffic, delivered peak-hour speed or the amount of capacity that survives a failure.
The distinction can be expressed as a stack. “Designed” capacity is what equipment and interfaces could carry under specified conditions. “Installed” capacity is what has been physically deployed. “Lit” capacity is provisioned and connected. “Powered” capacity is available while required sites have energy. “Operational” capacity is working and routed. “Usable” capacity is what customers can actually obtain after overhead, contention, congestion, policy and any failed component. “Sold” capacity is the aggregate of commercial commitments, which may exceed simultaneous demand where service is shared.
U2RED discloses offer rates but none of the other quantities in a form that can be reconciled across its footprint.
The 10 Gbps enterprise ceiling illustrates the issue. It may correspond to a capable interface, a bespoke circuit available in selected locations or a wholesale handoff that can be ordered. It does not mean every town has 10 Gbps of installed transport, every radio node can deliver it, or that 10 Gbps remains after a route failure. A dedicated service also needs a defined demarcation, committed information rate, latency and loss objectives, restoration terms and exclusions. The public page gives a strong starting point for a sales conversation, not enough detail for an independent capacity audit.
U2RED’s regulatory-information page publishes four performance statements: average download at 95% of the contracted plan, 99.5% service availability, average repair time below 24 hours and unplanned interruptions below 0.5% per month. These figures appear more precise than the 99% availability shown on the company history page and enterprise offer. They may refer to different service classes, rounding or periods. The page does not disclose the measurement window, number of lines, weighting, denominator, exclusion rules, geographic split, raw observations or external verification. The figures should therefore be attributed to U2RED and not presented as independently measured results.
The company also provides a customer speed-test page. A test facility can help an individual diagnose a connection, but it does not publish an aggregate distribution for U2RED customers. One test is sensitive to Wi-Fi, device limits, local traffic, test-server location and time of day. A credible capacity account would show distributions by access type and plan, including median and lower-percentile performance at busy hour, rather than a single average across an undefined population.
Address resources offer a different type of capacity evidence. Public registration pages for 107.149.165.0 and 64.204.90.0 show address ranges associated with current routing evidence and include U2RED-related registration information. IP space is not throughput. A /24 provides 256 IPv4 addresses before reservation and allocation choices, but customers may share addresses, use IPv6 or receive private addressing. Address holdings cannot establish subscriber count, interface speed, oversubscription or available transit.
The right interpretation is neither scepticism for its own sake nor acceptance of the largest number on the page. U2RED has made concrete offers and published operational targets, which is more informative than a vague claim of “high speed.” The unfilled gap lies between commercial ceiling and observed delivery. To close it, the company would need dated installed-capacity figures by major aggregation area; peak and 95th-percentile traffic; sold committed capacity; route-failure capacity; and customer-performance distributions by fibre and radio. The availability figures would need a common definition and service-class split.
Until then, 10 Gbps is a maximum product offer, 99% is a promise or rounded claim, and 99.5% is a self-reported metric whose population remains unknown.
One visible upstream leaves physical diversity unproven
U2RED’s public routing identity is active and recent. The RIPE Routing Information Service AS overview identifies the holder as U2RED S.A.S and shows AS273894 as announced. Its announced-prefix history for the observation window ending 17 July 2026 includes three IPv4 /24s—38.224.21.0/24, 107.149.165.0/24 and 64.204.90.0/24—and two IPv6 announcements, 2803:430::/32 and 2803:430:cafe::/48. A fourth IPv4 /24, 141.11.93.0/24, appeared only at the start of the window and ended on 4 July. This is evidence of active public routing and a small but non-trivial address presence. It is not a topology diagram.
The routing-status observation records first visibility for the AS in August 2024, three currently observed IPv4 prefixes representing 768 addresses, two IPv6 prefixes and one observed neighbouring network. The neighbour view identifies that visible neighbour as AS3356. Independent derived views at bgp.tools and CIDR Report similarly show AS3356 as the visible upstream relationship. These services are based on collector visibility and inferred path position. They can miss private interconnections, backup sessions that are idle, routes hidden by aggregation or arrangements not visible from their observation points.
Even if AS3356 is the only active upstream, its name does not reveal the physical circuit. Two sessions to the same upstream could be delivered over different carriers and entrances, or two logical sessions could share one local fibre, duct, pole line, powered handoff or distant aggregation site. Conversely, a single visible path can still be operationally robust if the upstream supplies physically diverse access and rapid repair. Public BGP data cannot distinguish those cases.
Route-origin security is a separate strength. RIPEstat returned valid RPKI origin validation for 38.224.21.0/24, 107.149.165.0/24, 64.204.90.0/24, the 2803:430::/32 allocation and the 2803:430:cafe::/48 more-specific on 17 July 2026. Valid origin authorisation reduces the chance that networks enforcing RPKI policy will reject these legitimate announcements and helps defend against some route-origin mistakes or hijacks. It says nothing about bandwidth, packet loss, tower power or physical path diversity.
Cloudflare’s public routing view offers another current logical observation, while its quality page is a potential measurement surface for internet experience. Neither exposes a U2RED last-mile map, customer-by-customer samples, fibre ownership or failover design. Such pages can corroborate whether an AS is seen and whether enough traffic exists for aggregate indicators; they should not be stretched into a physical-resilience claim.
The important finding is thus narrow but consequential. U2RED has crossed the threshold from an invisible retail brand to a network with its own public AS, active IPv4 and IPv6 announcements, and valid route-origin authorisations. That improves operator identity and routing control. At the same time, all readily visible collector evidence converges on one adjacent AS. Until U2RED or its supplier documents separate circuits, handoff sites and local approaches, the safe conclusion is “one visible upstream, physical diversity unknown.”
An adequate redundancy statement would identify whether there are two active upstream sessions, whether a backup is normally hidden, the cities or facilities of handoff at a coarse level, the local-access providers, and whether the paths avoid common poles, ducts, bridges, power domains and aggregation equipment. It would also state how much capacity remains after the loss of the largest path. Without that information, “redundant” point-to-point products cannot be generalised to the public internet edge or to every access site.
Power loss and road delay can turn redundancy into a shared failure
The failure path that matters most for U2RED is not one dramatic event but a sequence of dependencies becoming common. A customer may have a fibre drop or a radio link that is healthy. The serving box or node may also be healthy. Yet service can still disappear if a backhaul hop loses power, if a feeder fibre is cut, if the single visible upstream path is interrupted, or if a powered aggregation location becomes unreachable. Redundancy works only where alternate components avoid the original fault and remain available long enough to carry traffic.
The electricity context is not hypothetical at regional level, although no public document ties a specific interruption to a U2RED site. A Superservicios evaluation of Centrales Eléctricas del Norte de Santander describes the regional electricity provider and records how security conditions in Catatumbo affected projects and operational activity during its review period. The document is dated and cannot establish current grid performance, much less the condition of a particular telecom battery. It does show that electricity infrastructure, field access and security can interact in the same territory.
Road and security events can also affect restoration independently of the original fault. In 2024, Invías condemned an attack on a contractor working on the Transversal del Catatumbo, reporting a burned vehicle and threats to personnel on the Ocaña–Convención corridor. This was not a telecommunications incident and there is no evidence U2RED was involved. Its relevance is bounded: it demonstrates that travel and contractor safety on a corridor associated with the operator’s named service area can face constraints beyond ordinary drive time.
Consider how a shared failure can defeat apparently separate access links. Two customer radio links may point to different sectors on the same tower but share its battery and backhaul. Two towers may use separate radios but converge on one aggregation site. A fibre customer and a radio customer may look diverse at the last mile yet share the same upstream handoff. Two enterprise circuits described as redundant may terminate on different interfaces while following the same pole route out of town. A generator at one site may outlast its neighbour, only for the chain to fail at the neighbour first.
The least-enduring shared point sets the practical outage duration.
The reverse is also possible. A small regional provider can design meaningful independence without publishing it. A radio path can protect a fibre route; solar and batteries can keep a high site alive; a second transport provider can avoid the primary corridor; local technicians can carry standard spares; and remote configuration can restore service without travel. U2RED’s point-to-point offer explicitly uses the words “redundant” and “high availability,” suggesting that the company understands demand for protected links.
What is missing is the scope of that protection: whether it applies only to a custom enterprise circuit, whether access sites have alternate backhaul, whether the internet edge has another route, and whether all alternate elements are powered independently.
Recovery evidence should therefore be scenario-based. For grid loss, disclose tested autonomy at customer-shared sites and the refuelling threshold. For fibre cut, disclose the restoration objective and the portion of traffic that can move to radio. For radio failure, state whether a spare is held locally and whether customers can be repointed. For upstream loss, disclose the alternate path and surviving capacity. For blocked access, show which repairs can be completed remotely and which sites have enough endurance to wait safely.
U2RED’s 99.5% figure corresponds to roughly 3 hours and 39 minutes of unavailable time in a 30-day month if applied continuously to one service and calculated without exclusions. That arithmetic is illustrative, not a claim about how the company calculates its metric. A monthly average can conceal very different experiences: many short interruptions, one long rural outage, or high performance in a large urban group offsetting worse results at a small remote node. The company’s separate statement of less than 0.5% unplanned interruption appears mathematically related, but the denominator and service population are not published.
The resilience question is not whether the percentages look good. It is whether they include the hard-to-reach sites whose repair clock is most exposed to power and roads.
Recovery is a labour-and-spares problem
Networks recover through people making bounded decisions with the equipment available to them. U2RED advertises specialised and 24-hour support, and its local origin may shorten the distance between a report and someone who understands the terrain. But public customer-service availability is not the same as field-repair capacity. A call can be answered while the only trained fibre splicer is committed elsewhere, a tower visit awaits safe access, or a replacement power supply is stored in another town.
The company’s self-reported average repair time below 24 hours is useful because it acknowledges restoration as a measurable outcome. Yet an average alone is particularly weak for rural resilience. Ten faults repaired in two hours and one fault repaired in several days can still produce an acceptable-looking mean. Affected users care about the tail: the 90th or 95th percentile, the longest events, and the difference between urban fibre, rural radio, upstream and power faults.
They also need to know whether the timer begins when monitoring detects a failure, when the first customer reports it, when a ticket is opened or when a technician is dispatched.
U2RED’s PQR page gives customers formal categories for service quality, lack of availability, intermittency and compensation, among other issues. That is evidence of an established complaint channel, not evidence that any category is frequent. The channel can become a valuable operational record if results are aggregated without exposing customers: complaints per thousand lines, acknowledgement time, restoration time, recurrence within thirty days and compensation outcomes by access type. Public reporting of even a small set of those measures would make the availability claim more interpretable.
Colombia’s CRC has recognised that small fixed providers serving rural, remote and difficult-access areas face distinct operating conditions. Resolution 6755 of 2022 established differentiated requirements for providers with fewer than 30,000 residential fixed-internet accesses that serve qualifying areas. U2RED’s claim of 3,500 satisfied customers is not an official subscriber count, and the public evidence here does not establish its eligibility or regulatory treatment. The resolution is relevant because it identifies the class of constraints surrounding small rural providers while preserving customer-protection and reporting obligations.
Scarce labour affects both cost and resilience. A larger spare pool, more field crews and more geographically distributed depots improve recovery but raise fixed costs across a relatively small base. Stocking every radio, optical transceiver, power unit and fibre component at every town is inefficient; holding too little creates long waits. Training technicians across radio, fibre, IP routing, electrical safety and tower work is expensive, while relying on specialists increases scheduling risk. The operator’s optimal inventory depends on observed failure rates and travel times, data that are not public.
The most informative labour disclosure would not require naming employees. U2RED could publish the number of active field teams by broad zone, normal and emergency coverage hours, location of spare pools at municipality level, classes of repair completed in-house, and arrangements for tower or civil work. It could also disclose restoration percentiles by fault class and the share of incidents cleared remotely. Those measures would show whether local knowledge translates into faster recovery and where one simultaneous event could exhaust capacity.
There is a wider economic point. Rural service pricing must support not only bandwidth but readiness: batteries replaced before failure, vehicles maintained, technicians retained, spares held and routes inspected. Customers may compare headline Mbps across providers without seeing that readiness cost. U2RED’s strongest defensible differentiation could be the speed and competence of local restoration rather than the top advertised rate. But to make that case, it must turn “support 24/7” from an availability of contact into evidence about diagnosis, dispatch and restoration under the failures most likely in its actual footprint.
Households carry the impact when the repair clock lengthens
The affected population cannot be reconstructed from U2RED’s customer claim alone. “3,500+ satisfied customers” is self-reported, may refer to current or cumulative relationships, and does not distinguish households, businesses, active lines or users behind a connection. No public customer roster establishes that U2RED serves particular schools, clinics, government offices or emergency services, so none should be attributed to the company. The supported group is narrower: households and businesses in the named service areas that buy its fibre, radio, enterprise or point-to-point products.
The regional stakes are nevertheless measurable. DANE’s 2024 quality-of-life results for PDET municipalities found that 44.0% of households had internet access, compared with 51.8% in municipal centres and 37.9% in populated and dispersed rural areas. Among connected rural households, 29.5% reported fixed internet, 82.0% mobile internet and 11.4% both fixed and mobile. These are aggregate results across PDET municipalities, not a U2RED customer survey and not a Catatumbo-only estimate. They show that rural households are less likely to be connected and often rely on mobile service, sometimes alongside fixed access.
That combination changes the impact of a fixed-network outage. A household with reliable mobile coverage and affordable data may shift essential tasks to a phone. A household in terrain with weak mobile signal may have no effective substitute. A business can lose card payments, messaging, cloud access and customer communications even if voice service remains. Students and remote workers can postpone some activity but not always a live class or meeting. U2RED itself markets connectivity for work, study and entertainment; those are plausible uses, but the severity of an outage varies by user, location, time and backup options.
Shared-site failures can create geographically concentrated harm. A fibre-box fault may affect one street or cluster. A sector-radio fault may affect customers spread across a visible arc. A backhaul or power failure at an elevated node may affect several sectors at once. An upstream loss may reach much of the footprint if no alternate path exists. Without node-to-customer counts and outage records, the number affected in each scenario is unknown. The company’s aggregate availability percentage cannot tell a rural customer whether their particular serving site performs better or worse than the average.
Duration matters as much as reach. A brief interruption may be an inconvenience. An outage that lasts through a working day changes income and access to services. A multi-day failure can force travel to a connected town, increase mobile-data spending or interrupt business operations. This is why the road and power evidence belongs in an internet assessment: it influences how long a fault persists, and that duration determines the economic burden placed on users.
Transparency can reduce that burden even before infrastructure changes. A public status notice with the affected zone, detection time, broad fault class, estimated restoration window and update cadence lets users plan. A post-incident explanation can state what failed, how service was restored and what will prevent recurrence. Aggregate quarterly reporting can preserve security and customer privacy while showing availability and repair distributions by access type. U2RED’s complaint categories already create a route for individual redress; timely area-level information would complement it.
The local operator’s potential advantage is proximity. A national carrier may have greater capital and route diversity, but a regional team may know the road, customer landmarks, seasonal conditions and site history. That advantage is valuable only if it survives simultaneous faults and if spares and authority are close enough to act. For households deciding whether to rely on U2RED as their primary connection, the practical question is not simply “Does coverage exist?” It is “When my shared node loses power or backhaul, how quickly will the company know, reach it and keep me informed?”
What would turn U2RED’s claims into a resilience case
U2RED has already supplied several pieces that many small providers leave invisible. It identifies the operating company behind AS273894. It explains that rural radio depends on direct line of sight to shared elevated nodes. It acknowledges limited ports in pole-mounted fibre distribution boxes. It publishes named service areas, commercial rates, an availability figure, an average repair claim and a customer complaint route. Public routing observations confirm that its AS is active with IPv4 and IPv6 and that its current origins are RPKI-valid.
Together, those facts support a real mixed-access operator, not a name attached to an empty website.
They do not yet support the stronger statement that the network is resilient across Catatumbo and south Cesar. The decisive evidence can be organised without exposing exact tower coordinates or commercially sensitive contracts. First, U2RED could publish a dated inventory by municipality and site class: active radio nodes, fibre service zones, aggregation points and broad transport method. The inventory should distinguish installed, operational and planned assets and state which are owned, leased or shared.
Second, it could publish power evidence by site class. The useful measures are tested battery autonomy under current load, generator or solar availability, last endurance test, fuel or maintenance arrangements, and the proportion of customers behind each power class. A statement that “sites have backup” is limited public evidence because a network chain lasts only as long as its shortest critical power domain.
Third, it could clarify route independence. For each major service zone, the operator could state the number of active backhaul paths, their medium, whether they share a mast or aggregation point, and what capacity remains after the largest single failure. At the internet edge, it could explain whether the one publicly visible adjacent AS is reached over multiple physical circuits or whether another upstream exists but is normally invisible. Coarse handoff geography and diversity criteria would be enough; exact routes need not be exposed.
Fourth, capacity reporting should separate offer from operation. U2RED could disclose installed and lit transport, busy-hour utilisation, sold committed capacity and surviving capacity under failure. Customer results should be distributed by fibre and radio plan rather than compressed into “95% of plan.” The 99% and 99.5% availability statements should be reconciled with a shared definition, measurement period, exclusions, service classes and customer weighting.
Fifth, recovery performance should reveal the tail. Median, 90th-percentile and 95th-percentile restoration times by fault class would show whether the average below 24 hours conceals rare but severe rural events. The operator could report detection-to-diagnosis, diagnosis-to-dispatch and dispatch-to-restoration separately. It could also disclose the broad distribution of teams and spare pools, plus the fraction of incidents resolved remotely. None of this requires identifying staff or vulnerable sites.
Finally, incident reporting should connect infrastructure to users. Quarterly figures could show outage minutes and affected connections by fibre, radio, transport, upstream and power cause. Significant incidents could receive brief public explanations. A customer should be able to tell whether a “redundant” enterprise link avoids the same access node, backhaul, power supply and upstream handoff, and what service remains if one part fails.
There is no evidence here of a present crisis at U2RED. The company’s public routing is active, its legal identity is visible and its service descriptions are more physically candid than average marketing copy. The uncertainty is structural: the strongest numbers come from the operator itself, while the public record lacks the denominators, site inventory, power endurance, route geometry and restoration distribution needed to test them. That is why the title’s 99.5% matters. Precision invites a precise question.
For U2RED, rural resilience begins after the plan speed has stopped answering anything useful. It begins when a shared node loses power, when a backhaul path degrades, when a fibre is cut, or when a technician cannot immediately reach the site. The outcome depends on whether alternate paths truly avoid the same failure, whether power lasts, whether the alarm is understood, whether the spare is nearby and whether the road can be travelled. U2RED has established that it can reach difficult places.
The next evidentiary step is to show, with defined and segmented results, how long those places stay connected and how quickly the company restores them when they do not.