PowerTimestamp: Toward Global Event Ordering
Journal
Journal of Information Science and Engineering
Journal Volume
41
Journal Issue
6
Start Page
1695
End Page
1712
ISSN
10162364
Date Issued
2025-11
Author(s)
Chen, Yen-Shuo
Abstract
In distributed systems, event ordering is a critical issue, but there is still a lack of a complete solution. In 1978, Lamport introduced the concept of partial ordering, which made the construction of event orderings possible based on causal relationships between events. However, it is still unable to determine the chronological precedence for concurrent events, and the logical clock with a different concept of time is not widely accepted. In this paper, we propose a mechanism called PowerTimestamp to address this limitation by constructing a comprehensive ordering for all events in an environment where an event is earlier if it is certain time error ahead of NTP (Network Time Protocol) timestamp first or with higher priority such as the estimated computing power. To reach transitivity in total ordering, the time error must be well defined and dynamically adjusted according to the arrival time at the destination environment. PowerTimestamp adopts the General Proofof-Work model, invented to reach consensus on blockchain, ensuring both trustworthiness and fairness in the establishment of global event ordering for any event arriving in time. With PowerTimestamp, events can be ordered by occurrence, and synchronization can be deterministic in time with an error of hundreds to hundreds of thousands of microseconds, solving many fundamental problems in distributed systems.
Subjects
general proof-of-work
global event ordering
NTP
partial ordering
powertimestamp
proof-of-work
total ordering
Publisher
Institute of Information Science
Type
journal article