For systems in which extremely accurate synchronization with UTC is required, it is possible to equip the system with multiple receivers for WWV, GEOS, or other UTC sources. However, due to inherent inaccuracy in the time source itself as well as fluctuations in the signal path, the best the operating system can do is establish a range (time interval) in which UTC falls. In general, the various time sources will produce different ranges, which requires the machines attached to them to come to agreement.

To reach this agreement, each processor with a UTC source can broadcast its range periodically, say, at the precise start of each UTC minute. None of the processors will get the time packets instantaneously. Worse yet, the delay between transmission and reception depends on the cable distance and number of gateways that the packets have to traverse, which is different for each (UTC source, processor) pair. Other factors can also play a role, such as delays due to collisions when multiple machines try to transmit on an Ethernet at the same instant. Furthermore, if a processor is busy handling a previous packet, it may not even look at the time packet for a considerable number of milliseconds, introducing additional uncertainty into the time. In Chap. 10 we will examine how clocks are synchronized in OSF's DCE.