JSV TELECOM COMUNICAÇÃO EIRELI makes a local connectivity bill depend on upstream routes and field repair

Summary

• JSV Telecom is a current, locally anchored Brazilian broadband operator rather than a dormant registration. Its retail site offers 400, 600 and 800 Mbps fibre plans, business service, address-level availability checking and customer support; its 2024 service contract identifies CNPJ 31.549.677/0001-07 and a São Bernardo do Campo base; and the official IX.br São Paulo entity list shows AS270701 active on the exchange route servers for both IPv4 and IPv6.
• The access-network case is unusually tangible for a small provider. A 2020 federal notice records JSV among telecom companies whose infrastructure-sharing contracts with Enel Distribuição São Paulo were homologated, while JSV's current sales language says “100% fibre”. That supports a pole-borne local fibre reading, but it does not disclose route kilometres, neighbourhood coverage, ring design, splitter utilisation, spare fibres, cabinets, headends or whether separate streets have independent paths.
• The repair promise contains a material gap. The website advertises resolution within 24 hours, but clause 13.3 of the public contract allows up to 48 hours from a protocolled support request. The contract also gives JSV or its authorised technician exclusive control over provider-owned equipment and the external distribution network. A customer's recovery time therefore depends on triage, crew availability, access permission, spares and the location of the fault, not only on a call-centre answer.
• AS270701 was globally visible at the research cut with one allocated IPv4 /22, one IPv6 /32, six observed IPv4 announcements and seven IPv6 announcements. Multiple adjacent networks appear in public BGP views, and IX.br offers local reach, yet the aggregate routes were overwhelmingly visible through Ufinet AS52468 in a RIPE RIS spot check. Logical alternatives may exist, but public records do not prove physically diverse transit entrances, automatic failover under load, independent power or a second core site.

The invoice hides three clocks

A residential fibre bill looks like a price for speed. JSV's current homepage presents 400 Mbps for R$99.99, 600 Mbps for R$119.99 and 800 Mbps for R$149.99, each with fibre, support and included Wi-Fi. The same page speaks to gaming, families and home working, and offers a separate path for business plans. To a buyer, that creates one monthly choice: select a speed tier and expect the line to work.

Operationally, that bill buys access to three different clocks. The first is the packet clock, measured in milliseconds: how quickly traffic leaves the home, crosses JSV's access and aggregation network, reaches IX.br or a transit provider, and returns. The second is the congestion clock, measured in busy-hour minutes and hours: whether shared optical segments, uplinks and external capacity remain adequate when customers simultaneously stream, play and work.

The third is the repair clock, measured in hours or days: how long it takes to diagnose a failed optical terminal, find a cut cable, obtain access to a pole or property, dispatch a technician, splice fibre, replace a power supply or reroute traffic.

JSV's public record is strongest on the existence of the service and the packet-routing layer. It is much thinner on the second and third clocks. The company homepage says its specialist support resolves problems within 24 hours. The customer contract, however, says requests and questions will be resolved within 48 hours after a protocol is opened. Those statements are not necessarily irreconcilable: a company may aim to beat its contractual maximum. But a household or small business should treat 24 hours as a marketing target and 48 hours as the clearer written outer commitment, subject to the contract's force-majeure provisions.

That distinction matters because a local provider often wins customers through responsiveness rather than national scale. A 24-hour message can be more valuable than another 100 Mbps on the rate card if the operator has enough trained people, splicing equipment, replacement optical network terminals and after-hours authority to act. Conversely, a provider can answer quickly while a physical repair waits for daylight, safe pole access, a building manager, an electricity distributor or a replacement upstream circuit. JSV publishes no crew roster, mean time to repair, night-shift arrangement, spare-parts inventory or outage history.

The support topic is justified by the contract and sales promise; actual repair performance remains to be demonstrated.

A current operator under an older directory name

The directory entity retains the name JSV TELECOM COMUNICAÇÃO EIRELI. Internet-number records also use that name. Customer-facing material now uses JSV TELECOM COMUNICAÇÃO LTDA. The continuity key is the CNPJ, 31.549.677/0001-07, which appears on the current site, in the current contract, in the federal authorization notice and in Registro.br's records for AS270701. This is best understood as continuity of the same operating business through a change in legal form, not as two different networks.

The official Diário Oficial da União page for 7 January 2019 records Act 10,239 for JSV Telecom Comunicação Eireli and the same CNPJ, authorising the company to provide Serviço de Comunicação Multimídia, or SCM, for an indefinite period throughout Brazil. A national authorization is a legal permission, not a national footprint. JSV's visible operating centre is far narrower.

The contract places the company at Rua Vinte e Seis de Março, 392, São Bernardo do Campo, São Paulo. The Google Play listing gives the same street and postcode, names JSV Telecom Comunicação Ltda as developer, and says users can sign contracts, monitor internet use and pay bills. The Apple App Store listing carries the same corporate seller and customer-account functions. These are useful signs of a live billing and support operation. They do not define every served neighbourhood.

Public sales copy asks for a postcode, house number and mobile number before returning coverage. That is a more credible access-network posture than a national colour wash: orderability depends on an address. The contract reinforces it by making activation subject to prior technical feasibility and allowing up to ten days to start service after the adhesion document is signed. For a condominium, written permission from the building manager or owners may be required. Those terms reveal the practical edge of the network. A street can be inside the commercial area while a particular building lacks a viable drop, riser, port or permission.

São Bernardo do Campo is a substantial urban market. The IBGE municipal profile reports 810,729 residents in the 2022 census, an estimated 841,154 in 2025, and density of 1,979.65 inhabitants per square kilometre. Density can improve fibre economics because more potential customers sit near each metre of distribution plant. It can also complicate construction through crowded poles, road crossings, gated buildings, irregular street geometry and competition for riser space. The city-level numbers make a regional fibre business plausible; they do not reveal JSV's share of it.

The physical asset is local fibre on shared urban infrastructure

JSV's current homepage describes the product as 100% fibre optic. The phrase establishes the intended retail medium, but the contract is broader: it permits access by radio, cable, fibre or another applicable technology according to the selected plan. That language may preserve flexibility or cover older deployments. It does not prove that non-fibre access is currently sold. The honest conclusion is that current public sales emphasise fibre, while the legal document does not narrow every connection to fibre-to-the-home.

The stronger physical clue comes from the electricity sector. In the 8 May 2020 Diário Oficial da União, an ANEEL notice says infrastructure-sharing contracts between Enel Distribuição São Paulo and telecom providers named in the annex were homologated; JSV Telecom Comunicação Eireli appears in that annex. This does not map JSV's cable or establish that every route uses Enel poles. It does show a formal access relationship to the urban support infrastructure on which aerial fibre commonly depends.

That relationship turns a broadband subscription into a chain of physical permissions and assets. At the customer end, an optical drop must reach the property. It may cross a pole span, enter a façade, pass through a conduit and terminate at an optical network terminal. Upstream, distribution fibres converge through splitters, closures and cabinets toward aggregation equipment. The cable's path can rely on poles owned and maintained by the electricity distributor even when the glass, closures and telecom electronics belong to JSV.

A pole replacement, road collision, storm damage, unsafe loading or mandated cable reorganisation can therefore affect JSV without being initiated by JSV.

Brazil's regulators treat that dependency as material. Anatel's 2026 pole-contract data page says every SCM provider using shared utility poles had to submit information regardless of size. An April 2026 Anatel update says 3,428 providers reported 4,525 pole-use contracts covering 70.2% of fixed-broadband accesses reported to the agency, while warning that the submissions still required authenticity and correctness analysis. Those national numbers are context, not proof of JSV's 2026 filing status. JSV's company-specific evidence remains the 2020 homologation.

The 2020 record supports a pole-based asset boundary but leaves essential questions open. There is no published count of attachment points, no current pole contract, no neighbourhood route, no ring diagram and no distinction between owned fibre and leased strands. There is also no indication of underground segments or alternative routes around high-risk crossings. A provider can have thousands of valid attachments and still expose one neighbourhood to a single feeder. It can have a ring in the core while every customer drop remains a single radial path. “Fibre” describes the transmission medium; it does not by itself describe resilience.

Installed speed is not usable capacity

The speed tiers show what JSV is willing to sell at the edge. They do not reveal the ratio between sold capacity and capacity engineered through each shared segment. A 400 Mbps plan means the customer port is configured for that service level under specified conditions. It does not mean 400 Mbps of dedicated upstream capacity has been reserved from that home to every internet destination at every moment.

The capacity chain begins inside the property. The contract says provider-supplied equipment may be loaned or rented and remains JSV's property. It places responsibility on the customer to maintain suitable electrical conditions, including line filters and uninterruptible power supplies, and notes that Wi-Fi performance depends on router capability, client devices, walls, furniture and interference. It recommends wired testing and disclaims Wi-Fi measurements. That is a sensible technical boundary: an 800 Mbps service cannot deliver 800 Mbps to an old phone through a congested 2.4 GHz channel simply because the optical line is healthy.

The next boundary is the passive optical network. In a typical fibre access design, several premises share an optical line terminal port through passive splitters. The economics are attractive because one feeder and one port serve multiple subscribers. The operational risk is concentration: a damaged feeder, failed port or contaminated connector can affect a group at once, and simultaneous heavy use can expose how capacity was dimensioned. JSV does not publish its optical platform, split ratios, optical budgets, port occupancy or oversubscription policy. It would be speculative to assign a particular architecture.

What can be said is that any mass-market fibre provider must manage these shared resources, and the public speed card does not show how JSV does so.

At the internet edge, JSV's self-maintained PeeringDB profile declares a 10–20 Gbps traffic level, balanced traffic, an open peering policy, IPv4 and IPv6 support, and South American scope. That is a useful scale signal, but it is self-reported and not a capacity certificate. The same profile lists no public exchange points, no facilities and no port speeds. It also gives suggested prefix limits of 22 IPv4 and 32 IPv6 routes, which are routing-policy values rather than evidence of 22 or 32 allocated networks.

This distinction prevents three common errors. First, PeeringDB's traffic band is not the same as purchased transit capacity. Second, a 10–20 Gbps traffic declaration is not a guarantee that every neighbourhood feeder, optical port or edge router has equivalent spare headroom. Third, the count of BGP announcements is not capacity. JSV can announce an aggregate and several more-specific routes covering the same address space; announcing more routes does not create more addresses or bandwidth.

The practical capacity question is therefore not “does the company sell 800 Mbps?” It plainly does. It is whether JSV measures busy-hour utilisation at the optical port, aggregation link, core router, IX port and each transit circuit; what thresholds trigger upgrades; and whether failover capacity can carry normal traffic when the largest path is unavailable. None of those operating metrics is public.

AS270701 proves a routed network, not a redundant one

The internet-number record is strong. Registro.br's RDAP record for AS270701 identifies JSV Telecom Comunicação Eireli as the registrant and shows registration on 13 May 2020. The matching IPv4 record assigns 189.127.132.0 through 189.127.135.255, a /22 containing 1,024 IPv4 addresses. The IPv6 record assigns the much larger 2804:7010::/32. These allocations establish administrative control of address resources. They do not count customers because addresses may be shared, idle, dynamically assigned, used for infrastructure or divided in many ways.

RIPEstat's AS overview marked AS270701 as announced at the 10 July 2026 observation time. Its routing-status result showed 1,024 announced IPv4 addresses across six visible IPv4 prefixes, seven IPv6 prefixes representing 65,536 /48 units, and near-complete visibility from RIS peers in both address families. Its announced-prefixes result included the /22 aggregate, /23 and /24 more-specifics, the IPv6 /32 and more-specific /33 and /34 routes.

This is meaningful operating evidence. AS270701 is not merely reserved. It originates customer-facing address space into the global routing system, and reverse names observed in the IPv4 block use “dynamic” labels under jsvtelecom.net.br, consistent with an access provider. Cloudflare Radar's AS270701 overview also sees end-user traffic and publishes an estimated customer population. That estimate has varied across snapshots and is modelled from Cloudflare observations, so it should not be converted into a subscriber count. It is a signal that people use the network, not an audited customer total.

The routing security posture is less mature. RIPEstat's RPKI check for 189.127.132.0/22 returned “unknown” with no validating route-origin authorisations at the research cut; the check for 2804:7010::/32 did the same. “Unknown” is not “invalid”: it means relying networks do not have a cryptographic authorization that says AS270701 is permitted to originate those prefixes. Creating appropriate ROAs would reduce ambiguity and make origin validation possible, provided maximum lengths are chosen carefully for the more-specific announcements.

IRR information exists but also deserves caution. JSV's PeeringDB entry names AS-JSVTELECOM, while bgp.tools reported that set as not found at one observation and marked route matches against an unauthenticated IRR source. Public routing can work despite incomplete registry hygiene, but route filters and incident response become easier when the AS set, route objects, ROAs and published policy all agree.

IX.br shortens paths but does not replace transit

JSV is visibly connected to Brazil's largest exchange fabric. The official IX.br São Paulo entity page, updated on 10 July 2026, lists AS270701 as JSV Telecom Comunicação Eireli, classifies it as an internet service provider and shows participation in the multilateral route servers for IPv4 and IPv6. Reverse DNS for the listed interface resolves 187.16.217.11 to as270701.saopaulo.sp.ix.br, and bgp.tools identifies the same IPv4 interface plus 2001:12f8::217:11 for IPv6.

That matters to a São Bernardo provider. An exchange can let JSV reach content networks and other access providers without sending every packet through paid transit. Shorter local paths can reduce latency and transit cost, while route servers can make many bilateral opportunities operationally manageable. For gaming, video and software delivery, the result can be better performance when relevant content is present and routing policy selects the exchange path.

But an IX connection is not a universal upstream. It carries routes exchanged by entities according to policy; it does not automatically provide full internet reachability. JSV still needs transit for destinations not reachable through peering and for resilience if the exchange connection or its transport to the exchange fails. Nor does membership prove that the fibre from São Bernardo to the IX uses a route independent from the transit circuit. Both services could travel over the same metro carrier, duct, building entrance, optical shelf or power domain.

There is also a public-data mismatch worth resolving. IX.br's own page shows active route-server participation, while JSV's PeeringDB API record reports zero exchange points and zero facilities, and a netixlan query returns no entries. IX.br is authoritative for current participation on its fabric; PeeringDB is voluntary and can be incomplete. Still, buyers and peers would benefit if JSV updated the profile with the IX, port capacity, facility or transport arrangement, route-server policy and a working looking glass.

Public BGP paths point heavily to Ufinet

Multiple network relationships are visible around AS270701. RIPEstat's ASN-neighbours view saw left-side paths involving Ufinet AS52468 and Giga+ AS52613, among others, and a right-side relationship with Master Connect AS271331. bgp.tools has also identified Ufinet, Giga+, Jrfiber, Novell Internet and Click & Navegue in upstream-style positions across IPv4 or IPv6 observations. These names describe routing visibility, not contracts, fibre ownership or durable commercial roles.

The distribution of the paths matters more than the length of the neighbour list. In a spot check of RIPE RIS's public looking-glass result for 189.127.132.0/22, almost every populated path placed Ufinet AS52468 immediately before AS270701. The same concentration appeared for 189.127.134.0/24, while IPv6 showed more occasional alternatives but still had Ufinet immediately before JSV on the overwhelming majority of paths. This makes Ufinet the clearest currently visible route for the aggregate address space.

That observation should not be overstated. BGP selects best paths. A backup can be configured, healthy and intentionally less preferred, so it appears rarely or not at all until failure. Route collectors also do not see every private interconnection. Conversely, a long list of peers can be generated by an exchange route server without providing a second full-transit service. The public view therefore supports two conclusions at once: JSV has more than a single isolated BGP session, and global reachability for its main aggregate appears heavily concentrated through one upstream in ordinary conditions.

The decisive resilience evidence would be a controlled failover result. JSV would need to show that withdrawing or disabling the primary Ufinet path causes the full IPv4 and IPv6 tables to remain reachable through another provider within an accepted convergence time; that the alternative has enough capacity for peak traffic; and that its physical route, building entrance, termination equipment and power are independent. A route map alone cannot answer those questions.

The downstream appearance of Master Connect is similarly bounded. A public path with AS271331 to the right of JSV suggests JSV may provide reachability onward to that network. It does not establish a permanent customer contract, a shared facility or a legal relationship. ASNs and route paths are operational evidence, not a reason to recast every adjacent network as a corporate affiliate.

The first failure path is a cut in the access plant

For a household, the most likely visible outage begins closer than the global routing table. A drop cable can be pulled down. A connector can be contaminated. A splice can fail after water ingress or mechanical stress. A distribution cable can be cut during construction. A pole can be damaged by a vehicle or replaced by the electricity distributor. A cabinet can be opened, vandalised or flooded. In a condominium, a riser or shared telecom room can become inaccessible.

The formal Enel-sharing record is evidence that poles are part of JSV's asset environment, but it does not show whether any individual customer has an alternate feeder. Most residential drops are naturally single-homed: two fibres along the same poles would not create useful diversity if the pole line is the common point of failure. Resilience therefore depends on feeder rings or alternate aggregation routes farther upstream, plus the ability to isolate and repair a damaged segment quickly.

JSV's contract puts external-network work firmly on the provider side. When equipment needed for the connection belongs to JSV, maintenance and technical assistance are exclusive to JSV or an authorised service organisation. Customers may not alter the external distribution network or let an unauthorised person handle it. That protects network integrity, but it also makes crew availability a single operational gate. A customer cannot legally solve a provider-side splice problem by hiring a neighbourhood technician.

The public 24-hour claim and written 48-hour term should therefore be read against fault scope. Replacing an optical terminal at a reachable house is different from finding a feeder cut that affects dozens of addresses. The latter may require optical time-domain reflectometry, traffic control, safe pole work, a fusion splicer, closures, fibre stock and coordination with Enel or another infrastructure owner. If several cuts follow one storm, the number of crews and spare kits becomes the bottleneck. JSV publishes no information about those resources.

The second failure path is power

Passive fibre does not need electricity between endpoints, but the service does. The customer's optical terminal and Wi-Fi router require local power. JSV's optical line terminal, switches, routers, monitoring systems and any powered street cabinets require power at their sites. The IX and transit handoffs terminate on powered equipment. A fibre can remain physically intact while every packet stops because one critical rack lost electricity.

JSV's contract explicitly tells customers to provide suitable electrical conditions for loaned or rented equipment and mentions line filters and uninterruptible power supplies. That clause is useful because it identifies the customer-side boundary. It says nothing about runtime at JSV's own sites, generator coverage, battery maintenance, fuel, remote alarms or cooling during a prolonged outage. The website's “100% fibre” claim should not be misread as immunity from local electricity failures.

The correct resilience question is layered. How long will the customer's optical terminal and router run? How long will each aggregation location run? Does the core have batteries only, or a generator? Is there an automatic transfer switch? Are fuel and maintenance arrangements tested? Does the network have a second powered site, or does every route converge on one room? Public material does not answer these questions. Until it does, backup power remains an unverified dependency rather than a credited strength.

The third failure path is upstream loss or congestion

If the local optical signal remains healthy but Ufinet or the transport path to Ufinet fails, JSV needs another route. BGP can shift traffic, but only if an alternative session has a full table or default route, exports JSV's prefixes correctly, remains visible to the outside world and has sufficient capacity. IPv4 and IPv6 must both fail over. DNS, customer authentication, monitoring and management traffic must remain reachable too.

The IX.br route-server sessions help for destinations available at the exchange, but they cannot substitute for the entire internet. A second commercial transit relationship is the conventional answer. Public neighbour data suggests alternatives, especially Giga+ for IPv4 and other paths for IPv6, yet the ordinary aggregate route view is Ufinet-heavy. This is exactly where installed capacity can diverge from usable recovery capacity. A 10 Gbps backup does not protect a 15 Gbps busy hour without congestion; a nominal second carrier does not protect anything if both circuits use one fibre corridor.

Congestion can also be partial. Video may work because a cache is reached at IX.br while an overseas business application performs poorly through transit. IPv6 can follow a different path from IPv4. A more-specific route can be propagated differently from its aggregate. A route leak or filtering error can affect one prefix and not another. JSV's six visible IPv4 announcements and seven IPv6 announcements offer policy flexibility, but the benefit depends on disciplined filters, consistent route objects and active monitoring.

The absence of validating RPKI authorisations is relevant here. A correct ROA would not stop a fibre cut or buy backup capacity, but it would let networks reject an invalid origin during a routing incident. For a provider with a small and stable allocation, completing RPKI is a relatively direct way to improve the control plane while the harder physical work proceeds.

The fourth failure path sits inside the home or business

Not every poor-speed report is a network outage. JSV's contract usefully separates wired service from Wi-Fi conditions. Router processing power, client radio capability, walls, furniture, interference and the remote server can all affect a test. This matters commercially because the company includes Wi-Fi in the plan language, while the written terms place limits around what Wi-Fi measurements can prove.

A disciplined fault process should test in stages. First, determine whether the optical terminal has signal and whether the router has a wide-area address. Second, use a wired device capable of the purchased speed. Third, test to a nearby endpoint and then to representative external services. Fourth, compare IPv4 and IPv6. Fifth, check whether neighbouring JSV users are affected. The point is not to move every failure onto the customer; it is to locate the boundary before a crew is dispatched.

The public app and subscriber portal may help with billing, contract and consumption functions, but there is no public status dashboard showing area incidents. A provider of JSV's apparent size could improve customer trust by publishing outage notices, affected zones, start times, estimated restoration and final cause reports. That would also distinguish a local optical fault from a broad upstream issue and reduce duplicate calls during a major event.

Local labour is the hidden capacity reserve

Fibre economics are often described through homes passed, take-up and bandwidth. Repair labour belongs in the same calculation. Each new customer adds revenue, but the network also accumulates drops, connectors, ports, power supplies and possible fault points. If subscriber density grows faster than field capacity, routine installations can compete with restorations for the same people and vehicles.

JSV's website publishes weekday hours of 9 a.m. to 6 p.m. and Saturday hours of 9 a.m. to 1 p.m., alongside separate sales and support contact paths. Those are visible office or contact hours, not proof that network repair stops outside them. The “resolution within 24 hours” claim implies some ability to handle faults across a broader window, but no after-hours dispatch terms are stated. The contract's 48-hour language is the safer expectation.

Labour quality matters as much as headcount. A call-centre representative can reset a router; an optical technician can measure loss and replace a drop; a splicer can restore a feeder; a network engineer can diagnose BGP policy; and a person authorised to coordinate with the pole owner can manage shared-infrastructure incidents. One employee may cover several roles in a small operator, which can be efficient in normal conditions and fragile during simultaneous failures.

Unofficial signals suggest a real local staff presence but cannot measure capacity. Public professional profiles associate individuals with JSV in São Bernardo do Campo, and the company's own site speaks of a specialist team. Embedded customer-review material on an older landing page contains comments about named support interactions. These signals support the existence of local service work. They do not establish employment status, shift coverage, certifications, crew count, median restoration time or performance during a city-wide incident.

The question is not whether someone answers; it is how many independent failures the organisation can repair at once.

Price depends on density, poles and upstream bargaining

The retail ladder is economically revealing. Moving from 400 to 600 or 800 Mbps raises the monthly fee by less than the proportional increase in headline speed. That is typical where the incremental cost of a higher configured rate is lower than the cost of building the line in the first place. The expensive part is often reaching the address: pole attachment, fibre, closures, splitters, optical ports, installation labour, customer equipment, billing and support. Once the connection exists, selling a higher tier can improve revenue without rebuilding the drop, provided shared capacity has room.

Urban density helps amortise those fixed costs, but only when take-up is high enough. A street with many poles and few customers can be uneconomic even inside a dense city. A condominium can be attractive because many potential subscribers share one building entry, yet it may require permission, riser work and competition with incumbent cabling. JSV's address-check form and technical-feasibility clause are therefore part of the commercial structure, not mere administration: they prevent the price card from becoming a promise to serve every address at the same construction cost.

Pole payments are another recurring input. Anatel's March 2026 update on shared-pole reporting reported an average of R$8.40 per attachment point among early submissions, with a wide range from R$3.19 to R$38.13. Those are national, provisional figures, not JSV's price. They illustrate why an apparently small per-pole charge can become material across a distribution network and why clean records of attachments matter.

Upstream cost has a similar fixed-and-variable character. IX.br can reduce paid transit for traffic exchanged locally, but JSV still pays for transport to the exchange, ports or shared access, equipment and operations. Transit providers price capacity and commit levels. A small provider needs enough headroom to protect experience without buying idle capacity it cannot monetise. That balance is why PeeringDB's declared 10–20 Gbps traffic range is interesting, but only invoices, port telemetry and busy-hour measurements could establish the actual economics.

Who is affected when the chain breaks

The website itself identifies the primary affected groups: households using entertainment, families sharing access, gamers, people working or studying from home, and businesses. A feeder cut can disconnect a cluster of these users at once. A customer-premises power failure affects one address. A failed optical line terminal port may affect a splitter group. A core-site power failure or dominant upstream loss can affect much of the network. The scope expands as the failed component moves inward.

Business consequences can be disproportionate to the subscriber count. A shop may lose card processing and cloud software. A home worker may miss calls or remote sessions. A small office may depend on hosted telephony, messaging and identity services even if the local network itself does not sell those applications. A household can fall back to mobile data; a business with no independent second access may have no adequate substitute.

There is no public basis for claiming that JSV serves hospitals, government agencies, schools or other named critical sites. The company advertises business plans but does not publish customer references. Such users should not be inferred from an address block or route. The article's impact claim is therefore local and generic: residential and business subscribers in the served São Bernardo area depend on the same access, power, routing and labour chain.

What would prove resilience

JSV already publishes enough to establish identity, current retail activity, fibre intent, formal pole access, address resources, IPv4 and IPv6 routing, an IX.br presence and a written repair boundary. The remaining questions are concrete and answerable.

On the access side, the operator could show served neighbourhoods or an anonymised homes-passed map; the share of aerial and underground plant; feeder rings; single-feeder zones; optical platform and split-ratio ranges; spare ports; and the number of independently routed entries for important business buildings. A public map need not expose exact cabinet coordinates to explain whether a district can be isolated by one cut.

On power, it could publish battery runtime bands for access and core sites, generator coverage, maintenance cadence and the location of single power domains. On routing, it could identify full-transit providers by address family, external capacity bands, IX port capacity, physical transport carriers, route diversity and tested failover times. Updating PeeringDB and creating appropriate RPKI authorisations would make that posture easier to verify from outside.

On repair, the useful measures are median and high-percentile restoration time by fault class, after-hours coverage, number of simultaneous field teams, splicer and optical-test availability, spare optical terminals and routers, escalation ownership, and coordination arrangements with Enel. Publishing monthly availability without defining the measured service boundary would be less informative than these operating facts.

A buyer can ask for a smaller version of the same proof. Is the quoted address on a ring or radial feeder? Where does a second business circuit enter the building? Does it use a different pole route and carrier? What is the written restoration target? Are planned maintenance windows notified? Does the backup service have enough capacity to run essential applications? For a business, two lines from different brands are not diverse if both ride the same cable or terminate on the same powered equipment.

A credible regional ISP with a concentration problem still to resolve

JSV Telecom clears the first threshold for a serious company profile. Its current site sells actual fibre tiers and checks addresses. Its contract names the provider, CNPJ, service, location, equipment boundary and support commitment. Federal records connect the company to an SCM authorization and a formal Enel infrastructure-sharing arrangement. Registro.br assigns AS270701 and its IPv4 and IPv6 space. RIPE RIS sees those routes, and IX.br lists the company on São Paulo route servers for both protocols.

That is considerably more than a thin name in a routing directory. It supports all three analytical themes: regional ISP economics, because price and availability depend on local network density; local support labour, because JSV controls maintenance of its equipment and external network; and peering and transit, because AS270701 has live global routes and an IX.br connection.

The downgrade is about resilience, not existence. Public records do not reveal the fibre topology, facilities, core sites, backup power, crew depth or spare capacity. PeeringDB's 10–20 Gbps declaration has no accompanying port or facility inventory. Its exchange section is stale or incomplete relative to IX.br. Route-origin authorisations are absent in the RPKI checks. Most importantly, ordinary public views of JSV's aggregate routes are heavily concentrated through Ufinet even though other BGP relationships appear.

The title's “local connectivity bill” is therefore literal. The monthly amount pays for much more than a speed profile: lawful access to poles, a viable drop, working optical plant, powered electronics, maintained customer equipment, enough shared capacity, an exchange path, at least one dominant upstream and people able to repair the chain. JSV has made the chain visible enough to deserve scrutiny. The next step is to prove that one broken pole route, one failed power domain, one saturated backup or one unavailable crew cannot turn that simple bill into a long outage.