Nobody Knows What Time It Is
summary
Starts with sundials, hourglasses and atomic cesium seconds, then shows why clocks drift and NTP fails in distributed systems using a concrete ticket‑booking example where out‑of‑order timestamps break reservations. Presents Leslie Lamport’s concrete algorithm: each process keeps a counter, increments on events, sends the counter with messages, and on receive takes the max and increments, producing Lamport logical clocks that enforce the happened before partial order rather than real time. Mentions vector clocks as the stronger alternative.
AI-written summary · about
chapters (6)
- 0:00Introduction to time complexity
- 1:51Brief history of timekeeping
- 8:59Standardizing time zones
- 13:15Time in distributed systems
- 21:28Lamport clocks and causality
- 33:34Why time doesn't matter
related talks
47mAdventures in Spacetime
30mOk, it's time to synchronize our clocks...
27mDistributed Systems, Like It or Not
14mLightning Talk: Beyond NTP: High-accuracy clocks for distributed observability
48mAlright, Time to Synchronize Our Watches...
32mWas Leslie Lamport Right?
28mRock around the clock (synchronization
42mTime-sensitive protocol: Links From Design to Implementation