SkyWalker: A Locality-Aware Cross-Region Load Balancer for LLM Inference

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xia, Tian, Mao, Ziming, Kerney, Jamison, Jackson, Ethan J., Li, Zhifei, Xing, Jiarong, Shenker, Scott, Stoica, Ion
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908634823262208
author Xia, Tian
Mao, Ziming
Kerney, Jamison
Jackson, Ethan J.
Li, Zhifei
Xing, Jiarong
Shenker, Scott
Stoica, Ion
author_facet Xia, Tian
Mao, Ziming
Kerney, Jamison
Jackson, Ethan J.
Li, Zhifei
Xing, Jiarong
Shenker, Scott
Stoica, Ion
contents Serving Large Language Models (LLMs) efficiently in multi-region setups remains a challenge. Due to cost and GPU availability concerns, providers typically deploy LLMs in multiple regions using instance with long-term commitments, like reserved instances or on-premise clusters, which are often underutilized due to their region-local traffic handling and diurnal traffic variance. In this paper, we introduce SkyWalker, a multi-region load balancer for LLM inference that aggregates regional diurnal patterns through cross-region traffic handling. By doing so, SkyWalker enables providers to reserve instances based on expected global demand, rather than peak demand in each individual region. Meanwhile, SkyWalker preserves KV-Cache locality and load balancing, ensuring cost efficiency without sacrificing performance. SkyWalker achieves this with a cache-aware cross-region traffic handler and a selective pushing based load balancing mechanism. Our evaluation on real-world workloads shows that it achieves 1.12-2.06x higher throughput and 1.74-6.30x lower latency compared to existing load balancers, while reducing total serving cost by 25%.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24095
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SkyWalker: A Locality-Aware Cross-Region Load Balancer for LLM Inference
Xia, Tian
Mao, Ziming
Kerney, Jamison
Jackson, Ethan J.
Li, Zhifei
Xing, Jiarong
Shenker, Scott
Stoica, Ion
Distributed, Parallel, and Cluster Computing
Serving Large Language Models (LLMs) efficiently in multi-region setups remains a challenge. Due to cost and GPU availability concerns, providers typically deploy LLMs in multiple regions using instance with long-term commitments, like reserved instances or on-premise clusters, which are often underutilized due to their region-local traffic handling and diurnal traffic variance. In this paper, we introduce SkyWalker, a multi-region load balancer for LLM inference that aggregates regional diurnal patterns through cross-region traffic handling. By doing so, SkyWalker enables providers to reserve instances based on expected global demand, rather than peak demand in each individual region. Meanwhile, SkyWalker preserves KV-Cache locality and load balancing, ensuring cost efficiency without sacrificing performance. SkyWalker achieves this with a cache-aware cross-region traffic handler and a selective pushing based load balancing mechanism. Our evaluation on real-world workloads shows that it achieves 1.12-2.06x higher throughput and 1.74-6.30x lower latency compared to existing load balancers, while reducing total serving cost by 25%.
title SkyWalker: A Locality-Aware Cross-Region Load Balancer for LLM Inference
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2505.24095