arXiv 论文解析前缀缓存下 LLM 智能体服务的持续退化机制与解法
Bursts That Outlast Themselves: Persistent Degradation in Prefix-Cached Agent Serving
AI 导读
Wen Huang 与 Jack Liu 的 arXiv 论文（arXiv:2611.13412）表明，prefix caching 下的 LLM 智能体服务处于亚稳态，几分钟额外突发流量可使 vLLM 服务器进入退化区间，命中率降至约 10%、吞吐减半且积压持续增长 70 分钟。
正文 · 原文
Abstract:LLM agents drive inference engines with closed-loop sessions whose context grows every turn, and serving them economically depends on prefix caching: a hit costs an incremental prefill of about a thousand tokens, a miss recomputes a 10--40K-token context. We show that this makes prefix-cached agent serving metastable. With an identical base workload, a few minutes of extra arrivals leaves a vLLM server in a degraded regime--a ~10% hit ratio, about half the throughput, and a backlog that grows for the remaining 70 minutes--while its no-burst twin stays healthy (9 of 12 pairs; none in the reverse direction). The regime appears at loads the cache-warm server sustains, is absent without prefix caching, and also appears with Qwen3-32B on SWE-bench agent trace shapes and on SGLang, which tolerates larger triggers. Slower service keeps more sessions in the system, their working set outgrows the KV pool, and every turn becomes a recomputation; a first-order workload-balance condition, calibrated independently, brackets when the regime can drain. Getting out is the harder problem: throttling new admissions recovered 1 of 7 matched collapsed runs, while suspending admitted sessions against a capacity-based limit recovered all of them within minutes. A CPU offload tier delays the collapse but does not prevent it. A budgeter that bounds the resident KV demand of active sessions to the pool, with thresholds that need no per-workload tuning, prevented the failure in every run we tested.
| Comments: | 12 pages, 8 figures |
| Subjects: | Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.13412 [cs.DC] |
| (or arXiv:2611.13412v1 [cs.DC] for this version) | |
| https://doi.org/10.48550/arXiv.2611.13412 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wen Huang [view email]
[v1]
Tue, 10 Nov 2026 00:46:51 UTC (153 KB)
来源：arXiv cs.AI · arxiv.org