The fault-tolerant cluster-sending problem

TL;DR

This paper formalizes the cluster-sending problem in Byzantine distributed systems, establishes its lower bounds, and develops optimal protocols, advancing fault-tolerant communication primitives beyond consensus.

Contribution

It introduces the cluster-sending problem, proves its complexity bounds, and provides practical protocols that achieve optimal efficiency in Byzantine fault-tolerant systems.

Findings

Lower bounds established for crash and Byzantine failures.

Practical protocols matching lower bounds are developed.

Provides foundational insights for fault-tolerant distributed communication.

Abstract

The development of fault-tolerant distributed systems that can tolerate Byzantine behavior has traditionally been focused on consensus protocols, which support fully-replicated designs. For the development of more sophisticated high-performance Byzantine distributed systems, more specialized fault-tolerant communication primitives are necessary, however. In this paper, we identify an essential communication primitive and study it in depth. In specifics, we formalize the cluster-sending problem, the problem of sending a message from one Byzantine cluster to another Byzantine cluster in a reliable manner. We not only formalize this fundamental problem, but also establish lower bounds on the complexity of this problem under crash failures and Byzantine failures. Furthermore, we develop practical cluster-sending protocols that meet these lower bounds and, hence, have optimal complexity. As such, our work provides a strong foundation for the further exploration of novel designs that address challenges encountered in fault-tolerant distributed systems.