Time Synchronization and Logical Clocks

Distributed systems face a brutal challenge with time synchronization because every computer's internal clock drifts

at a different rate. Physical clocks can drift by about one second every eleven days,

which means without regular resynchronization your nodes will disagree on what time it is.

Cristian's algorithm solves this by having clients ping a UTC server and adjust for network delay,

while Berkeley's algorithm elects a master that polls all nodes and computes a fault tolerant

average to bring everyone into agreement. But here's the twist: for ordering events you don't actually

need real time at all. Lamport clocks use simple counters that increment with

each event and jump forward when receiving messages, giving you causality without physics.