Supervised Distributed Computing: Efficiency and Robustness under a Majority of Adversarial Workers

TL;DR

This paper extends supervised distributed computing to tolerate any fraction of adversarial workers less than one, achieving efficient and robust solutions with minimal honest worker effort using lightweight verification.

Contribution

It introduces a new framework that allows for robust supervised distributed computing with any adversarial fraction less than one, improving efficiency over prior methods.

Findings

Robust solutions are possible for any it<1.

Expected honest worker work is close to a single execution per task.

Lightweight verification enables correctness checking by honest workers.

Abstract

We consider a recently proposed \emph{supervised distributed computing} paradigm \cite{augustine2025supervised} that extends and refines the standard master-worker paradigm for parallel computations. In this paradigm, there is a supervisor, a source, a target, and a collection of workers. The distributed computation is given as an acyclic task graph that is known to the supervisor. The source initially stores the input and the target is supposed to store the output of the computation. The individual tasks of the computation are supposed to be executed by the workers under the guidance of the supervisor. The source, target and supervisor are assumed to be reliable, while a $\beta$-fraction of the workers might be adversarial, for some $\beta \in [0,1)$. This covers, for example, the case where a supervisor has to work with untrusted volunteers. In the standard master-worker approach, the master checks whether the workers correctly execute the assigned tasks, creating a severe bottleneck, whereas in the supervised approach, the supervisor outsources this checking to the workers. Prior to this work, only supervised solutions were known for the case that $\beta$ is a sufficiently small constant. We show that robust and efficient supervised solutions are possible for \emph{any} constant $\beta<1$ while the expected work for the honest workers is close to a \emph{single} execution per task, given that there is a lightweight verification mechanism that allows honest workers to check the correctness of task outputs, which is significantly better than all robust master-worker as well as peer-to-peer approaches known so far.