返回全部动态

NVIDIA BlueField-4 驱动 Scale-In 网络,加速代理式 AI 工厂基础设施

原标题:NVIDIA BlueField-4 Powers New Scale-In Network Infrastructure for Agentic AI Factories

NVIDIA Technical Blog一手来源产品发布质量 83

AI 摘要

NVIDIA 推出第五大 AI 网络支柱 Scale-In,专为代理式 AI 工厂设计,由 BlueField-4 DPU、DOCA 软件和 Spectrum-X 以太网驱动,提供安全、多租户、高性能的数据访问和基础设施操作。该架构将南北向网络整合为统一加速域,减轻主机 CPU 负担,提升安全性和效率。BlueField-4 还支持 NVIDIA BlueField Astra 扩展控制平面,并与 CMX 协同管理 KV 缓存。

以上摘要由 AI 生成,可能存在误差。事实请以原文为准。

正文节选

Traditional cloud infrastructure was designed for predictable, general-purpose workloads and standard interfaces. Agentic AI factories connect diverse users, agents, applications, data sources, and storage systems to massively accelerated compute at multi-terabit bandwidth per server, making dedicated DPU processing essential for line-rate networking, storage, and security. NVIDIA is introducing Scale-In network infrastructure, the fifth pillar of NVIDIA AI networking, bringing purpose-built acc


发布时间:2026-08-24 23:00
抓取时间:2026-08-24 23:18
来源机构:NVIDIA
阅读原文developer.nvidia.com