REACH: Reinforcement Learning for Efficient Allocation in Community and Heterogeneous Networks

TL;DR

REACH employs a Transformer-based reinforcement learning framework to optimize task scheduling in diverse, volatile community GPU networks, significantly improving completion rates and resource efficiency over existing methods.

Contribution

This paper introduces REACH, a novel RL-based scheduler that models scheduling as a sequence scoring problem, effectively handling heterogeneity and volatility in community GPU environments.

Findings

Task completion rates increased by up to 17%.

High-priority task success rate more than doubled.

Bandwidth penalties reduced by over 80%.

Abstract

Community GPU platforms are emerging as a cost-effective and democratized alternative to centralized GPU clusters for AI workloads, aggregating idle consumer GPUs from globally distributed and heterogeneous environments. However, their extreme hardware/software diversity, volatile availability, and variable network conditions render traditional schedulers ineffective, leading to suboptimal task completion. In this work, we present REACH (Reinforcement Learning for Efficient Allocation in Community and Heterogeneous Networks), a Transformer-based reinforcement learning framework that redefines task scheduling as a sequence scoring problem to balance performance, reliability, cost, and network efficiency. By modeling both global GPU states and task requirements, REACH learns to adaptively co-locate computation with data, prioritize critical jobs, and mitigate the impact of unreliable resources. Extensive simulation results show that REACH improves task completion rates by up to 17%, more than doubles the success rate for high-priority tasks, and reduces bandwidth penalties by over 80% compared to state-of-the-art baselines. Stress tests further demonstrate its robustness to GPU churn and network congestion, while scalability experiments confirm its effectiveness in large-scale, high-contention scenarios.