Learning to Optimize DAG Scheduling in Heterogeneous Environment

TL;DR

This paper introduces achesis, a reinforcement learning and graph neural network-based algorithm for optimizing DAG job scheduling in heterogeneous data centers, achieving significant reductions in makespan and improvements in speedup.

Contribution

It presents a novel task-duplication based learning algorithm that perceives job dependencies and assigns tasks considering heterogeneity, outperforming existing methods.

Findings

Achieves up to 26.7% reduction in makespan.

Improves speedup ratio by 35.2%.

Outperforms seven baseline algorithms.

Abstract

Scheduling job flows efficiently and rapidly on distributed computing clusters is one of huge challenges for daily operation of data centers. In a practical scenario, a single job consists of numerous stages with complex dependency relation represented as a Directed Acyclic Graph (DAG) structure. Nowadays a data center usually equips with a cluster of heterogeneous computing servers which are different in the hardware/software configuration. From both the cost saving and environmental friendliness, the data centers could benefit a lot from optimizing the job scheduling problems in the heterogeneous environment. Thus the problem has attracted more and more attention from both the industry and academy. In this paper, we propose a task-duplication based learning algorithm, namely \lachesis \footnote{The second of the Three Fates in ancient Greek mythology, who determines destiny.}, aiming to optimize the problem. In the proposed approach, it first perceives the topological dependencies between jobs using a reinforcement learning framework and a specially designed graph neural network (GNN) to select the most promising task to be executed. Then the task is assigned to a specific executor with the consideration of duplicating all its precedent tasks according to an expert-designed rules. We have conducted extensive experiments over standard workloads to evaluate the proposed solution. The experimental results suggest that \lachesisquad can achieve at most 26.7\% reduction of makespan and 35.2\% improvement of speedup ratio over seven strong baseline algorithms, including the state-of-the-art heuristics methods and a variety of deep reinforcement learning based algorithms.