Strata: Hierarchical Context Caching for Long Context Language Model Serving

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xie, Zhiqiang, Xu, Ziyi, Zhao, Mark, An, Yuwei, Mailthody, Vikram Sharma, Mahlke, Scott, Garland, Michael, Kozyrakis, Christos
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916917970731008
author Xie, Zhiqiang
Xu, Ziyi
Zhao, Mark
An, Yuwei
Mailthody, Vikram Sharma
Mahlke, Scott
Garland, Michael
Kozyrakis, Christos
author_facet Xie, Zhiqiang
Xu, Ziyi
Zhao, Mark
An, Yuwei
Mailthody, Vikram Sharma
Mahlke, Scott
Garland, Michael
Kozyrakis, Christos
contents Large Language Models (LLMs) with expanding context windows face significant performance hurdles. While caching key-value (KV) states is critical for avoiding redundant computation, the storage footprint of long-context caches quickly exceeds GPU memory capacity, forcing production systems to adopt hierarchical caching across memory hierarchies. However, transferring large cached contexts back to the GPU introduces severe performance bottlenecks: fragmented I/O from paged layouts prevents full bandwidth utilization, and existing schedulers fail to account for cache-loading delays, leaving systems loading-bound rather than compute-bound. We present Strata, a hierarchical context caching framework designed for efficient long context LLM serving. Strata introduces GPU-assisted I/O to combat KV cache fragmentation, decoupling GPU and CPU memory layouts and employs cache-aware request scheduling to balance compute with I/O latency and overlapping unavoidable stalls with complementary tasks. Built on SGLang and deployed in production, Strata achieves up to 5x lower Time-To-First-Token (TTFT) compared to vLLM + LMCache and 3.75x speedup over NVIDIA TensorRT-LLM on long-context benchmarks, without degrading short-context performance.
format Preprint
id arxiv_https___arxiv_org_abs_2508_18572
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Strata: Hierarchical Context Caching for Long Context Language Model Serving
Xie, Zhiqiang
Xu, Ziyi
Zhao, Mark
An, Yuwei
Mailthody, Vikram Sharma
Mahlke, Scott
Garland, Michael
Kozyrakis, Christos
Distributed, Parallel, and Cluster Computing
Large Language Models (LLMs) with expanding context windows face significant performance hurdles. While caching key-value (KV) states is critical for avoiding redundant computation, the storage footprint of long-context caches quickly exceeds GPU memory capacity, forcing production systems to adopt hierarchical caching across memory hierarchies. However, transferring large cached contexts back to the GPU introduces severe performance bottlenecks: fragmented I/O from paged layouts prevents full bandwidth utilization, and existing schedulers fail to account for cache-loading delays, leaving systems loading-bound rather than compute-bound. We present Strata, a hierarchical context caching framework designed for efficient long context LLM serving. Strata introduces GPU-assisted I/O to combat KV cache fragmentation, decoupling GPU and CPU memory layouts and employs cache-aware request scheduling to balance compute with I/O latency and overlapping unavoidable stalls with complementary tasks. Built on SGLang and deployed in production, Strata achieves up to 5x lower Time-To-First-Token (TTFT) compared to vLLM + LMCache and 3.75x speedup over NVIDIA TensorRT-LLM on long-context benchmarks, without degrading short-context performance.
title Strata: Hierarchical Context Caching for Long Context Language Model Serving
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2508.18572