Synchronous, asynchronous, and causally ordered communication

Abstract

This article studies characteristic properties of synchronous and asynchronous message communications in distributed systems. Based on the causality relation between events in computations with asynchronous communications, we characterize computations which are realizable with synchronous communications, which respect causal order, or where messages between two processes are always received in the order sent. It is shown that the corresponding computation classes form a strict hierarchy. Furthermore, an axiomatic definition of distributed computations with synchronous communications is given, and it is shown that several informal characterizations of such computations are equivalent when they are formalized appropriately. As an application, we use our results to show that the distributed termination detection algorithm by Dijkstra et al. is correct under a weaker synchrony assumption than originally stated.

Communication Patterns in Distributed Computations

“…In a classification of a hierarchy of communication patterns, Charron-Bost et al observed that a distributed algorithm designed to run correctly on asynchronous systems (called A-computations) may not run correctly on synchronous systems}an algorithm that runs on an asynchronous system may deadlock on a synchronous system [9]. A-computations that can be realized under synchronous communication are called realizable with synchronous communication (RSC) computations.…”

“…The communication patterns identified in this framework are shown to be generalizations of those used in areas or problems such as: formulating the temporal interactions of intervals [13], synchronous and causally ordered communication [9], determining size of logical clocks [8,20], designing distributed implementations for multilevel secure replicated databases and hierarchically decomposed databases [1,2], ordering of concurrent events without synchronization [1], transfer of knowledge [7], concurrency measures [10], determining necessary and sufficient conditions for a consistent global state [6,19] which is useful in checkpointing and recovery [3][4][5], and defining distributed deadlocks [14]. Thus, the paper shows that key concepts and structures in areas such as the above are instantiations of the communication patterns identified in the presented framework.…”

A Survey of Active Object Languages

“…Another synchronization constraint is related to message ordering [26]. Ensuring an order of message reception, like point-to-point FIFO, adds synchronization which can lead to a loss of efficiency and a gain in program properties.…”

Rely-Guarantee Based Reasoning for Message-Passing Programs

“…We adopt asynchronous message passing: send commands will not be blocked, while receives will be blocked if there is no message received on the designated ports. We assume the state of a port is a queue, and messages are transmitted following the FIFO-principle [3].…”