As you read this, your network of workstations and servers, each with their own clock, are time stamping files, email, transactions, etc., and all the while your server logs are recording every manner of transaction in the event you need that information. At some point during the day it is quite likely that automatic processes such as archiving, directory synchronization, cron jobs, etc. will execute and alter files based on time stamps. Fundamental to all of this is the belief that the time is correct. Even if the time is not absolutely correct there is often a belief that at least the time is “close enough.”

This paper describes why “close enough” is no substitute for accurate network time and why network time synchronization is critically important. Computer clocks are notorious for drifting. They are typically based on inexpensive oscillator circuits or battery backed quartz crystals and can easily drift seconds per day, accumulating significant errors over time. With increasing distributed computing and our interdependence on network infrastructures, having many clocks continuously drift apart puts the network infrastructure and the applications that run on it at risk. In particular, network operations and application related activities are most susceptible to problems related to the lack of time synchronization.

Read on to learn more about network time synchronization.