World map of NTP and NTS: coverage, operators, concentration

Par Richard DEMONGEOT · Campaign of 25 October 2026 · Original measurement from six distinct autonomous networks · Data CC BY 4.0

This map does not say what the state of NTP is in the world. It says what we measured of NTP in the world — not the same thing, and the difference is written wherever it matters. Every figure comes from a dated campaign, probed from six networks, and can be recomputed from the published CSV.

This page is a macro view: the world, the continents, concentration. It does not replace the detail — each server's name, the gap between what an operator documents and what it actually serves, the campaign history live on the continent pages. It recommends no server either: to choose a French source and configure it, go to the France guide.

Our scope, plainly: since 3 August 2026 we probe the whole world, and the geographic split is only a rendering filter. The campaign shown covers 5968 machines across 106 identified countries and 7 continents, plus 80 machines the geolocation does not place. Coverage remains very uneven — 3162 machines for the best-covered continent, a few dozen for the least. When a cell is empty, it means "we measured nothing there", never "there is nothing".

And we are one of the actors in this landscape: RDEM Systems operates 17 of the 310 measured machines serving NTS worldwide — 5.5 %, and 7.8 % of Europe's. On plain NTP our weight is negligible: these shares speak only of authentication. They are shown for every country with the same colour code as everyone else's — including when they are bad news for the landscape.

5968
machines measured
310
serve NTS
5718
distinct operators
106
identified countries (+ 80 unlocated machines)

The world, continent by continent

The queued column counts servers already in our inventory that we have not probed yet. It exists so that a zero row stays readable: a continent can show zero measured machines because we have not gone there yet, which has nothing to do with an absence of servers. Today 697 machines in 66 countries are waiting for a campaign.

These rows are aggregates of the same measurements. "World" includes global anycast and machines whose location we could not establish — which therefore belong to no continent.

ScopemachinesNTP IPv4NTP IPv6 NTS IPv4NTS IPv6 NTS number of machines operators number, all NTS operators number, namedRDEM share of NTS machines queuedthe detail
Africa 90 63 41 2 0 2 85 — 0.0 % 17 page not written yet — 17 servers queued
North America 1275 839 528 35 27 37 1237 — 0.0 % 132 page not written yet — 132 servers queued
South America 97 66 41 17 19 20 96 — 0.0 % 13 page not written yet — 13 servers queued
Asia 646 351 248 16 11 17 639 — 0.0 % 47 page not written yet — 47 servers queued
Eurasia 398 256 49 4 3 4 389 — 0.0 % 13 page not written yet — 13 servers queued
Europe 3162 2173 1494 202 187 217 2983 — 7.8 % 454 the European study
Oceania 218 155 92 10 9 12 215 — 0.0 % 21 page not written yet — 21 servers queued
World (everything we measured) 5968 3925 2501 287 257 310 5718 48 5.5 % 697 —

Stratum distribution

A country with forty servers all at stratum 3 depends on elsewhere; a country with five stratum 1 holds its own time reference. That is what this count shows, and nothing else — a stratum is a logical distance, not a quality.

ScopeS1S2S3S4+no answer
France 33 247 57 8 64
Europe 529 1813 324 46 450
North America 190 670 149 24 242
South America 22 50 2 0 23
Asia 56 330 93 21 146
Oceania 50 97 20 8 43
Eurasia 25 227 23 6 117
Africa 12 55 6 2 15
World 885 3255 624 111 1093

France first

We are French, the question concerns us, and the answer is not flattering. Out of 409 French machines measured, 35 serve NTS — and 45.7 % of those are ours.

This section says what we measured in France. It does not say which source to use: for that there is the guide to French NTP servers — choice by use case, access policies, chrony and w32tm examples.

Francemeasuredreading
Machines measured409—
NTP over IPv4 / IPv6261 / 206 IPv6 remains a minority among French sources
NTS over IPv4 / IPv634 / 30 see the note on the per-family split
Stratum 1 measured33 out of 409 machines — France holds few references of its own
Distinct operators387 of which 9 named operators serve NTS
RDEM Systems' share of French NTS 45.7 % that is us: French NTS depends on a private company
Largest actor's share 45.7 % RDEM Systems —

Why concentration is the real subject. A country showing thirteen NTS endpoints held by a single actor is not a well-served country. This is not a matter of technical redundancy — those machines may be numerous, spread across several sites and several AS, and often are. It is a matter of dependence: a country's authenticated time reference then rests on the decisions of a single organisation, most often a private company. A change of owner, of business model or of priorities is enough to move it, without any outage occurring.

That is why the map always shows two shares — ours and the largest actor's — and why they turn yellow above 33% and red above 50%. The threshold applies to us as to everyone else: a risk scale that spared whoever publishes it would be worthless. As of 25 October 2026, in France, the most concentrated actor is us — 16 NTS machines out of 35, i.e. 45.7%, above the yellow threshold and below the red one.

Country by country

The four counts are kept separate, never merged: a server answering only over IPv4 and a server answering only over IPv6 are not the same fact, and that is exactly what a coverage map must show.

This table gives the volumes. It does not give the names, nor what each operator documents versus what it actually serves, nor what has moved since the previous campaign: that takes prose, country by country, and lives on the continent page. The right-hand column points there when it exists.

Country machines NTP IPv4 NTP IPv6 NTS IPv4 NTS IPv6 NTS number of machines operators number, all NTS operators number, named RDEM share of NTS machines largest NTS actor the study
Germany 744 553 440 53 48 57 712 11 1.8 % 7.0 % PTB Europe
Netherlands 299 208 156 33 36 37 285 4 0.0 % 16.2 % SIDN Labs Europe
France 409 261 206 34 30 35 387 9 45.7 % 45.7 % RDEM Systems Europe
United States 1097 715 450 25 18 27 1064 2 0.0 % 7.4 % Wikstrom Telephone Company —
United Kingdom 408 260 190 20 20 21 392 3 0.0 % 9.5 % Terry Burton Europe
Brazil 41 28 25 17 18 19 41 0 0.0 % 5.3 % (unnamed) a.st1.ntp.br —
Switzerland 144 110 61 15 14 17 132 7 0.0 % 17.6 % Ueli Heuer Europe
Sweden 103 73 48 15 15 15 78 1 0.0 % 80.0 % Netnod Europe
Australia 176 128 75 10 9 12 176 0 0.0 % 8.3 % (unnamed) ntp.haveacry.com —
Canada 149 102 67 9 8 9 144 1 0.0 % 11.1 % (unnamed) ntp.quantum5.ca —
Austria 69 59 32 8 2 8 67 2 0.0 % 25.0 % BEV (Bundesamt für Eich- und Vermessungswesen) Europe
Japan 94 55 47 5 3 5 94 0 0.0 % 20.0 % (unnamed) pool-member-1474d457 —
China 53 28 23 4 1 4 53 2 0.0 % 25.0 % NEU —
Slovenia 21 19 12 3 3 4 19 1 0.0 % 50.0 % chrony.eu Europe
Finland 106 80 61 3 3 3 99 1 0.0 % 33.3 % (unnamed) ntp.gamermoments.org Europe
Russia 380 246 45 2 1 2 371 1 0.0 % 50.0 % (unnamed) time.nksp.su —
Poland 97 61 14 2 2 2 96 0 0.0 % 50.0 % (unnamed) ntp.asseco.cloud Europe
Singapore 93 45 53 2 2 2 93 0 0.0 % 50.0 % (unnamed) pool-member-4aa2ee7f —
Italy 63 36 32 2 2 2 62 0 0.0 % 50.0 % (unnamed) nettuno.ntp.irh.it Europe
Denmark 54 36 15 2 0 2 54 1 0.0 % 50.0 % DFM (institut de métrologie danois) Europe
South Africa 53 36 26 2 0 2 50 0 0.0 % 50.0 % (unnamed) ntp.as3741.net —
Spain 50 32 13 2 1 2 48 1 0.0 % 50.0 % ROA (Real Observatorio de la Armada) Europe
Czech Republic 49 32 21 2 2 2 47 1 0.0 % 50.0 % (unnamed) ntp.hdinternet.cz Europe
Belgium 47 35 16 2 2 2 43 1 0.0 % 50.0 % (unnamed) nts.teambelgium.net Europe
Turkey 18 10 4 2 2 2 18 0 0.0 % 50.0 % (unnamed) ntp.alastyr.com —
Vietnam 18 6 9 1 2 2 18 0 0.0 % 50.0 % (unnamed) pool-member-2f0df5fd —
Luxembourg 12 10 8 2 2 2 11 0 0.0 % 50.0 % (unnamed) pool-member-157e7714 Europe
India 50 32 20 1 1 1 49 0 0.0 % 100.0 % (unnamed) ntpd-rs.sidnlabs.nl —
Norway 41 28 22 0 1 1 39 0 0.0 % 100.0 % (unnamed) vps-trf2.orleans.ddnss.de Europe
Hong Kong 34 13 12 1 1 1 34 0 0.0 % 100.0 % (unnamed) anon-46b16b20 —
Romania 34 19 12 0 1 1 31 0 0.0 % 100.0 % (unnamed) vps-otp1.orleans.ddnss.de Europe
Ireland 27 18 14 1 1 1 25 0 0.0 % 100.0 % (unnamed) vps-dub1.orleans.ddnss.de Europe
Croatia 17 3 3 1 0 1 10 1 0.0 % 100.0 % UNIZG FER REMLAB Europe
Malaysia 17 9 8 1 1 1 17 0 0.0 % 100.0 % (unnamed) pool-member-2f3d384d —
Chile 14 9 7 0 1 1 14 0 0.0 % 100.0 % (unnamed) vps-zal1.orleans.ddnss.de —
Belarus 13 7 6 1 1 1 13 0 0.0 % 100.0 % (unnamed) ntp.datahata.by Europe
Costa Rica 13 12 6 1 1 1 13 0 0.0 % 100.0 % (unnamed) ntp-sjo2.impulsahosting.com —
Moldova 13 9 6 1 1 1 12 0 0.0 % 100.0 % (unnamed) vps-rmo1.orleans.ddnss.de Europe
Mongolia 1 1 0 1 0 1 1 0 0.0 % 100.0 % (unnamed) timeserver.mn —
Ukraine 92 54 21 0 0 0 89 0 — no NTS measured Europe
Taiwan 48 38 9 0 0 0 45 0 — no NTS measured —
Indonesia 45 26 18 0 0 0 43 0 — no NTS measured —
Bulgaria 42 24 16 0 0 0 42 0 — no NTS measured Europe
New Zealand 38 24 15 0 0 0 35 0 — no NTS measured —
Hungary 37 31 17 0 0 0 36 0 — no NTS measured Europe
Slovakia 33 18 13 0 0 0 31 0 — no NTS measured Europe
Thailand 32 19 8 0 0 0 32 0 — no NTS measured —
Philippines 29 3 12 0 0 0 29 0 — no NTS measured —
South Korea 22 12 5 0 0 0 22 0 — no NTS measured —
Latvia 21 18 4 0 0 0 21 0 — no NTS measured Europe
Greece 20 14 10 0 0 0 18 0 — no NTS measured Europe
Portugal 20 11 5 0 0 0 19 0 — no NTS measured Europe
Argentina 17 13 3 0 0 0 17 0 — no NTS measured —
Estonia 15 12 6 0 0 0 13 0 — no NTS measured Europe
Israel 15 5 4 0 0 0 15 0 — no NTS measured —
United Arab Emirates 13 8 4 0 0 0 13 0 — no NTS measured —
Afghanistan 12 12 6 0 0 0 12 0 — no NTS measured —
Kazakhstan 12 9 2 0 0 0 12 0 — no NTS measured —
Pakistan 12 5 1 0 0 0 12 0 — no NTS measured —
Cyprus 11 7 3 0 0 0 11 0 — no NTS measured Europe
Lithuania 11 9 1 0 0 0 10 0 — no NTS measured Europe
Nigeria 10 5 5 0 0 0 10 0 — no NTS measured —
Bangladesh 9 3 3 0 0 0 8 0 — no NTS measured —
Liechtenstein 9 7 4 0 0 0 9 0 — no NTS measured Europe
Mexico 9 4 4 0 0 0 9 0 — no NTS measured —
Iceland 8 5 2 0 0 0 7 0 — no NTS measured Europe
Uzbekistan 8 7 0 0 0 0 8 0 — no NTS measured —
Colombia 7 5 1 0 0 0 7 0 — no NTS measured —
Kyrgyzstan 7 1 1 0 0 0 7 0 — no NTS measured —
Bosnia and Herzegovina 6 4 1 0 0 0 6 0 — no NTS measured Europe
Maldives 6 6 0 0 0 0 6 0 — no NTS measured —
Serbia 6 2 0 0 0 0 6 0 — no NTS measured Europe
Ecuador 5 4 2 0 0 0 5 0 — no NTS measured —
Kenya 5 3 0 0 0 0 5 0 — no NTS measured —
North Macedonia 5 3 2 0 0 0 5 0 — no NTS measured Europe
Uruguay 5 5 3 0 0 0 4 0 — no NTS measured —
Cameroon 4 4 4 0 0 0 4 0 — no NTS measured —
Namibia 4 4 0 0 0 0 2 0 — no NTS measured —
Nicaragua 4 4 0 0 0 0 4 0 — no NTS measured —
Paraguay 4 2 0 0 0 0 4 0 — no NTS measured —
Saudi Arabia 4 2 0 0 0 0 4 0 — no NTS measured —
Venezuela 4 0 0 0 0 0 4 0 — no NTS measured —
Zimbabwe 4 4 3 0 0 0 4 0 — no NTS measured —
Albania 3 2 1 0 0 0 3 0 — no NTS measured Europe
Armenia 3 2 2 0 0 0 3 0 — no NTS measured —
Iran 3 1 0 0 0 0 3 0 — no NTS measured —
New Caledonia 3 3 2 0 0 0 3 0 — no NTS measured —
Azerbaijan 2 1 0 0 0 0 2 0 — no NTS measured —
Honduras 2 1 1 0 0 0 2 0 — no NTS measured —
Madagascar 2 2 0 0 0 0 2 0 — no NTS measured —
Monaco 2 2 0 0 0 0 2 0 — no NTS measured Europe
Morocco 2 1 1 0 0 0 2 0 — no NTS measured —
Réunion 2 1 0 0 0 0 2 0 — no NTS measured —
Angola 1 1 0 0 0 0 1 0 — no NTS measured —
Cambodia 1 1 1 0 0 0 1 0 — no NTS measured —
French Guiana 1 1 0 0 0 0 1 0 — no NTS measured —
Guadeloupe 1 0 0 0 0 0 1 0 — no NTS measured —
Ivory Coast 1 1 1 0 0 0 1 0 — no NTS measured —
Jordan 1 0 0 0 0 0 1 0 — no NTS measured —
Kuwait 1 1 0 0 0 0 1 0 — no NTS measured —
Mauritius 1 0 0 0 0 0 1 0 — no NTS measured —
Montenegro 1 1 0 0 0 0 1 0 — no NTS measured Europe
Panama 1 1 0 0 0 0 1 0 — no NTS measured —
Papua New Guinea 1 0 0 0 0 0 1 0 — no NTS measured —
Sri Lanka 1 0 0 0 0 0 1 0 — no NTS measured —
Tanzania 1 1 1 0 0 0 1 0 — no NTS measured —

The pool, stratum by stratum

This map counts servers published as time services. Alongside them lives another, far larger world: the machines volunteers lend to pool.ntp.org. They do not join the registry — “who publishes a service” and “who lends a machine” are two questions, and mixing them would sink the NTS share for a reason that has nothing to do with NTS. But their stratum breakdown says something the registry does not: what everyone's time actually rests on.

Stratummachinesshare of answerswhat it means
S060.2 %unsynchronised — announces no reference
S149714.3 %own clock (GNSS, atomic, radio)
S2242669.9 %follows a stratum 1
S347013.5 %follows a stratum 2
S4541.6 %four hops from the reference clock
S5180.5 %five hops from the reference clock
S620.1 %six hops from the reference clock
Silent105—did not answer our probe at harvest time

3578 machines lent to the pool, 3473 answered, 1106 reachable over both IPv4 and IPv6. Machines are counted once even when we know both of their addresses. Harvest of 2026-08-05. A pool zone does not mean the country: the project tops up a sparse national zone with servers from neighbouring countries, so a zone says where the operator registered, not where the machine sits.

Full data, per continent and per zone: pool-strata.json (CC BY 4.0).

The method, and what it does not say

Measured means: AUTHENTICATED time was actually obtained — a complete NTS-KE handshake (TLS, ALPN ntske/1 on port 4460, certificate valid for the queried name) FOLLOWED BY a successful NTS-authenticated NTP request, verified from six distinct networks. An open port is not enough, a negotiated ALPN is not enough, and a complete key exchange over a filtered UDP/123 serves no time at all.

Every server is queried from six vantage points in six different autonomous systems. A single point cannot tell an ACL, a rate limit or a geographic restriction from a real outage. The exact fractions are published in the CSV so you can recount with a threshold different from ours.

What this map does not prove:

  • That a country "without NTS" has none. It has none among the sources we found and measured. What was never published, what is reserved to a closed network, what an ACL keeps away from our six vantage points: none of it shows up here.
  • That the machine count reflects the size of a national fleet. It reflects our inventory depth, which is greater where we searched longer — France, for instance.
  • That a server listed here is a server you should use. We do not recommend any server: we measure, and the CSV carries everything needed to filter for yourself (declared access policy, families, stratum, NTS state and its evidence).

The published country is where the address is routed, measured by geolocation — the only indicator that describes the address itself. The prefix holder's country is published beside it, in registration_country, and is never a fallback: across our 8,016 addresses, 1,465 — 18% are not routed in their holder's country. A single column would file nearly one address in five at its hosting provider's head office: all of OVH would read as France, all of Cogent as the United States.

A name's ccTLD and the pool zone serving an address do not count as location: they describe a name, not a machine — Andorra has a pool zone without hosting a single server. A declared location is kept only with its evidence; those without it are quarantined, along with the value they carried and the reason. The CSV's country_source column says, for every entry, where the country comes from and on which date.

Is your server missing?

Declare it in our open repository — one YAML file, a single required field, the hostname — or by e-mail to ntp@rdem-systems.com. It enters the next campaign — and that is how a "queued" row in the table above eventually becomes a measurement. That repository is not a directory of recommended servers: an operator's declaration is worth existence and permission to measure, not a recommendation.

The data

What reproduces this page exactly are the sealed files of the campaign shown — signature, two authority timestamps and OpenTimestamps. Count the rows and the total lands exactly — but each file has its own spelling, and a recipe that does not say which one cannot be replayed:

Both return the same set of machines: it is the same campaign, projected twice.

One caveat, so the check succeeds first time. the files offered below are those of the campaign you are reading, not a “latest known” URL: their names change with every campaign, and that is what makes them citable.

All these files are published under CC BY 4.0:

The pages that take over from this one:

Free NTP Tools

Three independent tools to diagnose your time synchronization: