liveBook · Manning

Physical time

As explained in Think Distributed Systems

Time as measured by clocks and used to characterize the real-time occurrence and duration of events.

Physical time is time measured by clocks and used to describe when events occur and how long they take. Distributed systems use it alongside logical time, but physical time is inherently uncertain across components. section 2.0 section 2.2.3

A synchronous system assumes synchronized clocks or a known maximum deviation, plus strict timing or bounded delays. An asynchronous system provides no such timing guarantee, so a missing message may indicate failure or delay. section 2.0 section 2.1.2

Physical clocks create several practical concerns:

  • Clock skew is a difference in the time readings of two clocks; the clocks may retain a fixed offset. section 2.2.3
  • Clock drift is a difference in ticking frequency; clocks that initially agree can gradually diverge. section 2.2.3
  • Network Time Protocol mitigates skew and drift by having a server broadcast its reading while clients adjust their clocks. section 2.2.3
  • A hypothetical wall clock provides a perfect global reference, but components must use imperfect local physical clocks. section 2.2.3

Two clocks, one event

Suppose Clock 1 and Clock 2 begin five minutes apart and preserve that offset: this illustrates skew. If they begin together but slowly separate, this illustrates drift. Comparing their timestamps therefore requires explicit timing assumptions or synchronization. section 2.2.3

Figure 2.16 Clock skew
A black and white image of a clock

AI-generated content may be incorrect.

Use the clock-skew figure to see how two physical clocks preserve a fixed time difference. section 2.2.3

How Think Distributed Systems covers it

How deep it goes
Defines synchronous timing, clock skew, drift, wall clocks, and synchronization, while deferring detailed failure assumptions.
Examples
Uses fixed-offset clocks, gradually diverging clocks, wall-clock references, NTP synchronization, and time-of-day versus monotonic clock listings.
What it assumes
Assumes only basic awareness of distributed components, messages, events, and clocks; contrasts physical time with logical time.
What you can do after
Evaluate timestamp comparisons and timing-based failure interpretations under synchronized, bounded, or completely uncertain system assumptions.
How it connects
Connects physical time with synchronous systems, asynchronous systems, physical clocks, logical time, clock skew, clock drift, and happened-before relationships.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Think Distributed Systems. The section links point to the parts of the book it draws on.