English

OLAF: Programmable Data Plane Acceleration for Asynchronous Distributed Reinforcement Learning

Networking and Internet Architecture 2025-07-09 v1 Hardware Architecture

Abstract

Asynchronous Distributed Reinforcement Learning (DRL) can suffer from degraded convergence when model updates become stale, often the result of network congestion and packet loss during large-scale training. This work introduces a network data-plane acceleration architecture that mitigates such staleness by enabling inline processing of DRL model updates as they traverse the accelerator engine. To this end, we design and prototype a novel queueing mechanism that opportunistically combines compatible updates sharing a network element, reducing redundant traffic and preserving update utility. Complementing this we provide a lightweight transmission control mechanism at the worker nodes that is guided by feedback from the in-network accelerator. To assess model utility at line rate, we introduce the Age-of-Model (AoM) metric as a proxy for staleness and verify global fairness and responsiveness properties using a formal verification method. Our evaluations demonstrate that this architecture significantly reduces update staleness and congestion, ultimately improving the convergence rate in asynchronous DRL workloads.

Keywords

Cite

@article{arxiv.2507.05876,
  title  = {OLAF: Programmable Data Plane Acceleration for Asynchronous Distributed Reinforcement Learning},
  author = {Nehal Baganal Krishna and Anam Tahir and Firas Khamis and Mina Tahmasbi Arashloo and Michael Zink and Amr Rizk},
  journal= {arXiv preprint arXiv:2507.05876},
  year   = {2025}
}

Comments

17 pages, 11 figures

R2 v1 2026-07-01T03:51:11.591Z