Causality in autonomous mobile systems

Abstract

The analysis of causal relations among events in a distributed computation plays a central role in distributed systems modeling. Existing models for causal time-stamping are based on a known set of entities, either processes or data repositories, where events can occur. The advent of mobile computing settings that stimulate cooperation among arbitrary groups of nodes, possibly in isolation, precludes the use of a pre-established set of entity identifiers. Addressing this problem, the article proposes a causality definition and a time-stamping model that allows the analysis of those environments, while retaining compatibility with the classic causality model.