{"slug": "deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training", "title": "DeepSeek Elastic Compute:A Sandbox Infrastructure for Effective Agentic Training", "summary": "DeepSeek published a technical report on DeepSeek Elastic Compute (DSec), a production sandbox platform for large-scale agentic training and evaluation with large language models, submitted to arXiv on 19 Sep 2026. A single production-scale unit of DSec spans around 160 nodes, serves about 3 million sandboxes per day, supports over 380,000 concurrent sandboxes, and sustains over 5,000 sandbox creations per second. DSec exposes FnCall, container, microVM, and full-VM sandbox backends through a unified SDK, loads image data on demand from the Fire-Flyer File System (3FS), and is co-designed with the reinforcement learning framework to decouple stateful rollout execution from preemptible GPU training and mitigate agent misbehavior such as reward hacking.", "body_md": "# Computer Science > Distributed, Parallel, and Cluster Computing\n\n  [Submitted on 19 Sep 2026]\n\n# Title:DeepSeek Elastic Compute (DSec): A Sandbox Infrastructure for Effective Agentic Training at Scale\n\n[View PDF](https://arxiv.org/pdf/2609.22978)\n\n[HTML (experimental)](https://arxiv.org/html/2609.22978v1)\n\nAbstract:Large-scale agentic training and evaluation with large language models (LLMs) rely on isolated, stateful execution environments in which models inspect repositories, invoke tools, execute commands, and interact with task-specific services. These workloads create sandboxes in large bursts, span heterogeneous functionality and isolation requirements, retain state across long interactions, and draw from large image corpora with limited reuse. Supporting them therefore requires an elastic execution platform rather than a single sandbox runtime.\n\nThis report presents DeepSeek Elastic Compute (DSec), a production sandbox platform that exposes FnCall, container, microVM, and full-VM sandbox backends through a unified SDK. DSec coordinates placement and lifecycle management across the cluster, composes environments from independently versioned layers, combines memory sharing, reclamation, and CPU scheduling for high-density execution, and loads image data on demand from Fire-Flyer File System (3FS), a cluster-wide distributed filesystem. DSec is co-designed with the reinforcement learning (RL) framework, decouples stateful rollout execution from preemptible GPU training, coordinates sandbox lifecycle with training to preserve rollout state while reclaiming idle resources, and mitigates agent misbehavior such as reward hacking.\n\nA single production-scale unit of DSec spans around 160 nodes, serving about 3 million sandboxes per day; in production, it supports over 380,000 concurrent sandboxes and sustains over 5,000 sandbox creations per second. Our evaluation and deployment experience show that these mechanisms reduce environment setup and image-distribution overhead, improve memory efficiency, and preserve latency-sensitive performance under high-density overcommit.\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training", "canonical_source": "https://arxiv.org/abs/2609.22978", "published_at": "2026-09-22 19:14:41+00:00", "updated_at": "2026-09-22 19:24:15.194914+00:00", "lang": "en", "topics": ["ai-agents", "ai-infrastructure", "ai-research", "large-language-models", "mlops"], "entities": ["DeepSeek", "DeepSeek Elastic Compute (DSec)", "Fire-Flyer File System (3FS)", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training", "markdown": "https://wpnews.pro/news/deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training.md", "text": "https://wpnews.pro/news/deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training.txt", "jsonld": "https://wpnews.pro/news/deepseek-elastic-compute-a-sandbox-infrastructure-for-effective-agentic-training.jsonld"}}