Summary

  • RSI-NET warned that power work at MMTS-9 could interrupt communications services for 30 to 60 minutes in Novaya Shcherbinka and Kaskad Park.
  • That notice identifies a bounded dependency signal, not an outage result or a complete resilience model; customers still need a service-to-path restoration ledger.

One notice makes an invisible dependency visible

On 20 August 2026, RSI-NET published a notice saying that power work would take place the next day at MMTS-9 after power damage there. It warned of a possible communications interruption lasting 30 to 60 minutes and named two affected areas. This is stronger evidence than a generic reliability promise: it names a cause, place, time window and potential impact.

The limits matter just as much. The notice does not say whether the interruption occurred, which services would be affected, how many customers were exposed, what equipment depended on the power work, or whether traffic could move to another path. A planned window is not a measured outage and cannot be converted into an availability percentage.

The service catalogue creates several continuity questions

The official site presents internet, television and telephony, together with subscriber statistics, online support, account settings and a speed check. Its office page lists technical-support hours from 09:00 to 22:00 every day. Those hours show when a support channel is advertised as available; they do not establish round-the-clock restoration coverage or a restoration target. The functions may rely on different networks and systems. They may also converge on common transport, authentication, billing, power or control infrastructure. The public pages do not establish either design.

That distinction changes customer impact. Loss of an access path may stop internet connectivity while television continues. A shared authentication or account failure may block several functions even when transport remains available. A power event at an interconnection site may affect only one area, or it may expose a wider concentration. Each is a testable scenario, not a conclusion from the notice alone.

Registry and participant records identify, but do not assure

The RIPE NCC directory lists RamSvyazInvest LLC as a Russian member. A participant-maintained PeeringDB entry links the company name, RSI-NET website and ASN 59637. This research links to the company's existing BTW Media directory entry for reader navigation; that entry is not operational assurance. These are useful identifiers for further diligence. They do not establish current route visibility, traffic, capacity, physical diversity, customer-service use or restoration performance.

The evidence chain therefore needs separate steps: legal and institutional identity; registered resources; dated routing observation; internal mapping to services and areas; and tested continuity. Skipping a step turns identifiers into an assurance they cannot provide.

The practical deliverable is a restoration ledger

For each customer-facing service and locality, RSI-NET and a buyer could record the primary delivery path, alternate path, power domain, shared equipment, service owner, external dependency, monitoring point, escalation route, target restoration time and last failover test. Every entry should state its evidence date and scope.

The 30-to-60-minute notice could then be attached to the exact rows it informed. Unknown cells would remain unknown. Over time, notices and tests would show whether the same dependency recurs, whether alternatives are genuinely independent and whether published support channels align with restoration responsibility.

Sources