KVCache Cache in the Wild: Characterizing and Optimizing KVCache Cache at a Large Cloud Provider

TL;DR

This paper systematically characterizes KV$ workload patterns in large language model serving at a major cloud provider, revealing insights that enable workload-aware cache eviction policies to improve performance.

Contribution

It provides the first detailed analysis of real-world KV$ workloads in LLM serving, informing cache management strategies for better efficiency.

Findings

KV$ reuses are skewed across requests

Reuse patterns are predictable within request categories

Moderate cache size suffices for high hit ratios

Abstract

Serving large language models (LLMs) is important for cloud providers, and caching intermediate results (KV\$) after processing each request substantially improves serving throughput and latency. However, there is limited understanding of how LLM serving benefits from KV\$ caching, where system design decisions like cache eviction policies are highly workload-dependent. In this paper, we present the first systematic characterization of the KV\$ workload patterns from one of the leading LLM service providers. We draw observations that were not covered by previous studies focusing on synthetic workloads, including: KV\$ reuses are skewed across requests, where reuses between single-turn requests are equally important as multi-turn requests; the reuse time and probability are diverse considering all requests, but for a specific request category, the pattern tends to be predictable; and the overall cache size required for an ideal cache hit ratio is moderate. Based on the characterization, we further propose a workload-aware cache eviction policy that improves the serving performance under real-world traces, especially with limited cache capacity.