面向制造业的事件驱动 ML 流水线编排：AWS 工业实践经验

View PDF HTML (experimental)

Abstract:We present an industry experience report on three years of operating an event-driven cloud infrastructure for continuous machine learning training in automotive manufacturing. Our system orchestrates GPU-accelerated training of product-specialized model pairs, a physics prediction model and a reinforcement-learning control policy, across multiple plants, coordinating long-running GPU workloads triggered by manufacturing events. The architecture combines Amazon ECS with EC2 GPU capacity providers, SQS-based messaging with dead-letter queues, and an admission-controlled Lambda dispatcher that enforces cluster concurrency limits. A Conductor orchestrator on ECS Fargate initiates dependency-aware retraining chains on a weekly schedule. The entire infrastructure is codified in modular Terraform with multi-account separation. From 40000+ production training jobs we report a 72-78% cost reduction versus always-on GPU infrastructure. A discrete-event simulation confirms that admission control is necessary (naive dispatch loses 65% of jobs) and that queue-draining matches AWS Step Functions latency while eliminating per-job startup overhead. We provide lessons learned and release the simulator and Terraform module skeletons as open-source artifacts.

Comments:	Accepted in the 14th IEEE International Conference on Cloud Engineering (IC2E 2026). It will be hosted on October 13th-15th, 2026 at Santa Clara, California, USA
Subjects:	Machine Learning (cs.LG)
Cite as:	arXiv:2610.06890 [cs.LG]
	(or arXiv:2610.06890v1 [cs.LG] for this version)
	https://doi.org/10.48550/arXiv.2610.06890 arXiv-issued DOI via DataCite

Submission history

From: Zhengyang Gu [view email]
[v1] Thu, 24 Sep 2026 03:56:58 UTC (667 KB)