Cost-Efficient LLM Serving in the Cloud: VM Selection with KV Cache Offloading

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Kihyun, Kim, Jinwoo, Chung, Hyunsun, Cha, Myung-Hoon, Kim, Hong-Yeon, Kim, Youngjae
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913796334813184
author Kim, Kihyun
Kim, Jinwoo
Chung, Hyunsun
Cha, Myung-Hoon
Kim, Hong-Yeon
Kim, Youngjae
author_facet Kim, Kihyun
Kim, Jinwoo
Chung, Hyunsun
Cha, Myung-Hoon
Kim, Hong-Yeon
Kim, Youngjae
contents LLM inference is essential for applications like text summarization, translation, and data analysis, but the high cost of GPU instances from Cloud Service Providers (CSPs) like AWS is a major burden. This paper proposes InferSave, a cost-efficient VM selection framework for cloud based LLM inference. InferSave optimizes KV cache offloading based on Service Level Objectives (SLOs) and workload charac teristics, estimating GPU memory needs, and recommending cost-effective VM instances. Additionally, the Compute Time Calibration Function (CTCF) improves instance selection accuracy by adjusting for discrepancies between theoretical and actual GPU performance. Experiments on AWS GPU instances show that selecting lower-cost instances without KV cache offloading improves cost efficiency by up to 73.7% for online workloads, while KV cache offloading saves up to 20.19% for offline workloads.
format Preprint
id arxiv_https___arxiv_org_abs_2504_11816
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cost-Efficient LLM Serving in the Cloud: VM Selection with KV Cache Offloading
Kim, Kihyun
Kim, Jinwoo
Chung, Hyunsun
Cha, Myung-Hoon
Kim, Hong-Yeon
Kim, Youngjae
Machine Learning
Distributed, Parallel, and Cluster Computing
LLM inference is essential for applications like text summarization, translation, and data analysis, but the high cost of GPU instances from Cloud Service Providers (CSPs) like AWS is a major burden. This paper proposes InferSave, a cost-efficient VM selection framework for cloud based LLM inference. InferSave optimizes KV cache offloading based on Service Level Objectives (SLOs) and workload charac teristics, estimating GPU memory needs, and recommending cost-effective VM instances. Additionally, the Compute Time Calibration Function (CTCF) improves instance selection accuracy by adjusting for discrepancies between theoretical and actual GPU performance. Experiments on AWS GPU instances show that selecting lower-cost instances without KV cache offloading improves cost efficiency by up to 73.7% for online workloads, while KV cache offloading saves up to 20.19% for offline workloads.
title Cost-Efficient LLM Serving in the Cloud: VM Selection with KV Cache Offloading
topic Machine Learning
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2504.11816