Alcesski Information Technology Solutions makes a local connectivity bill depend on upstream routes and field repair

Summary

• The National Telecommunications Commission's 31 January 2025 list of internet-access providers with valid certificates names Alcesski Information Technology Solutions and gives Parañaque City and Las Piñas City as its service areas. The company's A.I.T.S. website publishes four residential plans from PHP800 for up to 40 Mbps to PHP1,500 for up to 70 Mbps, plus local contact numbers and a BF Homes address. Together, those are credible evidence of a local retail offer, but not of availability at every address or current subscriber volume.
• APNIC's AS153351 registration is active and identifies Alcesski at 137 C. Benitez Street, BF Homes, with Richard Palma in the administrative record. APNIC separately records 160.187.215.0/24 as active, portable IPv4 space assigned to the business. RIPEstat's routing status showed that single 256-address route visible to all 327 reporting IPv4 peers on 10 July 2026, with no originated IPv6 space.
• Public BGP observations show two logical ways into the network. RIPEstat's neighbour view found RISE AS63927 and Hurricane Electric AS6939 immediately before AS153351, and its BGP-state view found most collector paths arriving through RISE. That is better than a single visible adjacency, but it does not establish two physically independent circuits, two facilities, or enough surviving capacity after one path fails.
• Alcesski joined GetaFIX Manila in November 2025 and is marked as a route-server entity. Capacity records conflict: PeeringDB and Internet Society's IXP Tracker show 10 Gbps, while GetaFIX's current member export and member detail show a 100 Gbps interface. Neither figure is evidence of retail throughput, transit capacity, busy-hour traffic or failover headroom.
• The final network evidence grade is Medium. The regulator listing, current tariff page, active APNIC resources, sustained route history, RPKI-valid origin and exchange participation support a live small ISP. The grade cannot be Strong because the access medium, street topology, upstream circuit routes, powered sites, backup runtime, spare stock, field-team depth, service-level terms and measured restoration performance are not public. APNIC also marks the registered abuse mailbox invalid, while exchange and PeeringDB records disagree on speed and naming.

A PHP800 bill meets a disputed 100-gigabit port

A household in southern Metro Manila can buy Alcesski's lowest published plan for PHP800 a month. The advertised ceiling is 40 Mbps, with no data cap and unlimited use. At the other end of the network, a current GetaFIX record associates Alcesski's AS153351 with an interface whose nominal line rate is 100 Gbps. The ratio between those two figures looks enormous. It is also almost meaningless until the layers between them are understood.

The household is not buying a fraction of an exchange port. It is buying a working path from a router or other customer-premises device, across an undisclosed access medium, through local aggregation and metro backhaul, through the operator's edge and onward to the service being used. The IXP interface is relevant only for traffic that can reach an appropriate peer over that exchange. Other traffic needs transit. Every segment has its own contention, power supply, physical route and restoration arrangements. The slowest or failed segment determines what the subscriber experiences.

Alcesski is unusually instructive because the public record became much richer at the internet edge in late 2025 while remaining sparse at street level. GetaFIX's membership page dates Alcesski's entry to 12 November 2025. RIPEstat's routing history for 160.187.215.0/24 begins on 14 November. PeeringDB created its network and exchange records around the same period. These dates support a meaningful upgrade in routing autonomy, not the birth of the retail service itself: the company says it was established in 2018 and the regulator already listed it in early 2025.

That timing is the heart of the story. Alcesski appears to have moved from a locally marketed access business into a more visible independent routing role. The move can improve control, route security and access to local content. It cannot by itself strengthen a pole line, duplicate a feeder, add batteries to an aggregation point, or ensure a technician is free when several customers fail at once.

What can confidently be called Alcesski

The identity chain is coherent across several independent public records. The NTC list names Alcesski Information Technology Solutions as a value-added service provider offering internet access, with Parañaque City and Las Piñas City as the service areas. APNIC's organisation record gives the same name and the 137 C. Benitez BF Homes address. The administrative contact record names Richard Palma and repeats that address. PeeringDB links the same organisation, address, website and AS153351. The company's website supplies the same physical location and the [email protected] contact used in the number registry.

This is enough to identify the business behind the current public route. It is not enough to describe a corporate structure. The public pages reviewed here do not disclose audited accounts, a parent company, shareholdings, an employee roster or a legal document explaining whether every piece of local plant is owned, leased or operated under another party's facilities agreement. The defensible language is therefore operational: Alcesski markets the retail service, holds the internet-number resources recorded by APNIC and originates the route under AS153351.

The company's own history needs a small qualification. Its site says the provider was established in 2018 and founded by Richard B. Palma. The same paragraph says it was established during the COVID-19 pandemic to help people working and studying from home. Those statements do not fit neatly together because 2018 predates the pandemic. They may be compressing an earlier founding and a later expansion, but the page does not explain that. The founder and date should be treated as the business's own account, not as a fully documented chronology.

Current operation has better support than that history. The retail page is reachable and publishes order-facing prices, contact numbers, email addresses and a subscriber login. APNIC resources are active. The route is live. GetaFIX lists the ASN as a full member. Cloudflare's AS153351 quality page and APNIC Labs' Philippine customer-population measurements both observe traffic attributed to the network. These last two are useful external signals, but their estimated user populations are not billing records. They cannot be converted into a subscriber count, revenue estimate or address-level coverage figure.

Two cities are authorised, but the street footprint is not published

The NTC service area is precise at city scale: Parañaque and Las Piñas. The company's address and its own description narrow the centre of gravity to BF Homes in Parañaque. Its public site says the service is intended for residential customers and businesses using vendo WiFi machines. It does not publish a polygon, street list, building list or address checker that would establish how much of either city is actually reachable.

That distinction matters in a dense urban market. The Philippine Statistics Authority counted 182,216 households in Parañaque and 156,899 in Las Piñas in the 2020 census. Those 339,115 households define neither Alcesski's market share nor its build. They show why a city-level authorisation can encompass a very large opportunity while a small provider serves only selected streets, subdivisions, buildings or radio footprints. The official Barangay BF Homes page describes the barangay as one of Parañaque's largest, reinforcing the density opportunity without proving coverage.

The access technology is the largest missing fact. Alcesski calls itself an internet provider and PeeringDB classifies the network as Cable/DSL/ISP. Neither source says whether current homes receive fibre-to-the-home, Ethernet over aerial cable, coaxial service, fixed wireless, leased access from another network, or a mixture. The four retail plans do not identify upload speeds, access medium, installation charge, contract term, included router, public-address option or minimum guaranteed rate. The plan language says only "up to" 40, 50, 60 or 70 Mbps.

It would be easy to infer fibre from Metro Manila's market or from photographs on a commercial page. That would be unsafe. The NTC's fixed-broadband measurement rules explicitly distinguish FTTH, DSL, cable and fixed wireless. A provider can offer internet access through any compliant arrangement. Until Alcesski publishes an access description or a dated installation record establishes the medium at a particular premises, the plant should remain technology-neutral in any resilience assessment.

The /24 represents real routing control, not 256 customers

AS153351 is a recent but genuine internet edge. APNIC registered the number in October 2024 under the name RICHARD-AS-AP; the entity is active and identifies Alcesski in the description. The address block 160.187.215.0/24 was registered days later. APNIC labels it ALLOCATED PORTABLE, which is materially different from a small range merely delegated inside an upstream carrier's larger block. Portable space gives Alcesski a clearer ability to originate the same prefix through another network, subject to routing policy and contracts.

RIPEstat's AS overview marked the ASN announced on 10 July 2026. The announced-prefixes view showed only 160.187.215.0/24 throughout the preceding fortnight. The routing-status view counted one IPv4 prefix, 256 addresses, no IPv6 prefixes and complete visibility among its reporting IPv4 peers. A network-information lookup independently maps an address in that range to AS153351.

The route is also properly authorised. RIPEstat's RPKI validation returns valid, with a route-origin authorisation matching AS153351 and a maximum length of /24. Its routing-consistency view finds the prefix in both BGP and APNIC's routing records. This is good routing hygiene: other networks using RPKI validation have a cryptographic basis for rejecting an unauthorised origin of the same prefix.

None of those facts says how many households are connected. A /24 includes 256 addresses, but some can be used for network infrastructure, customer-facing translation, servers or management. Many customers can share public IPv4 addresses through carrier-grade translation; alternatively, a provider can assign public addresses only to selected users. PeeringDB's self-reported field says "24" IPv4 prefixes, apparently conflicting with the one route observed globally and perhaps reflecting a mask length entered into a prefix-count field. The route collectors are the stronger evidence for what is actually announced.

The absence of originated IPv6 is similarly narrow. Alcesski has an IPv6 address on the GetaFIX peering fabric, 2401:fdc0::15:3351:1, and PeeringDB says unicast IPv6 is supported at the exchange. Yet no IPv6 customer prefix is visible from AS153351. An exchange-interface address allows IPv6 sessions on that local fabric; it does not show that subscribers receive globally routed IPv6. Customers who need inbound reachability, many simultaneous sessions or clean attribution should ask about public IPv4, address sharing and IPv6 rather than infer them from the ASN.

Two visible neighbours improve the edge, with physical diversity unproven

The strongest resilience signal is the presence of two observed adjacent networks. On 10 July, RIPEstat classified both AS63927 and AS6939 on the left side of AS153351 in collected paths. The AS63927 overview identifies RISE; APNIC's AS63927 record names RISE Asia Technology Limited. The AS6939 overview identifies Hurricane Electric. Both are announced networks, and both appear as GetaFIX Manila members.

The distribution is uneven. In the BGP-state snapshot, 342 collected routes reached Alcesski immediately through RISE, while 20 reached it immediately through Hurricane Electric. Counts at route collectors are not traffic shares and should not be read as 94.5 percent versus 5.5 percent of bytes. They show that RISE supplied the dominant observed path propagation and Hurricane Electric supplied a smaller second path in that observation.

Another public view is less generous. The CIDR Report for AS153351 found one upstream adjacency, RISE, and no downstream networks. Its own explanation warns that "upstream" is relative topology, not a verified provider-customer relationship. The difference from RIPEstat is normal for observers with different collection points, but it means the second path should not be treated as universal.

There are at least three possible readings. Alcesski may buy transit from RISE and peer with Hurricane Electric at GetaFIX. It may receive routes from both under other commercial terms. A route server may mediate some exchange reachability. Public BGP establishes reachability paths, not invoices or contracts, so this article does not turn those observations into a corporate claim.

Physical diversity is a separate question. Two BGP neighbours could use one access circuit to the same exchange switch. Two circuits could leave BF Homes along the same pole line, enter the same duct, terminate in the same building or depend on the same power domain. A second logical path that shares the first path's vulnerable segment will not protect against a cut on that segment. Public sources do not identify Alcesski's backhaul provider, route kilometres, building entrances, handoff rooms, optical protection, leased wavelengths or automatic failover tests.

GetaFIX improves local reach, but the speed field is unsettled

Internet exchange participation can be valuable to a small ISP. Alcesski's route-server flag means it can exchange routes with participating members under the exchange's arrangements rather than establish a separate bilateral session with every network. When content and networks are reachable locally, packets can avoid a longer transit path, reducing some cost and latency. The GetaFIX member list includes content networks and other Philippine providers, so the attachment expands the set of potential local paths.

The Alcesski record is specific. PeeringDB gives exchange IPv4 address 103.104.19.159, the IPv6 fabric address, operational status and route-server participation. The exchange's current member export confirms those addresses, the AS macro and an active connection. PeeringDB's GetaFIX exchange record identifies seven facilities and ties Alcesski's exchange side to VITRO Makati 2. Alcesski has not separately published a facility presence, so that association should be understood as the exchange-side record, not proof of a company-owned router, rack or crew at the building.

The line rate conflict is too large to ignore. PeeringDB lists 10,000 Mbps, as does Internet Society Pulse. The GetaFIX detail page and automated member export list 100,000 Mbps. The exchange detail page also labels the ASN "M-NET ICT" while its membership list, export, APNIC and PeeringDB identify Alcesski. This could be a recent port upgrade with lagging external data, a bundled or logical interface, or a display/configuration mismatch. There is no public explanation.

Even a verified 100 Gbps physical port would not mean that Alcesski sends 100 Gbps, buys 100 Gbps of transit or can deliver that rate to customers. It would be the interface ceiling before protocol overhead and before limits elsewhere. PeeringDB separately reports 10-20 Gbps traffic and a heavy-inbound ratio, but those are operator-supplied categories rather than measured public graphs. The safe conclusion is that Alcesski has an active exchange attachment whose recorded interface rate is inconsistent across reputable sources.

Exchange access also cannot replace all transit. A subscriber visiting a service that does not peer at GetaFIX still needs a path beyond the exchange. An exchange outage may leave transit working; a backhaul cut between BF Homes and the metro edge may remove both exchange and transit if they share transport. An IXP brings routing options close to the edge. Its resilience value depends on how Alcesski physically reaches it and what alternative remains when that path fails.

Installed, usable and surviving capacity are different numbers

Capacity is often discussed as one figure, but a local ISP has at least four. Installed capacity is what ports, optics, radios or access systems can theoretically carry. Contracted capacity is what the operator is entitled to use on transit or transport. Usable capacity is the lower amount that can move across the end-to-end path after overhead, contention and equipment limits. Surviving capacity is what remains after the largest plausible failure.

The public record supplies fragments from different categories. A 10 or 100 Gbps IXP interface is installed exchange capacity. PeeringDB's 10-20 Gbps band is a self-description of network traffic, not a contract or independent measurement. The residential plans are customer product ceilings. The single /24 is address capacity, not bandwidth. None reveals the size of RISE transit, the capacity of the Hurricane Electric path, the transport link from Parañaque to Makati, an access split ratio, a wireless sector load or a neighborhood aggregation uplink.

That missing distinction becomes critical during failure. Suppose the normal edge can use a large exchange port and a smaller transit circuit. If the exchange path is lost, every flow may move to transit. The backup can be technically alive and commercially valid yet congest immediately. Conversely, two 10 Gbps paths can be more resilient than one 100 Gbps path if they are physically independent and each carries the actual peak load. Port labels alone cannot answer the question.

The four retail plans create another unknown: oversubscription. Shared broadband networks are economically viable because customers do not all use their advertised peak simultaneously. The ratio between sold peak rates and upstream capacity can be sensible when measured and managed. It becomes a service problem when busy-hour demand outruns the access or upstream bottleneck. Alcesski publishes no subscriber count, aggregate sold rate, busy-hour utilisation, congestion policy or performance distribution, so no public calculation can establish headroom.

The strongest evidence would be a dated traffic series by interface, a diagram separating access, transport, transit and peering, and a failover test showing throughput while the primary path is disabled. Until that appears, the exchange number should be treated as an interface attribute rather than a promise to the PHP800 customer.

The first failure path starts inside the home

The lowest layer is the customer premises. A household's service can fail because the provider's network is down, but it can also fail because the local router has lost power, a power adaptor has failed, a connector is damaged, a WiFi channel is crowded or a drop cable has been disturbed. The company's website does not state which customer-premises equipment is supplied, who owns it, whether replacement is stocked locally or whether the handoff continues to work on a customer's backup power.

That boundary affects diagnosis. A provider may see an edge router and upstream circuits operating normally while one street reports poor service. A customer may experience a WiFi bottleneck below the subscribed rate even when the wired handoff is healthy. Conversely, all lights can look normal on a home device while a feeder, radio sector or upstream session is unavailable. Effective support needs a way to distinguish these states quickly.

Alcesski advertises a customer-relations team reachable through group chats, calls and text messages, and says its repair teams are efficient and effective. This is relevant because the contact method is designed for a local customer base. It is not a service-level commitment. The page does not give support hours, triage targets, mean time to repair, escalation rules, appointment windows, outage credits or a status page with timestamps.

For a household relying on work or study, those omissions are practical. The deciding questions are whether a first-line responder can see the customer's access signal, whether a replacement power adaptor or router is available the same day, and whether a technician can isolate an in-home issue without waiting for a wider outage to clear. A low monthly fee is attractive, but recovery has its own cost in travel, stock and skilled time.

The second failure path runs through the neighborhood

After the premises comes the shared access plant. Because Alcesski has not disclosed its medium, the risk has to be framed without choosing a technology. A wired service may depend on an aerial or underground drop, splitters or switches, building entrances and local cabinets. A fixed-wireless service may depend on line of sight, customer radios, sectors, towers or rooftops. A mixed network can have both sets of dependencies. Every design needs powered active equipment somewhere, unless the local segment is entirely passive up to a more distant powered site.

BF Homes and the two-city service area create a dense but administratively complex setting. New aerial cable, rehabilitation and expansion along major Metro Manila roads are subject to permit requirements under MMDA Resolution 07-11. Executive Order 32 sets a national policy to streamline permits for telecommunications and internet infrastructure. These rules do not prove where Alcesski has cables or poles. They show that physical expansion and repair can cross rights-of-way, utilities, local authorities and property access rather than occur solely at the operator's discretion.

The 2025 Open Access in Data Transmission Act defines last mile, middle mile, backbone, active infrastructure and passive infrastructure, including poles, dark fibre, cable entrances, ducts and utility corridors. Those definitions are useful here because Alcesski's public evidence spans the categories unevenly. The backbone-facing route is visible. The passive last-mile assets and the middle-mile transport are not.

A neighborhood failure can therefore be broad without reaching BGP observers. One damaged feeder, failed local switch, water ingress, misaligned radio, inaccessible cabinet or shared power problem can disconnect many customers while AS153351 and its /24 remain globally visible. Route uptime is not access uptime. The only way to assess this layer is through topology, alarms, maintenance records and customer-side measurements that Alcesski has not published.

Power can preserve the route while darkening the access network

Meralco's consumer guidance lists broken or fallen poles, damaged wires, transformer problems and other incidents that can cause a brownout or safety concern. This is broad local context, not evidence of an Alcesski outage. It illustrates why communications and power failures can overlap when cables and electricity share streets or structures.

A resilient ISP has to decide which sites receive batteries, generators or alternative feeds, and for how long. The edge router at a professionally operated facility may remain powered while a local aggregation point in the service area exhausts its battery. A passive fibre segment may remain intact while the device that lights it goes dark. A wireless customer unit can fail when the house loses power even if the provider's tower survives. Each layer can have a different runtime.

Alcesski publishes no powered-site inventory, battery duration, generator arrangement, fuel plan, dual-feed facility claim or test schedule. PeeringDB's GetaFIX facility data also leaves the field for diverse serving substations unspecified. It would be wrong to turn a data-centre location into a blanket claim of power resilience, and equally wrong to assume that street equipment has no protection. The evidence is simply absent.

The recovery test is not whether batteries were once purchased. It is whether they are monitored, maintained and sized for real load; whether a prolonged outage can be refuelled safely; and whether the backhaul and access endpoints survive together. A battery at one end of a circuit cannot keep the other end alive. For customers, the useful disclosure is expected runtime by service layer and what happens when that runtime is exceeded.

The third failure path sits between BF Homes and the internet edge

The public route proves that AS153351 can announce 160.187.215.0/24. It does not locate the router making that announcement or the transport path carrying customer traffic to it. PeeringDB's exchange-side association points to VITRO Makati 2, while Alcesski's public address is in Parañaque. Some form of metro transport must bridge the operating footprint and the exchange edge, but its supplier, capacity and route are not disclosed.

That transport can be the common point behind apparently diverse upstreams. If RISE and Hurricane Electric sessions both terminate over one link to GetaFIX, a cut before the exchange removes both. If RISE transit is delivered separately and Hurricane Electric is reached only at the IXP, the paths may share a building or a final conduit. If there are two independently routed circuits, the public record does not say so. Logical diversity is positive evidence; physical diversity remains an open question.

Routing failure has its own modes. A configuration error can withdraw the /24; a router can reboot; an RPKI or filter mistake can reject a valid route; one neighbour can stop propagating it; or the route can remain visible while return traffic follows a poor path. RPKI validity reduces unauthorised-origin risk, but it does not protect against an authorised misconfiguration or equipment failure. A route-server session increases reach, but it does not guarantee that every peer accepts the route or that traffic returns symmetrically.

The route history since November 2025 is encouraging. It shows repeated global visibility over roughly eight months, with no long disappearance in the sampled intervals. Fluctuations in the number of collectors seeing the route are not an outage log because collector membership and internet paths also change. The history supports continuity of the announcement, not a measured retail uptime percentage.

The right resilience evidence would identify where the edge routers are, whether there are two, which facilities and power domains they use, how RISE and Hurricane Electric are reached, and what the network carries during a simulated failure. In its absence, Alcesski should be credited with two visible routing options and not with a verified dual-path architecture.

Field labour is part of the product, not an after-sales detail

Alcesski's website places repair teams and customer relations at the centre of its offer. That emphasis is sensible for a local ISP. Large upstream networks cannot repair a customer's drop, gain access to a gated street, replace a damaged local device or trace a neighborhood cable. Those tasks fall to the party operating the local service, whether its staff perform them directly or a contractor does.

Small-provider economics make the promise difficult. Skilled technicians, vehicles, test equipment, safety gear and spares cost money whether or not a fault occurs. A lean team can be responsive on an ordinary day because travel distances are short and local knowledge is deep. It can also become the bottleneck when a power event, storm, construction cut or equipment fault creates several jobs at once. The public site gives no crew count, depot location, on-call rota or contractor arrangement.

Spares determine whether a visit restores service. A technician may diagnose a failed customer device immediately and still be unable to replace it. An access repair may need compatible optics, a radio, a switch, a sealed enclosure, cable, connectors or a power supply. Without a published equipment architecture, even the relevant spare list is unknown. The useful question is not whether a repair team exists but whether it has authority, access, instruments and stock for the likely failure modes.

The contact record adds one warning. APNIC's incident-response contact was last changed in May 2026 and carries a remark that [email protected] is invalid. The same address appears on the company site and across PeeringDB's public roles. This does not prove that customer calls or the Gmail address fail, and it is not an outage metric. It does mean that the public email channel for network-abuse coordination was not validated at the registry when checked. A provider with its own ASN benefits from keeping NOC and incident contacts distinct, monitored and current.

The tariff sits inside a demanding local cost structure

Alcesski's four plans are tightly spaced. An extra PHP200 buys another advertised 10 Mbps at each step from 40 to 70 Mbps. All are described as unlimited and without data caps. The menu is simple, and the PHP1,200 middle plan sits close to the Philippine Statistics Authority's reported PHP1,290.90 average monthly internet spending in the National Capital Region in 2024.

That comparison is contextual, not a price-performance ranking. The survey covers household spending across technologies and providers; it does not say what Alcesski customers pay after installation or what speeds they receive. The same PSA release says 68.7 percent of NCR households had internet at home and that fixed wired broadband was the most commonly available home connection nationally among connected households. It describes a dense, competitive market in which local providers have to control costs while maintaining support.

Density can help. Shorter drops and more potential customers per kilometre of plant can lower the construction cost per premises. Local referrals and a compact service radius can reduce sales and travel expense. Exchange peering can reduce the amount of traffic paid for as transit when popular destinations are reachable locally. Owning portable address space and an ASN can reduce dependence on one upstream's numbering arrangement.

Density can also concentrate failure. A single aggregation point can serve many paying homes. Congestion can appear quickly when take-up grows faster than feeder or upstream capacity. Road works or a pole incident can affect several streets. A compact technician team faces simultaneous calls from the same event. The business decision is how much of the tariff to spend on duplicate routes, batteries, spare equipment and standby labour rather than on the cheapest normal-state delivery.

No audited Alcesski financial statements, subscriber totals, transit invoices or cost allocations are public. The article therefore makes no revenue or margin estimate. The plans show the retail constraint; the infrastructure records show recent investment in routing and peering. They do not reveal whether the recurring price funds enough recovery capacity.

The operator boundary matters when something breaks

Philippine regulation provides a useful clue about this boundary without answering it for Alcesski. NTC Region 2's public VAS licensing requirements include a system description, equipment list, configuration, schedule of rates and copies of valid facilities or network lease agreements with authorised providers. The NTC's 2025 internet-security circular likewise defines a facilities lease agreement as the contract under which a VAS provider or ISP leases telecommunications facilities or bandwidth from a public telecommunications entity.

The regulator's valid-certificate list establishes that Alcesski is authorised for the named cities. It does not publish the underlying agreements or indicate which assets are leased. The APNIC records establish control of the ASN and portable /24. GetaFIX establishes exchange participation. The missing middle consists of the local access rights, backhaul, facilities and field responsibilities that make those public internet resources usable by households.

This division affects restoration. If a customer's drop is Alcesski's responsibility but a feeder is on another operator's plant, two repair queues can govern one incident. If metro transport is leased, Alcesski may diagnose the failure but depend on the supplier for access and repair. If a facility houses the router, the facility controls some power and physical-access conditions while Alcesski controls configuration. A clear service arrangement assigns who opens each ticket, who communicates with customers and when escalation occurs.

The public site does not explain those boundaries. Its local support promise implies that Alcesski is the customer's front door, which is appropriate. Customers and business buyers still need to know whether the provider has enforceable restoration terms with the parties behind that door. A local support number is most valuable when the responder has visibility into every dependency and authority to escalate it.

Who feels the outage

The company identifies residential users, people working and studying from home, and businesses using vendo WiFi machines as intended customers. These are the groups that can be named without speculation. The service areas also contain many other businesses and institutions, but no public contract establishes that Alcesski serves a school, hospital, government office or particular enterprise, so those should not be presented as customers.

For a household, a failure interrupts communication, remote work, school tasks, entertainment and access to online services. For a small business, the same failure can stop cloud applications, card terminals, messaging and customer WiFi. A vendo WiFi operator has a second-order dependency: its own customers lose access when the Alcesski circuit fails, while the business remains responsible for explaining the loss.

The breadth depends on the failed layer. One damaged drop affects one premises. A local switch, radio sector or feeder can affect a street or cluster. Loss of the metro backhaul can affect the whole access footprint even while the ASN remains visible elsewhere. Withdrawal of the sole /24 can make every public address in the block unreachable, although customers using another numbering or upstream arrangement could behave differently. Public evidence does not show whether such an alternate arrangement exists.

Communication is therefore part of recovery. The company advertises group chat, call and text support, channels suited to a local service. A resilient operation also needs an authoritative outage notice that separates individual, neighborhood and upstream faults; gives a start time; describes the affected area without exposing sensitive details; and updates the estimated restoration time. No public status history was found, so past communication and restoration performance cannot be graded.

What would demonstrate resilience rather than merely suggest it

Alcesski has already supplied several positive signals: a valid regulator listing, order-facing plans, its own portable IPv4 space, a valid RPKI origin, sustained route visibility, two observed BGP neighbours and an exchange membership. The next evidence should connect those internet-edge facts to the household.

For the access layer, the company could state which technologies are used, publish an address-availability check, distinguish aerial and underground areas, and explain whether neighborhoods are fed radially or by a protected ring. It need not expose sensitive route coordinates. A broad architecture and the number of independent feeder paths would let customers understand shared failure risk.

For transport and upstreams, it could reconcile the 10 and 100 Gbps exchange records, state whether RISE and Hurricane Electric are transit, peering or route-server paths, and say whether the circuits use separate entrances, facilities and power domains. A failover result should show surviving throughput at busy hour, not merely that a BGP session comes up.

For power and repair, useful disclosures include backup runtime by site class, the number of maintained spare customer devices and critical network units, support hours, severity targets and measured restoration times. The APNIC incident mailbox should be revalidated or replaced with a monitored role address. An outage archive would make continuity and communication observable.

For customers, the tariff needs fuller terms: expected upload rate, minimum or typical performance, installation and equipment charges, contract and cancellation terms, public-address options, IPv6 availability, fault reporting, appointment windows and credits. The NTC provides a public telecommunications complaint form, but a local provider should resolve most service questions before the regulator becomes the escalation path.

These are not demands for a large carrier's bureaucracy. They are the minimum facts that convert a local trust proposition into an assessable infrastructure service.

Watch the route, but watch the repair layer more closely

The most visible near-term watchpoint is routing maturity. AS153351 has less than a year of observed announcement history. Continued visibility, stable RPKI status and the appearance of an originated IPv6 prefix would show further development. A sustained second path visible across more collectors would strengthen logical diversity. None would prove physical independence, but all would make the public edge more robust.

The exchange records should converge. A documented upgrade would explain the 100 Gbps GetaFIX figure; a correction would explain the 10 Gbps external records. The "M-NET ICT" label on the member detail should be reconciled with Alcesski's identity. PeeringDB's claimed 24 IPv4 prefixes should be corrected if it is intended to describe one /24. Clean public network data matters during faults because peers and responders use it to find the right operator and limits.

The regulator evidence also needs renewal over time. The January 2025 NTC list is strong proof that the certificate was valid at that date, and live routing and retail pages carry the operating case into 2026. A newer NTC publication that again names the two cities would remove the date gap. It would still establish authorisation rather than measured service quality.

The hardest watchpoints are local: a published access description, an outage archive, response-time evidence, backup-power policy and proof of independent backhaul. Those are less glamorous than an exchange port, but they decide whether a customer waits minutes or days after a fault.

A credible small ISP with an ungraded last mile

Alcesski Information Technology Solutions clears the threshold for a live regional ISP. Its NTC certificate covered internet access in Parañaque and Las Piñas in January 2025. Its website offers current-looking residential plans and local support. APNIC assigns the business an active autonomous system and portable /24. That route is globally visible, RPKI-valid and observed through RISE and Hurricane Electric. GetaFIX gives the network a direct place in Manila's peering fabric.

The evidence is strongest where customers are least likely to see it: at the public routing edge. It is weakest where the monthly bill becomes a physical service: the drop, pole or duct, local powered node, metro transport, spare shelf and field crew. The exchange record cannot bridge that gap. Nor can a route collector tell whether the same trench carries both apparent paths.

The correct grade is Medium. Alcesski should receive credit for current operation and for taking control of internet-number resources that many micro-providers never acquire. It should not receive unearned credit for a resilient ring, independent upstream circuits, 100 Gbps of usable capacity, tested power reserve or rapid repair. Those remain questions.

For a customer, the practical judgement is straightforward. The PHP800 plan may be good value, and a locally reachable support team may outperform a distant call centre. Reliability depends on what happens after the first shared cable, powered device or metro path fails. Until Alcesski publishes that layer, the bill buys a credible connection with a visible modern edge and an invisible recovery chain.