paperbot · PL 论文追踪

全部论文 待读 收藏

P2 LeapIO: Efficient and Portable Virtual NVMe Storage on ARM SoCs

LeapIO:在 ARM SoC 上高效且可移植的虚拟 NVMe 存储
ASPLOS Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Systems2020匹配度 0/10 · 该论文研究基于 ARM 协处理器的虚拟 NVMe 云存储卸载与系统性能优化,属于体系结构/存储系统方向,与用户关注的 PL 与 AI 交叉、静态分析、浮点程序分析、贝叶斯程序分析及 LLM/Agent 等方向均无直接关联。引用 59 inbox
Huaicheng Li, Mingzhe Hao, Stanko Novakovic, Vaibhav Gogte, Sriram Govindan, Dan R. K. Ports, Irene Zhang, Ricardo Bianchini, Haryadi S. Gunawi, Anirudh Badam
crossref: 2026-09-12

AI 速览

一句话:LeapIO 利用 ARM 协处理器卸载复杂云存储服务,向未修改 guest VM 提供虚拟 NVMe,并保持接近裸机的性能。

云存储栈非常耗费资源,会占用数据中心 10–20% 的 x86 核心,是云厂商必须承担的“存储税”,但现有 IO 加速器尚未完全具备卸载复杂云存储栈的条件。LeapIO 提出利用基于 ARM 的协处理器来卸载复杂存储服务,并通过一组 OS/软件技术及新的硬件特性,在 x86 与 ARM 核心之间提供统一地址空间,同时向未修改的 guest VM 暴露虚拟 NVMe 存储。该工作还应对硬件可替换性、软件可移植性、虚拟化、可组合性与效率等部署挑战。摘要报告其性能可与裸金属服务器竞争。

问题
云存储栈资源消耗巨大,占用大量数据中心 x86 核心,且难以完全卸载到当前 IO 加速器。
方法
LeapIO 借助 ARM 协处理器卸载复杂存储服务,结合 OS/软件技术与硬件特性,在 x86 和 ARM 核心间建立统一地址空间,并向未修改 guest VM 暴露虚拟 NVMe。
评估
摘要报告 LeapIO 的性能可与裸金属服务器竞争,并应对硬件可替换性、软件可移植性、虚拟化、可组合性与效率等部署挑战。
原文摘要(Abstract)

Today's cloud storage stack is extremely resource hungry, burning 10-20% of datacenter x86 cores, a major "storage tax" that cloud providers must pay. Yet, the complex cloud storage stack is not completely offload-ready to today's IO accelerators. We present LeapIO, a new cloud storage stack that leverages ARM-based co-processors to offload complex storage services. LeapIO addresses many deployment challenges, such as hardware fungibility, software portability, virtualizability, composability, and efficiency. It uses a set of OS/software techniques and new hardware properties that provide a uni- form address space across the x86 and ARM cores and ex- pose virtual NVMe storage to unmodified guest VMs, at a performance that is competitive with bare-metal servers.

链接与引用

DOI 原文 · PDF(开放获取) ·

BibTeX
@inproceedings{paperbot5412,
  title = {LeapIO: Efficient and Portable Virtual NVMe Storage on ARM SoCs},
  author = {Huaicheng Li and Mingzhe Hao and Stanko Novakovic and Vaibhav Gogte and Sriram Govindan and Dan R. K. Ports and Irene Zhang and Ricardo Bianchini and Haryadi S. Gunawi and Anirudh Badam},
  booktitle = {ASPLOS},
  year = {2020},
  doi = {10.1145/3373376.3378531}
}