Summary

  • Monitoring’s website describes three access architectures with materially different powered and shared components: GPON fibre to the premises, an individual point-to-point optical line, and FTTB with a building switch between the fibre and the customer.
  • AS25129 had full captured IPv4 RIS visibility and near-full IPv6 visibility on 28 August 2026. That establishes an observable routing footprint, not the availability, power autonomy or restoration performance of any customer access path.
  • A useful continuity record must identify the access technology, every powered element, shared ducts and aggregation, backup-power ownership, incident command, restoration evidence and the remedy attached to the affected service.

One fibre label, three different failure boundaries

Monitoring’s current website does not present access as a single technical product. It describes GPON fibre to the premises, point-to-point fibre to the premises and fibre to the building. Those names may look like implementation detail beside a tariff, but they decide where electrical power, shared equipment and field intervention enter the customer’s path.

The GPON description says the optical distribution path contains no intermediate active equipment. That is a meaningful architectural property. A splitter in the distribution network does not need the local electrical supply required by an Ethernet switch. Removing powered electronics from that portion of the route can remove one class of failure and one maintenance dependency.

It does not make the service passive from end to end. The optical line terminal that serves the passive network is powered. So are upstream aggregation and routing systems. The customer’s optical terminal, router and business equipment need power as well. A passive distribution segment therefore moves the power boundary; it does not erase it.

Monitoring describes point-to-point FTTH differently: an individual optical line runs from the communications node to the subscriber, terminating through a media converter or optical module. This can reduce sharing in the access strand, but an individual fibre is not automatically an independent service. Two fibres may share the same duct, building entrance, aggregation chassis, backhaul or power source. Independence is a property of the complete failure path, not of the word “individual”.

FTTB makes a powered dependency explicit. Fibre reaches the building and a shared switch distributes service over copper to apartments. That switch needs electricity and becomes a common component for the customers behind it. Whether it has battery support, remote monitoring, spare capacity and a tested replacement procedure matters as much as the fibre entering the building.

The public descriptions are valuable precisely because they make these questions possible. They are not a measured inventory of Monitoring’s deployed network, and they do not show which architecture serves a particular address. They do show why a buyer should not ask only whether a connection is “fibre”. The operational question is which architecture applies, where the powered components sit and who is responsible when one fails.

The product table reveals segmentation, not resilience

Monitoring’s tariff page says the current listed plans took effect on 12 April 2026. Home tiers range from advertised access speeds of up to 75 Mbit/s to 900 Mbit/s. Business tiers run from up to 40 Mbit/s to 1,000 Mbit/s. The page also lists public IP-address options.

That table reveals a commercial design. The same provider serves households and organisations with different speed and price points, and it can attach address resources to the service. It does not reveal the contention behind a tier, provisioned backhaul, peak utilisation, restoration target or the cost of keeping field teams and spare equipment available.

The distinction matters for smaller businesses. A 1,000 Mbit/s access headline can coexist with a single powered building switch, a shared backhaul or an untested customer power arrangement. Conversely, a lower-speed service can be operationally adequate when its power, path and restoration controls are understood. Speed and continuity are separate dimensions.

The public tariff also cannot establish whether a higher-priced business plan buys a different access architecture, priority repair, a contractual service level or merely a different speed ceiling. Those terms require a current order form or contract. Procurement that treats the tariff name as a continuity specification imports an assumption that may become visible only during an outage.

A routing footprint is a different kind of evidence

The public registry chain is clear. RIPE Database identifies Scientific-Production Center “Monitoring”, Ltd as the organisation behind ORG-RM4-RIPE and links AS25129, named MONITORING-AS, to that organisation. The organisation record describes an LIR associated with Moldova and a Bendery address. The aut-num object publishes import and export policies involving several neighbouring autonomous systems.

RIPEstat provides a dated control-plane observation. At its 28 August 2026 08:00 UTC query time, it reported 48 observed IPv4 originated prefixes covering 8,192 IPv4 addresses and one IPv6 originated prefix covering 65,536 /48 units. It saw AS25129 at all 327 IPv4 RIS peers in the returned set and 319 of 320 IPv6 peers, with six observed neighbours.

That is strong evidence that AS25129 had a widely observable public routing presence at the captured moment. It says nothing direct about whether an optical terminal had power, a building switch was reachable, a customer received the contracted speed or a repair team restored service within a target. A backbone route can remain visible while a local access node is unavailable.

The announced-prefixes response also needs careful reading. It returned 48 records that include aggregates and more-specific routes from the same address space. The record count is not a customer count, a measure of capacity or a map of physical diversity. Routing policy can change the number of visible prefixes without changing the number of subscribers or the condition of the access network.

The RIPE aut-num policy is similarly bounded. It records intended import and export relationships. It does not prove that every session was established at capture, that each path is physically diverse or that traffic was balanced among them. The captured PeeringDB query returned no network record for AS25129. That absence does not prove the absence of peering or facilities; it means there was no matching self-maintained PeeringDB network object in that API response.

The contract separates access from the wider Internet

Monitoring’s published customer offer identifies the provider, describes prepaid data-access services and commits to service and telephone support under the document’s terms. It also distinguishes the provider’s access service from the availability of external Internet segments. This is an important responsibility boundary: no access operator controls every remote network or server a customer may try to reach.

The boundary should not be used to make every incident external. A customer still needs a way to establish whether the failure sits in premises power, the access line, a building switch, the optical termination, aggregation, an upstream path or the remote service. Each point has a different evidence owner.

The document is a published contractual source, not proof of present performance. It does not provide a current distribution of repair times, customer-affecting minutes, service credits or root causes. It also should not be assumed to be the complete current contract for every product without checking the order and applicable terms. Its value here is structural: it shows why access delivery and the rest of the Internet must be diagnosed separately.

Resilience begins with an inventory of responsibility

For a Monitoring customer, a useful continuity record begins with the installed access architecture. GPON, point-to-point FTTH and FTTB should not share one undifferentiated field. The record should identify the customer terminal, building equipment, optical distribution, serving node, aggregation, backhaul and upstream dependencies.

Each powered component then needs an owner and a test. The relevant questions are not merely whether backup power exists, but what it supports, for how long, under which load, how it is monitored, and when the result was last demonstrated. A battery specification without a dated load test is an input, not an outcome.

Path diversity requires the same discipline. Two commercial services can share a duct, building entrance, node, backhaul or operating team. A route policy naming multiple neighbours can improve control-plane options without proving physical separation. A continuity claim becomes useful only when the customer can link the claimed alternative to a verified failure-domain map.

Monitoring’s public architecture descriptions and AS25129 records therefore form a good starting layer. They establish identity, offered access types and a visible network-resource footprint. The missing layer is performance evidence: a technology-specific inventory, incident chronology, restoration results and an accountable remedy.