What is the difference between internal and external clock synchronization in distributed systems?

Question

What is the difference between internal and external clock synchronization in distributed systems?

Answer 1

Clock synchronisation in distributed systems usually aims for one or both of the following goals:

1. Making sure that all the nodes in the distributed system have the same internal clock
2. Making sure that the distributed system is in sync with another external clock.

Internal synchronisation is usually done through sync protocols which allow a compute cluster to sync up their local clocks. The machines agree on using a common time. However, the time they agree on does not necessarily need to be in sync with an external clock i.e. with the time in a certain timezone.

External synchronisation makes sure that computing systems sync their clocks with an external time source such as a time providing server using the NTP protocol. The goal is to have the computing system sync up with the time in a certain time-zone. If a highly accurate time is required, usually NTP systems that generate time from an atomic clock are used.

In both internal and external synchronisation the protocol NTP can be and is widely used.

Answer 2

External synchronisation each process clock is synchronized with respect to an external time source, such us UTC time.

Example: Christian's algorithms and NTP.

Internal synchronisation does not use external source, instead the process in the group themselves use each others clock to synchronize

Example: Berkeley algorithm.