Optimal Resilience in Systems that Mix Shared Memory and Message Passing

TL;DR

This paper characterizes the minimal failure threshold for systems combining shared memory and message passing, linking it to the resilience of algorithms implementing shared objects and solving fundamental distributed tasks.

Contribution

It establishes a precise connection between failure resilience and the ability to implement shared objects and solve key distributed problems in hybrid communication models.

Findings

Minimal failure number equals the resilience of shared object algorithms

Resilience bounds apply to m&m-model and hybrid cluster models

Implications for designing fault-tolerant distributed systems

Abstract

We investigate the minimal number of failures that can partition a system where processes communicate both through shared memory and by message passing. We prove that this number precisely captures the resilience that can be achieved by algorithms that implement a variety of shared objects, like registers and atomic snapshots, and solve common tasks, like randomized consensus, approximate agreement and renaming. This has implications for the m&m-model and for the hybrid, cluster-based model.