SARA: SLO-Aware Resource Allocation for Disaggregated Agentic LLM Services

原文：SARA: SLO-Aware Resource Allocation for Disaggregated Agentic LLM Services

作者：Shicong Liu, Xianghao Yu, Zhen Gao, Jun Zhang

来源：arXiv cs.DC（分布式并行）

正文

Computer Science > Performance

arXiv:2609.26763v1 (cs)

[Submitted on 22 Sep 2026]

Title:SARA: SLO-Aware Resource Allocation for Disaggregated Agentic LLM Services

Authors:Shicong Liu, Xianghao Yu, Zhen Gao, Jun Zhang

View a PDF of the paper titled SARA: SLO-Aware Resource Allocation for Disaggregated Agentic LLM Services, by Shicong Liu and 3 other authors

View PDF

HTML (experimental)

Abstract:Recent advances in large language models (LLMs) are driving the emergence of multi-modal and agentic services for mobile users through cloud and edge infrastructures, where long-context workloads pose daunting challenges for inference latency. Existing disaggregated LLM serving systems largely rely on hardware profiling, configuration enumeration, or heuristic scheduling, offering limited analytical guidance for cost-efficient resource allocation. In this paper, we propose SARA, a Service level objectives (SLOs)-Aware Resource Allocation framework for disaggregated agentic LLM serving systems, which maximizes goodput under a deployment cost constraint and a series of quantile-based SLO constraints. By capitalizing on queuing theory, we first model the prefill, KV cache transfer, and decode stages as an M/G/k queue, an M/G/1 queue, and a generalized birth-death process, respectively. The analysis reveals that the prefill and decode stages are dominantly limited by computational capacity and high-bandwidth memory (HBM) resources, respectively. With these mathematical models, we further derive tractable tail behaviors of different stage-wise service level metrics for both light- and heavy-tailed workloads. These characterizations explicitly map workload, model architecture, and hardware parameters to stage-wise SLO constraints and minimum resource requirements. Finally, we develop an effective resource allocation framework to maximize system goodput under limited cost budgets. Simulation and hardware results demonstrate that the proposed framework accurately predicts the stage-wise SLO with mean errors below 5%, and improves system goodput by 26.6% on average over state-of-the-art baseline methods under the same deployment cost.

Comments:

17 pages, 12 figures

Subjects:

Performance (cs.PF); Distributed, Parallel, and Cluster Computing (cs.DC)

Cite as:

arXiv:2609.26763 [cs.PF]

(or

arXiv:2609.26763v1 [cs.PF] for this version)

https://doi.org/10.48550/arXiv.2609.26763

Focus to learn more

arXiv-issued DOI via DataCite (pending registration)

主题

cs.PF · 分布式计算

由「前沿雷达」于 2026-09-23 采集。正文取自原文页面，已保留出处链接。标题与正文版权归原作者所有。