Summary
- Platon says it guarantees 99.9% service availability, but the reviewed public pages do not define the measurement window, exclusions, service credits or achieved result.
- The company describes two independent optical routes to SIX and backup power at its own data centre; public wording alone does not establish separate ducts, entrances, termination equipment or other common failure domains.
- RIPE records connect Platon Technologies, s.r.o. to ORG-PTS36-RIPE, AS208429 and an IPv4 allocation, providing a traceable network-resource footprint rather than proof of end-to-end resilience.
At 99.9%, the missing tenth of one percent equals roughly eight hours and 46 minutes over a 365-day year. That arithmetic is simple. The contract behind it is not.
Platon Technologies says on its homepage that it guarantees 99.9% availability of services. The same page describes the company's own server room, servers and network. Those are attributable claims from the provider, not an independently measured availability series. A buyer still needs to know whether the denominator is a month or a year, whether planned maintenance is excluded, where monitoring occurs, which products are covered, when an incident begins, and what remedy follows a breach.
The distinction matters because a percentage can describe different operating realities. A service measured only at a host interface is not necessarily measuring DNS, upstream reachability, customer access, storage, support response or application recovery. A credit formula may compensate an invoice without restoring a customer operation. Even a precisely calculated availability figure says little about the concentration of risk unless its scope and dependencies are visible.
Platon's infrastructure description provides a more concrete starting point. Its server page says its server room is connected to SIX by two independent optical routes. Its data-centre page describes a 36-square-metre facility about 400 metres from SIX, 10 Gbit/s domestic connectivity to the exchange, 1 Gbit/s international connectivity through GTS CE, generator-backed electricity, dual filtered cooling and monitoring data updated every five minutes. The company also says on its about page that it built the facility in 2009 and provides web-hosting and server-hosting services to customers in Slovakia and the United States.
Each statement is useful, but none resolves the failure map by itself. Two optical routes may be operationally diverse, or they may share a duct, street crossing, building entrance, power source or termination device. A generator may reduce utility risk, but its value depends on rated load, fuel autonomy, maintenance, transfer design and test history. Five-minute monitoring can expose conditions, but it does not establish where probes sit or whether a customer-facing service remained usable. The reviewed pages do not answer those questions, so this article does not treat the published specifications as proof of resilience.
Public number-resource records add a separate evidence layer. The RIPE organisation object records ORG-PTS36-RIPE under the name Platon Technologies, s.r.o., country code SK and organisation type LIR. The RIPE aut-num object records AS208429, the AS name PLATON-AS, status ASSIGNED, and registered import and export policies involving AS44185 and AS5578. A RIPE inetnum search records 45.137.164.0–45.137.167.255 as ALLOCATED PA under SK-PLATON-20190805. At the evidence-capture time, RIPEstat network-info returned 45.137.164.0/22 with origin AS208429.
That alignment makes the network footprint traceable. It does not turn a registered routing policy into a live topology diagram, prove continuous diversity, identify the physical path of every service, or validate the 99.9% claim. The aut-num object's policy dates from 2019, while the RIPEstat response is a timestamped observation. Current upstream relationships and actual traffic engineering require current route and contractual evidence.
Platon's contact page identifies the legal supplier as Platon Technologies, s.r.o., company number 36 279 129, at Hlavna 3 in Sala. That gives a procurement team a named counterparty. The next step is to bind that legal identity to the precise service, facility, network path, incident process and remedy being purchased.
The diligence request is therefore specific. Ask for the operative SLA and its calculation examples; the monitoring points and raw availability history; maintenance and force-majeure exclusions; the service-credit and termination rights; current upstream and prefix-origin inventories; diagrams showing conduit, entrance, termination and power separation; generator runtime and load-test records; recent failover tests; recovery objectives; and incident ownership across facility, network and managed-server layers.
Platon has published more operational detail than a bare hosting label would provide. The defensible conclusion remains narrower than the marketing promise: the company exposes a traceable legal and network-resource footprint and describes several continuity controls, while the independence, tested performance and contractual enforceability of those controls remain matters for verification.
Member Briefing
Deeper Profile Context
Sign in with the right membership level to unlock the full briefing and source notes.
Only for Strategic Circle
Strategic Circle
Open to all readers. Unlock profile briefings after joining and signing in.
Join Strategic CircleOnly for Leadership Alliance
Leadership Alliance
For qualified IP-asset owners and management; sign in to unlock alliance briefings.
Join Leadership Alliance