Introduction: Time is an essential concept in real-time systems, and keeping time using accurate clocks is thus required to ensure the correct operations of these systems. The master source for time is Paris’s International Atomic Time (TAI), an average of several laboratory atomic clocks in the world. Since the earth’s rotational rate slows by a few milliseconds each day, another master time source called the Universal Coordinated Time (UTC) performs leap corrections to the time provided by TAI while maintaining TAI’s accuracy,making the time length of every natural solar day constant.
UTC is used as the world time, and UTC time signals are sent from the National Institute of Standards and Technology (NIST) radio station, WWVB, in Fort Collins, Colorado, and other UTC radio stations to specialized receivers. Selected radios, receiver-clocks, some phone-answering systems, and even some VCRs have the capability to receive these UTC signals to maintain accurate clocks. Some computers have receivers to receive these UTC signals and thus the time provided by their internal clocks is as accurate as UTC. Note that depending on the location of the receiver-clock, there is a delay in receiving the UTC signal. For instance, it takes around 5 ms for WWVB’s UTC signal to get from Fort Collins to a receiver-clock in my Real-Time Systems Laboratory in Houston, Texas.
For computers whose time is kept by quartz-based computer clocks, we must ensure that these clocks are periodically synchronized such that they maintain a bounded drift relative to UTC. Software or logical clocks can be derived from computer clocks [Lamport, 1978]. In this text, when we refer to wall clock or absolute time, we refer to the standard time provided by a bounded-drift computer clock or UTC. Thus there is a mapping Clock: real time→standard clock time.
