Supervised Distributed Computing

2025-03-14

AI Summary

In distributed computing with malicious worker nodes, the central master node becomes a scalability and reliability bottleneck. Method: This paper proposes a supervisor-based distributed computing framework that introduces four distinct roles—supervisor, source, target, and worker—and models computation as a directed acyclic graph (DAG). The supervisor performs data-agnostic scheduling solely based on task dependencies, eliminating centralized coordination. We design a fault-tolerant scheduling algorithm and a redundancy execution strategy tailored for adversarial environments. Contribution/Results: Our framework achieves asymptotically optimal communication and computation overhead even when the fraction of malicious workers is arbitrarily high—a first in the literature. Empirical evaluation on supervised matrix multiplication and sorting tasks confirms that both communication and computational complexity remain invariant to the proportion of malicious workers, thereby overcoming the fundamental scalability limitations inherent in conventional master–slave architectures.

Technology Category

• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

We introduce a new framework for distributed computing that extends and refines the standard master-worker approach of scheduling multi-threaded computations. In this framework, there are different roles: a supervisor, a source, a target, and a collection of workers. Initially, the source stores some instance $I$ of a computational problem, and at the end, the target is supposed to store a correct solution $S(I)$ for that instance. We assume that the computation required for $S(I)$ can be modeled as a directed acyclic graph $G=(V,E)$, where $V$ is a set of tasks and $(v,w) in E$ if and only if task $w$ needs information from task $v$ in order to be executed. Given $G$, the role of the supervisor is to schedule the execution of the tasks in $G$ by assigning them to the workers. If all workers are honest, information can be exchanged between the workers, and the workers have access to the source and target, the supervisor only needs to know $G$ to successfully schedule the computations. I.e., the supervisor does not have to handle any data itself like in standard master-worker approaches, which has the tremendous benefit that tasks can be run massively in parallel in large distributed environments without the supervisor becoming a bottleneck. But what if a constant fraction of the workers is adversarial? Interestingly, we show that under certain assumptions a data-agnostic scheduling approach would even work in an adversarial setting without (asymptotically) increasing the work required for communication and computations. We demonstrate the validity of these assumptions by presenting concrete solutions for supervised matrix multiplication and sorting.

Problem

• Extends master-worker model for distributed computing.
• Handles adversarial workers without increasing computational overhead.
• Demonstrates framework with matrix multiplication and sorting.

Innovation

• Supervised distributed computing framework introduced.
• Tasks scheduled via directed acyclic graph model.
• Data-agnostic scheduling works in adversarial settings.