Timestamping Algorithms: A Characterization and a Few Properties

Abstract

Abstract. Timestamping algorithms are used to capture the causal order or the concurrency of events in asynchronous distributed computations. This paper introduces a formal framework on timestamping algorithms, by characterizing some conditions they have to satisfy in order to capture causality. Under the proposed formal framework we derive a few properties about the size of timestamps and local informations at processes obtained by counting the number of distinct causal pasts which could be observed by an omniscient observer during the evolution of a distributed computation.