Time Synchronization in Distributed Systems without a Central Clock
Time synchronization and a common understanding of the concept of time is an essential and important aspect of modern computing. Among the drawbacks of current and popular methods, such as the Network Time Protocol, (NTP), or the Precision Time Protocol, (PTP), are the need for live synchronization, and the use of a central reference clock. While these methods are ideal in an environment with constant, uninterruptible, and reliable internet connection, they may prove impracticable in a number of scenarios including Internet of Things (IoT) devices. In cases where the is no access to a central clock, or where the accuracy of two clocks is paramount yet there exists no reliable network access, new methods for synchronization are needed. Our proposed method utilizes existing hardware, without the need for onerous algorithms, in order to establish time-accuracy between two devices. The method effectively establishes a master-slave relationship between two computers, and endeavors to find the difference in time between the two devices, over a period of time. This reduces the reliance on constant, and on-going connection to a central clock. Further, in cases where methods such as NTP, or PTP may be preferred, our proposed solution can bolster those protocols.
Cite this version of the work
Takin Tadayon (2019). Time Synchronization in Distributed Systems without a Central Clock. UWSpace. http://hdl.handle.net/10012/14465