arXiv cs.AI· Yunpeng Dong, Jingkai He, Shiqi Liu, Yuze Hou, Dong Du, Zhonghu Xu, Si Yu, Baochuan Yang, Yubin Xia, Haibo Chen·· 4 小时前AI 评分45
DeltaBox:毫秒级沙盒检查点与回滚技术,面向有状态 AI 智能体
DeltaBox: Millisecond-Scale Sandbox Checkpoint and Rollback for Stateful AI Agents
AI 导读
研究者提出 DeltaBox,一个具备毫秒级检查点/恢复能力的 AI 智能体沙盒,通过 DeltaState 抽象只复制相邻检查点之间的变化而非完整状态。其中 DeltaFS 以写入时复制方式检查点文件状态,DeltaCR 基于冻结模板进程快速恢复进程状态。
正文 · 原文
Abstract:Emerging LLM-powered AI agent workloads, such as test-time tree search and reinforcement learning (RL), require high-frequency state exploration, relying on rapid checkpoint/restore (C/R) of the complete sandbox state, including files and process state. Existing mechanisms duplicate the entire state, causing hundreds of milliseconds to seconds of latency per C/R, which severely bottlenecks deep search and large-scale fan-outs. We observe that consecutive sandbox states between checkpoints differ only marginally. Therefore, instead of full duplication, a sandbox should only duplicate the changes between consecutive checkpoints (Key Insight), yet the lack of OS support makes this hard to realize. We propose a new OS-level abstraction, DeltaState, that delivers change-based C/R for agents through co-designed OS mechanisms. First, DeltaFS checkpoints file state by freezing the writable layer and inserting a new one, turning updates into copy-on-write and rollback into a layer switch. Second, DeltaCR quickly restores process state by forking from a frozen template process. A coordinator pairs the two so each checkpoint is one consistent (filesystem, memory) state. Building on these mechanisms, we present DeltaBox, an agent sandbox with millisecond-level C/R. On SWE-bench, DeltaBox hides its checkpoint work (10.83 ms) under inference and rolls back in a mean 1.86 ms, with the same warm-template fork driving low-cost RL rollout fan-out. DeltaBox is open-sourced at this https URL.
| Comments: | 16 pages, 9 figures, 3 tables. Accepted to SIGOPS ATC 2026 |
| Subjects: | Operating Systems (cs.OS); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.22781 [cs.OS] |
| (or arXiv:2605.22781v3 [cs.OS] for this version) | |
| https://doi.org/10.48550/arXiv.2605.22781 arXiv-issued DOI via DataCite |
Submission history
From: Yunpeng Dong [view email]
[v1]
Thu, 21 May 2026 17:36:17 UTC (517 KB)
[v2]
Mon, 8 Jun 2026 12:58:55 UTC (393 KB)
[v3]
Fri, 13 Nov 2026 11:30:26 UTC (480 KB)
来源:arXiv cs.AI · arxiv.org