Global Time Coordination: How the World Stays in Sync

When you check the time on your phone, you are seeing the end of a long chain of infrastructure that spans continents, institutions, and centuries of physics. Your phone is within a few tens of milliseconds of every other synchronized clock on Earth. That did not happen by accident. It is the product of a deliberately built stack: atomic clocks, national time laboratories, a coordinating body in Paris, a protocol called NTP, and a database of time-zone rules maintained by volunteers. Here is how that stack fits together.

Layer 1: Atomic clocks

At the bottom of the stack are atomic clocks. The SI second is defined by the cesium-133 atom: exactly 9,192,631,770 oscillations of the microwave radiation it emits during a specific electron transition. A cesium clock counts these oscillations and produces a frequency so stable it would take millions of years to drift by a second. National metrology institutes — NIST in the US, NPL in the UK, PTB in Germany, and dozens of others — operate banks of these clocks and report their readings to the BIPM. For a deeper look at how these clocks work, see Atomic Clocks and UTC.

Layer 2: TAI and UTC

The BIPM aggregates the readings of about 450 atomic clocks in 80-plus laboratories into a weighted average called International Atomic Time (TAI). TAI is a continuous atomic timescale — it never has leap seconds. From TAI, the BIPM derives Coordinated Universal Time (UTC), which is TAI adjusted by leap seconds to stay within 0.9 seconds of mean solar time at Greenwich. UTC is the official reference for civil time worldwide. When your phone says it is 14:32, it is ultimately tracing back to UTC(k) — the local realization of UTC maintained by your national time laboratory.

The leap-second decisions are made by the International Earth Rotation and Reference Systems Service (IERS), which monitors Earth's rotation and announces adjustments about six months in advance. For more on UTC itself, see What Is UTC?.

Layer 3: Time signal distribution

UTC(k) — the national realization of UTC — has to get to users. Historically this was done by radio time signals: stations like WWV (US), MSF (UK), and DCF77 (Germany) broadcast time codes that receivers can decode to within a few milliseconds. These are still used in clocks and watches that display a small radio-tower icon. For higher precision, GPS satellites carry atomic clocks and broadcast timing signals that receivers can use to synchronize to within tens of nanoseconds. (GPS itself is steered by the US Naval Observatory, which keeps its own realization of UTC.) Galileo, GLONASS, and Beiwei do the same.

Layer 4: NTP — the internet's time protocol

For computers, the workhorse is the Network Time Protocol (NTP), designed in the 1980s by David Mills and still in widespread use. NTP organizes time servers into a hierarchy called strata:

• Stratum 0 devices are the reference clocks themselves — atomic clocks, GPS receivers, radio receivers.
• Stratum 1 servers are directly attached to a stratum 0 device. National laboratories and large tech companies run stratum 1 servers.
• Stratum 2 servers sync from stratum 1. Your company's time server is probably stratum 2.
• Stratum 3 and below sync from the level above. Your laptop, syncing from a stratum 2 or 3 server, is typically stratum 4.

NTP is designed to tolerate network delay and jitter. It exchanges timestamps with its upstream servers, estimates the round-trip delay, and adjusts the local clock gradually (slewing rather than stepping) to avoid discontinuities that could break applications. A well-configured NTP client on a normal internet connection keeps the local clock within a few tens of milliseconds of UTC. For tighter requirements, PTP (Precision Time Protocol, IEEE 1588) can achieve sub-microsecond accuracy on a local network with hardware support.

Layer 5: Time zones and the IANA database

UTC gives the world a common instant; time zones give humans a local clock. The mapping between UTC and local time is defined by the IANA Time Zone Database, maintained by volunteers and shipped with virtually every operating system. When a government changes its time-zone rules, the database is updated and downstream systems pick up the change through normal updates. This is the layer that turns "UTC is 14:32" into "it is 9:32 AM in New York on March 15." For more, see How Time Zones Work.

Why all this matters

The coordination stack is invisible when it works and catastrophic when it fails. A few examples:

• Finance. High-frequency trading systems timestamp orders to the microsecond, and regulators (including the SEC's Consolidated Audit Trail) require clocks synchronized to within 100 microseconds of UTC. A clock that drifts can make a firm's records non-compliant.
• Power grids. Phasor measurement units on electrical grids timestamp readings to the microsecond; without synchronized clocks, grid operators cannot detect instabilities in time to prevent cascading failures.
• Telecommunications. Mobile networks and CDMA base stations require sub-microsecond synchronization to hand off calls between towers.
• Distributed systems. Database replication, log ordering, and conflict resolution all depend on clocks that agree. A 2022 Cloudflare outage was triggered by a leap-second handling bug.

How to keep your own systems in sync

1. Run NTP on every server. Most operating systems do this by default. Make sure it is pointing at reliable upstream servers — your cloud provider's internal time service, or the NTP Pool Project.
2. Never set the clock by hand in code. If you need a timestamp, ask the system clock, which NTP is keeping honest. Manually setting a clock is how you get drift.
3. Store UTC, convert at the edge. This is the single most important rule for time-zone correctness. See What Is UTC? for why.
4. Keep tzdata updated. A server that has not updated its time-zone database will eventually produce wrong local times as governments change their rules.

For day-to-day cross-zone work, our World Clock and Timezone Converter sit on top of this whole stack, so you do not have to think about it.

The bottom line

Global time coordination is one of the great invisible achievements of modern infrastructure. A cesium atom in a laboratory in Boulder, a satellite in orbit, a server in Frankfurt, and the phone in your pocket are all kept within a fraction of a second of each other by a layered system of physics, protocols, and volunteer-maintained databases. You do not need to understand every layer to use it, but knowing the stack exists — and where it can break — will make you a better engineer and a calmer scheduler.