Summary
- Sencinet publicly describes a regional estate of five teleports and four data centres, while the strongest Colombia-specific evidence identifies one Bogotá teleport and two data-centre sites; it does not map five independent recovery paths from a Colombian branch.
- AS14187 was actively announcing 43 prefixes on 17 July 2026, but route visibility, address space and observed neighbouring networks do not reveal branch access diversity, satellite reserve, gateway headroom or the physical separation of fibre paths.
- A credible failover test must begin at the customer site and follow power, local access, terminal, teleport, data-centre and operating authority end to end; the public record leaves branch runtime, spare satellite capacity and restoration responsibility unresolved.
A storm turns a terrestrial loss into a test of the branch edge
Imagine a remote Colombian operations site shortly after a heavy storm. The primary terrestrial circuit has stopped carrying traffic. The edge appliance still has power, detects loss or unacceptable performance on that path and selects a satellite alternative. A dish outside has a clear enough view of the sky; its modem and radio are energised; the chosen satellite beam can reach an available gateway; the gateway has functioning power and terrestrial backhaul; and the service beyond it can still reach the applications that the branch needs.
Only when that entire chain holds does the word “failover” describe restored service rather than a change of status on an appliance.
This is an analytical scenario, not a report of an outage at a named Sencinet customer or facility. It is nevertheless a realistic way to test the offer. Colombia faces recurrent floods and landslides as well as other natural hazards, according to the World Bank's account of national disaster and climate risk. That country-level exposure does not prove that a particular branch will fail. It does explain why a remote-access design should be assessed against disrupted roads, local electricity loss and damaged terrestrial plant, not only against a clean demonstration in normal conditions.
Sencinet markets satellite as end-to-end remote connectivity and explicitly includes backup for high-availability applications among its uses. Its satellite service page says that five regional teleports support services across C, Ku and Ka bands and operate with multiple satellites. Its Secure SD-WAN page describes continuous link monitoring, application traffic steering, packet duplication and multiple access options. Those are meaningful operating capabilities. They also sit at different layers. An edge system can select only among paths that remain usable, and a second transmission medium is independent only to the extent that it avoids the first medium's failed power, duct, building entrance, access supplier, aggregation point and remote operating dependency.
Satellite introduces its own environmental and engineering constraints. The current ITU-R P.618 recommendation provides prediction methods for propagation effects on Earth-space systems, including attenuation associated with precipitation. It does not say that a Sencinet link will fail in rain; it shows why band, elevation, fade margin, coding and site diversity belong in an availability design. “Satellite” is therefore not a single uniform backup class. A C-band path, a Ku-band path and a Ka-band path may have different link budgets, terminal sizes, licensing conditions and weather sensitivity.
Power is the common dependency easiest to overlook. A fibre handoff and a satellite modem beside it may appear diverse while sharing one branch distribution board, one small uninterruptible supply or one generator that has not been exercised under load. The CISA resilient-power guidance for critical facilities is written for a different national setting, but its engineering principle is general: independent communications services should minimise common failure and backup power must be treated as an operating system with fuel, maintenance and endurance requirements. The first question for Sencinet Colombia is consequently not “Are there five teleports?” It is “What remains alive at this branch, and what independent chain can carry the load from here?”
The Colombian company is older than the Sencinet brand
The legal entity behind this Colombian operating surface is not a newly incorporated regional brand. A recent Cámara de Comercio de Bogotá certificate identifies SENCINET LATAM COLOMBIA S.A., NIT 800255754-1, domiciled in Bogotá. It records incorporation in March 1995 as Comunicaciones Satelitales de Colombia S.A., a later COMSATCOL name, the change to BT Latam Colombia in 2008 and the change from BT Latam Colombia S.A. to Sencinet Latam Colombia S.A. in December 2020. Its registered activities include wired telecommunications, satellite telecommunications, information-technology consulting and data processing or hosting. This sequence matters because it ties the present name to a long-running Colombian company rather than treating every regional Sencinet claim as automatically attributable to a generic overseas parent.
The same chamber certificate records a control relationship in which CIH Telecommunications Americas LLC controls the Colombian company indirectly through Sencinet Inc., formerly BT Latam Inc., from 1 October 2020. A public United States regulatory ownership attachment independently describes Sencinet Inc. as directly and wholly owned by CIH Telecommunications Americas and sets out the holding-company chain above it. These documents support a corporate boundary; they do not allocate each teleport, fibre pair, satellite contract or data hall to the Colombian balance sheet.
The transaction that created the Sencinet brand supplies the regional context. BT's completion announcement says that selected Latin American domestic operations and infrastructure were sold to CIH in 2020 and would operate as Sencinet. It lists two owned fibre networks totalling 650 kilometres, 2,000 kilometres of leased fibre lines, four data centres and five teleports within the transaction. The sentence is useful because it is the clearest public provenance for the now-familiar asset counts. It is equally important for what it does not say: it does not divide the owned fibre, leased fibre, teleports or capacity by country, nor does it state that the Colombian company owns every item.
A Sencinet launch announcement distributed in 2020 added that the new business began with 500 employees in 16 countries and connected more than 25,000 locations. Those were launch-era, region-wide numbers, not a current Colombian customer count. They demonstrate the scale and integration challenge inherited from BT; they do not reveal how many Colombian branches use satellite backup today, which company signs each capacity agreement or which operating centre has final restoration authority.
That distinction protects the analysis from two opposite errors. One would be to reduce the Colombian company to a reseller merely because Sencinet operates regionally. The legal history, registered activities, local facilities and active ASN show substantive Colombian operations. The other would be to place all regional infrastructure inside the Colombian company because a regional website uses “we” and “our.” The public record supports a local operating company within an international control structure.
Asset title, service operation, wholesale access and customer accountability still have to be separated facility by facility and contract by contract.
The regional estate is large, but the Colombian boundary is narrower
Sencinet's offices page lists a Bogotá office at Calle 113 No. 7-21 in the Teleport Business Park, alongside offices in Brazil, Argentina, Mexico and the United States. An office address establishes a place of business. The word “Teleport” in the building name does not establish that an earth station is on the office roof, that the office has independent power or that it is an operating centre. Certification evidence later places an actual Colombian teleport elsewhere in Bogotá's free zone, which is why a map should not convert a corporate address into a network site.
The company's corporate description presents Sencinet as an integrator of networking and satellite services and displays relationships with satellite, network and equipment companies. A partner list is evidence of an ecosystem, not evidence that every named partner supplies live capacity to every Colombian service. It does, however, reinforce the right unit of analysis: Sencinet combines facilities, carriers, satellites, equipment and operating staff rather than relying on one transmission technology.
One Sencinet article offers a more granular regional snapshot. Its discussion of enterprise WAN connectivity says the company had 112 network points, more than 160 managed-access providers, four data centres—one in Brazil, one in Argentina and two in Colombia—and five teleports across the continent. It also says that its end-to-end service-management team had a presence in 15 countries. These are company statements without a publication date or live inventory attached, so they should not be read as an audited 2026 count. The two-Colombia data-centre statement is nevertheless consistent with named-site certificates for Nimbus and Naos.
The terrestrial claim is more specific in one place. Sencinet's networking service page describes Metro Ethernet fibre infrastructure in the metropolitan areas of Bogotá and Buenos Aires, corporate Internet options and an MPLS network across 15 countries. It does not publish a Bogotá route map, kilometre count, duct ownership schedule, access-building list or current upstream capacity. “Fibre infrastructure” can include owned plant, leased strands, capacity purchased from another operator and cross-connects within third-party facilities. The BT transaction's regional split between owned and leased fibre makes that uncertainty material.
The result is a layered Colombian footprint rather than a single network polygon. At the most securely located layer are the Bogotá office, the Zona Franca Bogotá teleport, the Nimbus data centre in that free zone and Naos in Tocancipá. At the service layer are Bogotá metro fibre, corporate Internet, MPLS, satellite, SD-WAN and cloud products. At the regional layer are five teleports, four data centres, hundreds of access relationships and the inherited fibre estate. At the customer layer are individual branch circuits and terminals, many of which can be provisioned by local suppliers.
Public evidence joins these layers commercially, but it does not reveal every physical junction between them.
That matters for a remote Colombian customer. A national contract may name Sencinet and carry a Sencinet service-management commitment while the first kilometre belongs to another carrier, the satellite capacity belongs to a satellite operator, the building belongs to a free-zone developer and the data-hall electrical plant belongs to a facility owner. None of those arrangements is inherently weak. Managed integration is the product. Resilience depends on knowing where the arrangements converge and on ensuring that somebody has both the information and authority to restore each link.
Two Colombian data centres expose the difference between operation and ownership
The most reliable Colombian facility map comes from certificates that name sites and scope. Sencinet's ISO/IEC 27001 certificate, valid within its stated certification cycle to September 2027, covers the provision, operation and support of data centres at the principal Bogotá office, Nimbus at Carrera 106 No. 15A-25 in Zona Franca Bogotá, and Naos at the Briceño-Zipaquirá road in Zona Franca Tocancipá. “Provision, operation and support” is significant evidence of an operating role. It is not identical to title over the land, building, switchgear or every rack.
The site list is corroborated and extended by Sencinet's ISO 9001 certificate. That document names Telepuerto ZFB at Carrera 106 No. 15A-25, Manzana 8, Bodega 55; IDC Nimbus at the same wider free-zone address but Manzana 4, Lote 38; and IDC Naos in Tocancipá. This is the strongest public basis for saying that a Colombian teleport exists in Zona Franca Bogotá and is distinct from the Nimbus data-centre site. It does not say that the teleport and Nimbus have separate utility feeds, fuel systems, carrier entrances or flood zones.
Independent certification records illuminate the facility boundary. The TIA listing for Sencinet Naos Sala 1 records an active ANSI/TIA-942-C Constructed Facility certification at Rating 4, issued in June 2026 and expiring in June 2029. The current Nimbus TIA certificate names Desarrolladora de Zonas Francas S.A. and the “SENCINET Nimbus” second-floor facility, records Rating 3, and runs from December 2025 to December 2028. The distinction is not cosmetic: Nimbus's certificate names the free-zone developer as the certified-facility company, while Sencinet's ISO certificate names Sencinet's operating and support scope at the site.
The Uptime Institute's Colombia award list similarly associates DATA CENTER BT NAOS with Cotel S.A. and records Tier IV design and constructed-facility awards. It associates the BT Nimbus floor with Desarrolladora de Zonas Francas and lists a Tier III design award, with management-and-operations recognition for the wider DC38 floors. These entries support resilient facility design at specific scopes. They do not certify the remote branch, the metro path into the campus, the satellite gateway chain, the customer application or the amount of spare compute and network capacity at the moment of a regional failure.
Cotel's Naos project account makes the ownership line unusually explicit: it describes Naos as owned by Cotel and serving Sencinet and C&W. The contractor says the installation includes 3 MVA and 2.5 MVA transformers, eighteen 537 kW generators in a 2N arrangement, and 650 kW of UPS capacity expandable to 1,300 kW. It also says capacity filled to 100 percent in four years. These are valuable engineering disclosures, but they come from a project case study. “Capacity” is not defined as racks, sellable power, occupied space or contracted load, and the account does not date the point at which the facility was full. The numbers should not be converted into current Sencinet headroom.
A Colombian government procurement document supplies a customer-side view. The DAFP study describing services at BT Naos places the facility in Tocancipá and describes redundant UPS systems and at least 24 hours of electrical autonomy for the service then provisioned. This is better evidence than a generic availability slogan, but its scope is still the documented service and date. It cannot establish present fuel stocks, generator maintenance, every tenant's power path or branch-to-Naos connectivity.
Finally, the TIA-942 standard description states that the standard addresses data-centre telecommunications, power, cooling, architecture, fire protection, safety and physical security. That clarifies what a facility rating can tell a buyer. It also clarifies the limit: a highly rated data hall can sit behind a customer access circuit, campus entrance or remote-site power system that has not been shown to be equally diverse. Naos and Nimbus add real resilience options, but the branch benefits only if its traffic can reach the intended site and if the surviving site has capacity to accept the displaced work.
Sencinet's hybrid-cloud service page says it operates four Latin American data centres in Argentina, Brazil and Colombia, with two Tier III and one Tier IV facility, and offers disaster recovery, backup and managed services. The page invites customers to define recovery-time and recovery-point objectives. Public evidence does not disclose a common recovery architecture for every customer. A buyer therefore needs to verify whether its secondary application, security policy, DNS, identity service and management access are actually outside the same failure domain as the primary branch path.
Five teleports are a footprint count, not five independent recoveries
Five teleports are more than a marketing abstraction. An earth station concentrates antennas, radio-frequency equipment, baseband systems, satellite capacity and terrestrial backhaul, and it gives a managed provider a place from which to operate remote terminals. Sencinet states that its five teleports use multiple satellites and frequency bands and can host multiple satellite providers. That diversity can reduce dependence on one spacecraft, one transponder or one band.
But the public statement does not name the five current sites, map their beams, identify which can serve a given Colombian terminal, or state how much failover traffic each can accept.
Colombia-specific certification identifies Telepuerto ZFB, but the remaining map is incomplete. A historical World Teleport Association directory entry for BT, last updated in 2013, listed La Calera and Zona Franca teleports in the Bogotá metropolitan area among a much larger collection of BT earth stations. That old entry is useful provenance for the inherited Colombian satellite estate. It cannot prove that both sites were included in the five assets sold in 2020, that both remain active, or that Sencinet currently routes a particular customer's traffic through either one. The fact that the old BT directory listed more than five Latin American earth stations is itself a warning against matching today's count to yesterday's names without a current asset schedule.
The branch-to-teleport relationship is also constrained by physics and service design. A remote antenna may be commissioned for one orbital position, band and hub platform. Shifting traffic to another teleport may require compatible hub equipment, network configuration, satellite coverage and capacity rights; it is not equivalent to changing an Internet route. Multiple satellites at one teleport protect against some failures but not the site's utility supply, antenna field access, shared baseband room or common terrestrial exit.
Two teleports in one metro may protect against equipment faults while remaining exposed to a regional power, carrier or weather event.
Even when alternate gateways are technically available, reserve capacity is decisive. The normal load on a satellite network can fit comfortably while a mass terrestrial outage causes many branches to switch at once. The relevant number is not the aggregate spectrum that exists somewhere in the regional estate. It is the capacity contractually and technically available on the compatible beam and hub at the time of failure, after higher-priority traffic and contention rules are applied.
Sencinet does not publicly disclose per-teleport gateway throughput, committed satellite bandwidth, oversubscription, hub-port headroom or the number of terminals that can fail over concurrently.
The five-site claim should therefore be read as installed geographic and operational reach. It raises the ceiling of what a resilient design could do. It does not, on its own, establish five-way diversity for a Colombian branch. A customer-specific design must name the primary and alternate gateways, show that they do not share the critical terrestrial exit or operating dependency, and demonstrate that the terminal can use the surviving path. Without those facts, the fifth teleport may be valuable to Sencinet's regional portfolio while remaining irrelevant to the single branch being tested.
SD-WAN can choose paths only after the paths really diverge
Sencinet's SD-WAN proposition is built around a sensible idea: measure available links, apply application policy and steer traffic over the path that best meets the required performance. The public service description adds centralised visibility, zero-touch deployment, security functions and 24-hour management. That can shorten detection and decision time. It can also provide a consistent way to use Internet, MPLS, mobile and satellite access at many branches.
The logical choice must be backed by physical separation. Two services can enter one building through the same duct. Fibre and a mobile service can share a nearby aggregation site or regional power failure. A terrestrial circuit and a satellite terminal can share customer-premises power and the same edge appliance. Two access contracts can be fulfilled by the same wholesale carrier. A branch can have distinct last miles that meet again at one Sencinet point, one upstream or one data-centre security service.
None of those possibilities is established as a flaw in Sencinet's network; they are the dependencies that a diversity claim has to exclude.
Sencinet has publicly acknowledged the supplier layer. A 2021 announcement seeking regional ISP partners said it wanted to expand distribution, particularly of SD-WAN, through regional providers. The broader WAN article says thousands of qualified ISP suppliers complement the transport layer. This is a normal way to reach enterprises across a large and geographically difficult region. It means the service boundary includes ordering, assurance and restoration relationships with companies outside the Sencinet group.
One public procurement file makes that boundary concrete. In documents relating to a government service, Media Commerce Partners notified Sencinet of activation of a dedicated 10 Mbps data channel in Pereira. That single record does not describe Sencinet's entire access strategy, and it does not show the current state of the circuit. It does prove that at least one Sencinet-delivered Colombian service relied on a named third party for an access channel. It also shows why the customer needs to know who receives the alarm, who can test the physical circuit and who controls field dispatch.
A robust acceptance test would remove the primary link under realistic load, not simply unplug a lab cable. It would record the time needed to detect degradation, move each application class, establish or expand the satellite session, preserve security controls and return cleanly after restoration. It would repeat the test while branch mains power is absent, and it would verify that monitoring remains reachable over the backup. It would also test failure of the preferred satellite gateway or control service. The result should distinguish traffic that continued, traffic that degraded and traffic that was intentionally dropped.
The commercial consequence is straightforward. Managed SD-WAN can reduce the skill burden on a remote customer and coordinate many suppliers. It cannot manufacture diversity after procurement. Sencinet's value depends on its ability to know the underlay, reserve compatible capacity and exercise authority across providers. Buyers should pay for a tested service chain, not for a count of icons representing access types.
AS14187 proves an active routed edge, not a resilient access route
Sencinet Colombia has an observable Internet identity. The LACNIC RDAP record for AS14187 identifies the registrant as SENCINET LATAM COLOMBIA S.A., gives a Bogotá address and records the autonomous system's registration in November 1999. That continuity fits the legal history from Comsatcol through BT Latam Colombia to Sencinet. It supports attribution of the ASN to the Colombian company; it does not show which customer products use it.
On 17 July 2026, the RIPEstat announced-prefixes response returned 43 visible prefixes: 41 IPv4 and two IPv6 in the response list. The RIPEstat AS overview identified the holder as Sencinet Latam Colombia and marked the ASN announced on that date. These observations establish a live routed footprint. They are not a bandwidth meter. A /22 does not carry four times the traffic of a /24 merely because it contains four times the address space, and an IPv6 /32 says essentially nothing about installed port capacity.
The view of external connectivity requires similar restraint. The RIPEstat neighbour response showed nine left- or right-side neighbour observations in its current result, including a repeated AS on opposite sides. Those records describe positions observed in collected AS paths, not nine physical interconnects. They do not identify cities, buildings, routers, fibre entrances, commercial roles or simultaneous usable capacity. A path collector can miss private peering and can show multiple logical observations that traverse shared infrastructure.
A second observation source, bgp.tools for AS14187, reported 41 IPv4 and two IPv6 originated prefixes and listed three upstreams at the time captured: AS52320, AS52468 and AS23520. This strengthens the conclusion that the public Internet edge is not visibly single-homed at the AS level. It still cannot prove that the three adjacencies terminate in separate facilities, use separate metro routes, avoid one submarine or long-haul corridor, or have enough spare capacity for a failover.
The PeeringDB record returned for AS14187 presents a different limitation. It retains the older “BT LATAM COLOMBIA” name and has no facilities or Internet-exchange entries populated. PeeringDB is maintained by participating networks and is not a compulsory registry of every interconnection. The blank facility fields must therefore be read as missing public detail, not as evidence that Sencinet has no facility presence or exchange connectivity. The mismatch between the old name there and current LACNIC attribution also shows why no single directory should control the conclusion.
Terrestrial radio authority adds one more piece. A 2021 MinTIC fixed-service frequency assignment report includes Sencinet Latam Colombia among applicants granted assignments in the SHF process. That is evidence that the Colombian company participated in regulated point-to-point spectrum use. It does not locate every radio, state current authorisation status, disclose channel throughput or connect a particular branch to AS14187.
Taken together, the network evidence is positive but bounded. AS14187 is active, has a substantial set of visible prefixes and has more than one observed external adjacency. Colombia-specific fixed-service activity is documented. What remains absent is the physical topology connecting branches, Bogotá metro infrastructure, Telepuerto ZFB, Nimbus, Naos and external networks. A customer cannot infer local-path independence from the ASN any more than it can infer satellite reserve from the number of teleports.
Installed equipment and usable failover capacity are different inventories
Capacity is where apparently precise numbers can become most misleading. The 2020 transaction disclosed 650 kilometres of owned fibre, 2,000 kilometres of leased fibre, four data centres and five teleports across the divested Latin American business. Those are installed or contracted estate measures. They do not state lit fibre pairs, port speeds, satellite bandwidth, customer commitments, power sold, free racks or spare staff. They also predate six years of growth and change.
A 2021 Sencinet account of a Colombian customer migration offers rare satellite figures. In the Certicámara case study, Sencinet said its regional infrastructure had 358 MHz of spectrum and 106 Mbps of VNO capacity. The terms are not accompanied by satellite, beam, band, teleport, utilisation or measurement definitions, and the figures appear in company-authored background to a data-centre story. They are useful as a dated disclosure, not as a current Colombian failover ceiling. It would be especially unsafe to compare 358 MHz directly with terrestrial Mbps or to assume that 106 Mbps was unused reserve.
Naos supplies a different class of installed number. Cotel's case study lists transformer, generator, UPS and cooling ratings. Those ratings describe equipment capability at a particular facility scope. The same account's claim that capacity filled in four years creates an apparent tension with the expandable UPS figure: expansion capability is not the same as sellable headroom, and “full” can mean commercial occupancy rather than electrical saturation. Only current facility and service records could show the load and reserve available to Sencinet customers.
The public-sector Naos description gives a minimum 24-hour autonomy for the documented service, while the Pereira access record gives one 10 Mbps circuit. Both are useful because they attach a number to a real service context. Neither can be extrapolated to every data-centre room, branch, carrier path or satellite gateway. A ten-megabit link may be ample for one administrative location and negligible for a site carrying video, telemetry and operational applications. Twenty-four hours at a data centre offers little reassurance if the remote terminal loses power after thirty minutes.
AS14187's 43 visible prefixes form a fourth inventory: routable address blocks. They demonstrate active use but reveal no throughput and no congestion margin. Similarly, an SD-WAN appliance may report two healthy tunnels in normal operation without revealing whether the backup can carry the primary load. Installed capacity answers “What exists?” Usable capacity answers “What can carry this customer's priority traffic now, under the same failure that removed the primary path, for the required duration?”
For satellite failover, the usable inventory should include contracted bandwidth on compatible beams, gateway and hub-port headroom, terminal licence and radio capability, contention or priority policy, alternate terrestrial egress from the teleport and the number of branches expected to switch together. For terrestrial failover, it should include committed rates, handoff and aggregation capacity, upstream headroom and physical path separation. For cloud recovery, it should include reserved compute, storage, security and application dependencies at the alternate site.
For power, it should include battery condition, generator load, fuel on site and refill access.
None of those customer-specific inventories is public. That absence is not proof that Sencinet lacks reserve; such details are often commercially sensitive and design-specific. It does mean the regional asset figures cannot close the question. A procurement decision should convert each capacity claim into a dated, scoped quantity and a demonstrated failure test. If the provider cannot state whether a number is installed, contracted, occupied, available or reserved, the number should not be used to price resilience.
Power links the remote terminal, teleport, control plane and data centre
Communications services are electrical systems before they are network diagrams. At the branch, the terrestrial termination, satellite modem, outdoor radio, edge appliance, local switch, authentication equipment and monitoring path all need power. Some components may sit on different circuits; others may be fed by one small unit. A generator may support the main building but omit the antenna heater, outdoor cabinet or carrier demarcation. A fuel plan may assume a road that the same storm has blocked.
At the teleport, the antenna and radio chain, baseband, timing, control, cooling and terrestrial backhaul have their own power domains. At the data centre, the public record for Naos shows substantial redundant electrical plant. At Nimbus, current facility certification supports a high infrastructure rating. Yet no public document links a particular Colombian branch backup service to a named teleport power design or proves that its management and application services remain available through a common utility event.
This is why “separate power equipment” belongs in the physical inspection. The customer should trace each load from the device to the distribution board, measure runtime under realistic consumption and verify automatic transfer. It should confirm whether the carrier's demarcation equipment is included, whether remote alarms survive on backup and whether satellite transmission is permitted at the available reduced-power state. A battery label is not a runtime result; a generator nameplate is not proof of fuel, start reliability or supported load.
Power diversity must also extend beyond the customer fence. Satellite removes dependence on the local fibre route but not on the gateway's electricity and backhaul. A second data centre removes some facility risk but can still depend on the same metropolitan utility area, carrier corridor or identity service. Cloud applications may remain unreachable if the branch can reach the Internet but cannot reach the security or name-resolution services required to establish a session. A complete test should therefore begin with loss of utility supply and follow user transactions, rather than stopping when the backup interface turns green.
The public evidence supports two different confidence levels. Confidence is high that Sencinet operates in two named Colombian data-centre environments with independently certified facility infrastructure and that Naos has documented redundant power features. Confidence is low on branch and teleport runtime because no public asset schedule, one-line diagram, fuel endurance or tested-restoration result is available. The correct response is not to assume failure. It is to make runtime and common-power tests contractual acceptance evidence.
Recovery authority passes through contracts, suppliers and people
When the primary circuit fails, recovery begins with detection but ends with authority. Someone must decide whether the condition is a local equipment fault, an access-carrier problem, a wider terrestrial event or a satellite impairment. Someone must be able to change traffic policy, engage the access supplier, allocate satellite capacity, dispatch a technician, enter a free-zone facility and approve return to the primary route. A 24-hour service desk is valuable only if those rights and contacts remain available during the event.
Sencinet's home page advertises integrated network management and 24/7 support. That is a broad service commitment, not a response-time schedule. The regional ISP programme and the Pereira circuit record show that restoration can cross company boundaries. The data-centre certificates show that facility ownership or certification can sit with Cotel or a free-zone developer while Sencinet provides, operates and supports customer services. Satellite service adds satellite operators and possibly separate teleport or hub suppliers. The recovery path is therefore a chain of accountabilities.
Public evidence also shows a Colombian field-operations function. A Universidad Distrital repository record describes a 2023 project to design and validate operating manuals for equipment used in Sencinet Colombia's satellite services, with validation by the field-services line manager and operations and engineering personnel. A student project cannot establish staffing levels, response time or current procedure quality. It does support the narrower conclusion that satellite equipment operations and field practice exist within the Colombian organisation rather than only as regional advertising.
The Certicámara migration provides a customer-impact example. Sencinet reported that some work required face-to-face migration at Naos and that maintenance windows had to be coordinated because the customer's certification services could not be interrupted. That story concerns a planned migration during the pandemic, not an unplanned branch failure. Its lesson is still relevant: physical access, customer approval and operational sequencing can dominate recovery even when the technical destination is a resilient data centre.
The strongest service arrangement would name one incident commander for the customer, while preserving direct escalation into each dependency. It would specify who can declare disaster mode, reserve or increase satellite capacity, alter application policy, dispatch to the branch, authorise free-zone access, obtain generator fuel and approve return. It would carry current contact paths that do not depend on the failed network. It would also preserve evidence from exercises: detection time, traffic moved, capacity consumed, manual interventions, unresolved alarms and successful user transactions.
No public document establishes that full recovery arrangement for Sencinet Colombia, and it would be unreasonable to expect customer-specific operating terms on a public website. The gap nevertheless changes due diligence. Buyers should ask for responsibility and test evidence, not infer authority from ownership language. The integrator can remain accountable even when it does not own the local loop or building, but that accountability must be matched by enforceable supplier terms and access rights.
The users at risk—and the questions a buyer should settle
Sencinet says its satellite services support Internet access, high-availability backup, multicast, SCADA and machine-to-machine applications for energy, oil and gas, agricultural connectivity and mobile backhaul. It describes a customer base spanning natural resources, government, banks, telecommunications and retail. These are not all equally sensitive to the same failure. A retail branch may tolerate reduced bandwidth while preserving payments. A remote industrial site may need low-volume telemetry and voice more than bulk transfers. A mobile-backhaul site can affect many downstream users.
A certification service at Naos can affect digital identities and transactions far beyond one building.
The common customer is the organisation that has bought continuity, not merely connectivity. Its affected users include branch staff, field operators, central security and network teams, customers using public services, and third parties whose systems exchange transactions with the branch. The impact mechanism runs from a physical fault through loss or degradation of applications: telemetry becomes stale, voice becomes unreliable, authentication times out, transactions queue, or a remote site loses safe access to central expertise. The service priority should be built around those outcomes.
Before treating the five-teleport estate as a Colombian branch safety net, a buyer should settle at least ten questions:
- Which legal company contracts for the Colombian service, and which company remains accountable for every local access, satellite, data-centre and security dependency?
- What are the exact primary and alternate paths from the branch, including building entrance, access supplier, aggregation point, teleport, data centre and external network?
- Which dependencies are physically separate, and which share power, duct, pole route, metro facility, carrier, gateway, control service or operating team?
- Which satellite band, beam, teleport and hub can the installed terminal use, and what is the tested alternate if the preferred gateway fails?
- How much bandwidth is reserved for failover, what priority applies during a mass switchover and how many branches are assumed to move at once?
- What remains powered at the branch and carrier handoff, for how long under measured load, and how will fuel or replacement batteries reach the site?
- Which applications are permitted to degrade, which must continue, and what measured loss, latency and recovery targets trigger traffic movement?
- Who has authority to change policy, allocate capacity, dispatch field staff, enter each facility, contact each supplier and declare service restored?
- When was the full chain last exercised with terrestrial and utility power removed, and what user transactions succeeded over the surviving path?
- What evidence will be supplied after an incident to distinguish access failure, satellite impairment, power loss, congestion, control failure and application failure?
The public record makes a positive case for taking Sencinet Colombia seriously. The company has a legal and operating history dating to 1995, two named Colombian data-centre sites, a named Bogotá teleport, active regulated and routed network resources, regional satellite infrastructure, a managed-access ecosystem and visible field expertise. Naos and Nimbus add certified facility capability. AS14187 adds an active Internet edge. Multiple access types and regional teleports create real options.
The same record refuses an easy conclusion. It does not publish a current map of all five teleports, a Colombian allocation of the inherited fibre estate, a physical route from branch to gateway, current satellite reserve, simultaneous failover assumptions, branch runtime or a complete restoration-authority chain. Those are not small omissions when resilience is the product being bought.
Sencinet's five teleports can protect a remote Colombian site only through the path that begins at that site's handoff. If the terminal and fibre equipment share failing power, if two contracts use one local carrier, if alternate satellite capacity is not reserved, or if recovery waits for an organisation without field authority, regional scale converges on a local single point. The decisive proof is therefore a witnessed end-to-end failure exercise: remove the terrestrial path and normal power, observe the applications, trace every handoff and measure how the people and capacity respond. Five teleports describe reach.
The branch test determines resilience.