Laser: Unlocking Layer-Level Scheduling for Efficient Multi-SLO LLM Serving
Jianxiong Liao, Quanxing Dong, Yunkai Liang, Zhi Zhou, Xu Chen
2026年份
1被引次数
摘要
Engaging applications with diverse SLO requirements has become indispensable for production-scale LLM serving systems. However, existing systems rely on iteration-level scheduling, which enforces inflexible, unified execution across multi-SLO workloads, significantly constraining the serving efficiency.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
相关 Paper
- Libra: Flexible Request Partitioning and Scheduling for Serving Unbalanced and Dynamic LLM WorkloadsChaoyi Ruan, Yinhe Chen, Dongqi Tian, Yandong Shi 等NSDI 2026 · 被引用 5 次
- JITServe: SLO-aware LLM Serving with Imprecise Request InformationWei Zhang, Zhiyu Wu, Yi Mu, Rui Ning 等NSDI 2026 · 被引用 29 次
- AdaServe: Accelerating Multi-SLO LLM Serving with SLO-Customized Speculative DecodingZikun Li, Zhuofu Chen, Remi Delacourt, Gabriele Oliaro 等EuroSys 2026
- Janus: Multi-LLM Serving at Production ScaleTianbao Zhou, Yi Wang, Yu Zhou, Zirui Liu 等SOSP 2026
- DuetServe: Harmonizing Prefill and Decode for LLM Serving via Adaptive GPU MultiplexingLei Gao, Chaoyi Jiang, Hossein Entezari Zarch, Daniel Wong 等ICML 2026 · 被引用 4 次
